Inside the Minds of Government Chief Data and Artificial Intelligence Officers:

Dr. Adita Karkera recently sat down with Neil Cooke to discuss Texas’s data leadership journey and his work as the state’s Chief Data Officer (CDO).

Data Alone Isn't the Story

Texas's chief data office is led today by Neil Cooke, whose path to the role runs through data analyst, business analyst, and systems analyst positions in both the UK and the US — practical experience at the intersection of business, systems, and the data underneath them. That foundation taught an early lesson that still shapes the office's approach: data alone is only part of the story. Understanding a program's goals, its business processes, and how information is captured, reported, and used for decision-making matters just as much. That perspective drives a recurring set of questions the office still asks today: what matters most, how is it measured, where does the data live, does it exist, who owns it, and can people trust it? Bridging the gap between business and technical teams remains one of the office's biggest accomplishments — translating business needs into practical, useful data solutions. Recurring challenges around siloed data, inconsistent definitions, and low trust in the numbers reinforce a core lesson for the office: leading data in Texas is rarely about the technology itself. It's about governance, communication, relationships, and creating real business value.

Leading Without Losing Momentum

Texas's chief data office is now four years into Cooke's tenure as CDO, after he stepped in as interim CDO at a moment when the office already carried strong momentum and a long history of achievement. The first priority was protecting that momentum while shifting from the deep program knowledge of an agency data officer to the broader, statewide view required of a state CDO. That has meant staying visible, listening to teams, and removing blockers, while keeping delivery clear on priorities like data governance, data sharing, open data, and data literacy — all while continuing to support the data management officers who realize statewide priorities.

A Structure Unlike Any Other State

The Texas chief data office sits within the Texas Department of Information Resources (DIR), the state's IT agency, which reports to a board that in turn reports to the governor — a structure Cooke notes is unique compared to other states.

Trust Is the Real Skill

Effective data leadership in Texas goes well beyond technical expertise, Cooke says — business acumen, governance discipline, communication, and change management matter just as much, since the work is really about translating data into decisions, outcomes, and actions across executives, program owners, IT, legal, privacy, security, and sometimes lawmakers. Public sector data leadership carries a heavier emphasis on trust, transparency, stewardship, and statutory responsibility, along with the coordination challenge of working across agencies and programs where definitions, systems, and authorities are not always aligned — a contrast with the private sector, where the pressure is more often about speed, customer insight, operational efficiency, revenue growth, and measurable ROI. The two sectors are converging, however, around data quality, cybersecurity, AI governance, interoperability, and measurable outcomes. Cooke notes that the most important skill for a CDO today may simply be the ability to build trust — because if people don't trust the data, the governance, or the decisions being made, the data doesn't get used.

About Molly Just Behr

•  Adita Karkera, Chief Data Officer for Deloitte’s Government and Public Services, recently sat down with Molly Just Behr, Databricks’ Global Head of Public Sector for Go-to-Market, to explore what it takes to move government agencies from fragmented data environments to a truly AI-ready foundation. Molly brings 15 years of experience at the intersection of government and technology, having served in government, a public-private partnership, and management consulting — a combination that gave her a grounded understanding of both public sector constraints and the innovation happening in industry. Today, she leads Databricks’ global efforts to help governments modernize with a unified data and AI platform.

The Databricks Platform: Doing More with Less

• A recurring theme in the conversation was the pressure on government agencies to “do more with less.” Molly described how Databricks addresses this through Genie — an AI agent that lets any employee chat with their data and get trusted answers instantly, without needing to submit a request and waiting weeks for a formal report to come back. The broader Databricks platform combines data storage, analytics, and AI into a single governed environment through four key products: Genie for conversational data access; Lakebase, the database backbone for the AI era; Agent Bricks for building AI agents on agency data; and Lakewatch for AI- owered cybersecurity threat detection. Together they serve over 80 percent of federal executive departments and more than 400 federal, state, local, and education organizations.

The First 60–90 Days: From Strategy to Impact

• For agencies starting their AI journey, Molly’s practical advice is to secure executive buy-in and identify a single mission outcome to pilot. Real-world results underscore the value: the U.S. Postal Service centralized oversight of 130 billion pieces of mail annually on Databricks, cutting total cost of ownership by 40 percent and enabling real-time fraud detection. At the Centers for Medicare and Medicaid Services, unified beneficiary and provider views reduced fraud, waste, and abuse while improving care coordination. A successful pilot, Molly noted, generates the momentum that propels the entire AI journey forward.

Speaker + Episode/Part details

• Host: Dr. Adita Karkera, Chief Data Officer,
  Government & Public Services, U.S. Deloitte Consulting LLP
• Guest: Molly Just Behr, Global Head of Public
  Sector, Go-to-Market, Databricks
• Episode: Part 1 of 2 — Building the Government
  AI Foundation

Closing the Data Readiness Gap

• In the second part of her conversation with Adita Karkera, Molly Just Behr of Databricks turns to the deeper enablers of a successful government data and AI strategy: governance, trust, and workforce readiness. Molly argued that the biggest gap between wanting AI and realizing AI is data readiness — and that closing it is more achievable than most agencies think. The key steps include metadata tagging, PII identification through automation, and establishing the right identity access controls. Tools like Databricks’ Unity Catalog accelerate this by centralizing the data catalog and providing governance across both structured and unstructured data.

Building Trust Through Governance

• Molly pushed back on the idea that moving fast means cutting corners on accountability — in her view, a strong governance foundation is precisely what enables agencies to act with speed and confidence. Databricks invests in compliance frameworks including FedRAMP and DoD Impact Level 5 to give customers a platform they can trust. Clear decision rights, strict identity access controls, and documented guidelines on who can approve datasets, use models, and access use cases are all essential. That governance infrastructure has earned the trust of more than 20,000 global customers and over 400 federal, state, and local public sector organizations.

Investing in the Workforce

• Workforce readiness is equally critical, and Molly was clear that technology alone will not drive change. Databricks backs this belief with a $100 million investment in the next generation of data and AI talent and the launch of Databricks Free Edition, giving everyone from students to senior executives free access to the platform and training through the Databricks Academy. The goal: let government teams build skills and experiment with real-world use cases on the same platform they will use in production.

Looking Ahead: A Shared Data Future

• Looking three to five years ahead, Molly’s vision is a public sector where data is a shared asset — not a siloed resource — and where public servants use AI as a practical daily tool. Her two most important moves: invest in a unified government data foundation and tie every AI initiative to a clear mission objective. The conversation closed with a signature reframe: instead of building an AI strategy, agencies need a data strategy with AI built in. “Start with the data,” Molly said, “the AI will follow naturally when the foundation is right.”

Speaker + Episode/Part details

• Host: Dr. Adita Karkera, Chief Data Officer, Government & Public Services, U.S. Deloitte Consulting LLP
• Guest: Molly Just Behr, Global Head of Public Sector, Go-to-Market, Databricks
• Episode: Part 2 of 2

Adita Karkera recently sat down with Mike Capone, former Chief Executive Officer of Qlik, to discuss his leadership journey and his perspective on what it really takes for organizations to turn AI investments into measurable outcomes. Mike described himself as a non-traditional CEO whose path did not follow the customary route through finance or sales. He started his career as a programmer and went back to school to earn his MBA at night, volunteering for assignments where he could showcase his ability to merge technology with business outcomes. He transitioned from hardcore programming to packaged software implementations, then to leading enterprise IT at ADP, and ultimately to a CTO role overseeing a billion-and-a-half-dollar IT organization driving product development and operations. Mike noted that it is not often that CIOs or CTOs cross the chasm to become CEOs, though he believes that will become more common in the future. He took a deliberate step to a smaller, life-sciences-focused company, Medidata, to prove himself on the operating side of the business — a move that ultimately led him to Qlik. The thread running through every chapter of his career, Mike explained, has been the ability to synthesize what technology can really do to transform a business — and in today's world, that means data, analytics, and AI. Reflecting on the broader market, Mike observed that while everyone is embarking on an AI journey founded in data and analytics, not everyone is getting value. Disappointment around AI outcomes is widespread, and he believes the market is now discovering that the barrier to AI success is not the models themselves. As Mike put it, plugging in models is the easy part — the hard part is the data and the systems around the model. He emphasized that leaders must be urgent about building the right platform: data integration that harnesses data from everywhere, plus governance, indexing, and lineage. Mike cautioned against the "walled garden" approach where vendors say, "give me all your data and I'll give you the outcome you want" — an approach that rarely works. Instead, he advised meeting customers where they are, starting with a real business problem, and working backwards to see how data and AI can help. If the foundation is right, he said, the models take care of themselves. On the topic of agentic AI, Mike shared that organizations are closer to agentic than they think. Customers who have already built strong analytics foundations — even those running on dashboards or reports — can build on top of that platform to reach agentic capabilities faster. The alternative, tearing down trusted partnerships and ripping apart existing foundations to start something new, is both expensive and risky. Mike's advice: take the trusted workflows and analytics enterprises already have, and enable agentic capabilities on top of them.

Adita Karkera continued her conversation with Mike Capone, former Chief Executive Officer of Qlik, exploring his work on responsible AI, the role of literacy in scaling adoption, and how Qlik has embedded trust into its strategy and culture. Drawing on his experience as an advisor to the White House on responsible AI competitiveness, Mike shared a clear-eyed view of what it takes for AI to earn trust at both the enterprise and societal levels.

AI, Trust, and First Impressions

• Mike emphasized that trust is foundational and fragile. As he explained, if someone's first experience with a piece of technology is bad, it takes a long time to win them back — and AI is no different. If the first answers a user gets from an AI platform are inaccurate or not credible, it turns people off. Before rolling AI out to the masses, leaders must make sure the technology is trusted, the use cases are useful, and quick wins are visible. Mike cautioned against AI "haves and have-nots" inside an organization, and stressed that AI strategy has to lift the entire company — not just developers or one function.

Democratizing AI Beyond the Largest Companies

• Reflecting on his White House advisory role, Mike spoke to the broader societal responsibility of the private sector. The value of AI cannot accrue only to a handful of large companies — it has to reach the average citizen. He sees a clear obligation for the private sector to do its share in deploying AI responsibly so that capabilities are available to everyone.

From Data Literacy to AI Literacy

• Mike traced the evolution from general literacy to data literacy and now to AI literacy. Qlik built an early mission around helping people interpret, work with, and even argue with data — recognizing that data can be manipulated through scale changes or chart choices. That mission has now extended into AI literacy: training people to read the signals from a model, evaluate transparency, and maintain critical thinking. As Mike put it, the goal is acceleration without seeding complete control — using AI thoughtfully and responsibly without giving up the muscle to interrogate its outputs.

Building Trust Into Culture, Not Just Policy

• Mike described how Qlik built an external AI Council made up of advisors from outside the tech industry — people who work in government, education, and ethics — to ask the hard questions and keep the company honest. He warned against the failure mode of treating AI governance as a side committee that reviews policies but never shapes culture. To work, trust has to be embedded in everyday practice. At Qlik, his Chief Customer Officer holds a 30-minute weekly town hall with every employee in customer success to talk transparently about how AI is being used, how it will reshape jobs, and what employees can expect. Mike's principle: never tell people what they want to hear if it isn't true. Be transparent about what is changing — and support people through the journey.

In the closing part of her conversation with Mike Capone, former Chief Executive Officer of Qlik, Adita Karkera explored where AI is delivering the most value inside Qlik today, how AI has reshaped board-level conversations, the importance of trusted partnerships, and what Mike sees on the horizon for the analytics and data integration market.

Where AI Is Delivering Value Today

• Mike highlighted two areas where AI has delivered the strongest results inside Qlik. The first is customer success, where predictive AI tools observe how customers are using Qlik's products, identify signs of friction, and surface insights to customer success representatives — sometimes prompting outreach, sometimes offering five talking points before a conversation, and sometimes fully automating renewals. As Mike shared, roughly 80% of Qlik's renewals are now automated through agents, with the ability to spot opportunities for additional capabilities or capacity. The second is sales, where AI has been game-changing by giving salespeople a richer outside-in view of customers — what they are doing, how their business is performing, what vendors they use — so sales teams can tailor proposals around the customer's actual problems rather than generic product pitches.

AI in the Boardroom

• Mike described how AI has reshaped his board conversations. With Qlik backed by a private equity firm that owns 60 technology companies, his board is highly literate in AI and contributes actively to strategy — sharing best practices across portfolio companies and pushing for rapid innovation. Mike's advice for any CEO: do not appear to be lagging on AI. While he resists the idea that every CEO must personally be the Chief AI Officer, he believes leaders should be deeply engaged, hire smart people around them, and spend hours each day on AI strategy and external learning.

The Power of Trusted Partnerships

• Mike reflected that survival in today's environment requires a rich ecosystem of trusted partners. He highlighted the partnership between Qlik and Deloitte as an example — combining Qlik's technology and innovation with Deloitte's transformation and strategy capabilities to deliver real outcomes for customers in the public sector and beyond. Mike's view is that customers want fewer, better, more trusted partners — and that walking into a customer engagement together with a shared vision of doing the best thing for the customer is what makes the difference.

Looking Ahead: The Next Three to Five Years

• Closing the conversation, Mike offered a clear prediction for the analytics and data integration market. The next phase of AI will belong to organizations that can make AI useful across their entire enterprise — pressure-tested under real conditions, explainable when challenged, durable rather than fragile, and adaptable to a world that will keep changing in ways no one can fully predict. As Mike put it, the only thing we know for sure is that we don't know everything that's coming — so flexibility, openness, and the ability to adapt will define the companies that succeed.

Adita Karkera recently interviewed Deepak Shah to discuss his experiences and role as the Chief Data Officer (CDO) for the U.S. Department of War, Western Hemisphere Command. Deepak Shah started his career over 30 years ago leading data and technology initiatives in global investment banking, asset management, and capital markets, which involved developing timely reporting systems across multiple asset classes to bridge the gap between business and technology programs. This involved building enterprise data hubs and warehouses that standardized and streamlined data across his organization, leading to larger stakeholder alignment. These experiences molded him into a “data centric leader” and Deepak mentioned a key turning point in his career that happened during his time at Carnegie Mellon’s CDO Executive Program when he published an article on data governance that gained attention of leaders within the Department of Defense. This recognition ultimately led to his appointment as CDO of Forces Command which has now transitioned into the Western Hemisphere Command. In data management, the vision defines the enterprise-level outcome you’re steering toward, while the mission translates that vision into operating commitments. Deepak shared that this is where most projects either succeed or “quietly fail” before they even start. He kicks every initiative off with a clear vision, mission, and set of goals, and without them, execution stagnates. Deepak states: “vision is the north star - without it, initiatives become disconnected projects.” Deepak emphasized that vision establishes three things: it sets the enterprise framework, translates discipline execution, and aligns organizations around a very common direction (clear vision across domains reduces duplication and gives teams something to rally around). Deepak followed this up by explaining that mission turns a vision into a flight plan and without it “vision stays aspirational, execution shifts, and priorities will fragment”. He affirmed that for a CDO, mission is simple – deliver trusted, decision-ready data so leaders can operate with speed, confidence and accountability. Finally, goals allow organizations to measure adoption, quality, and efficiency, which builds trust and sustains momentum. Clear vision, mission, and goals help an enterprise data strategy become an engine that turns data and AI into true enterprise capabilities. Deepak shared his interpretation of the phrase “boiling the ocean” with Adita as he explained “you hear this phrase a lot in transformation efforts, but it only works, in my experience, if leaders apply real discipline behind it. Don't boil the ocean doesn't mean that the ocean is not important. It only means that trying to fix everything in one shot is neither practical nor realistic”. Deepak advised that he always starts his initiatives with use cases to address this issue by establishing a standardized use case intake template with consistent parameters tailored to different domains. That framework can then be published to business leaders across the organization. This structure creates discipline from the start”. By using a very transparent scoring model (involving business value, regulatory impact, risk reduction and operational efficiency), leaders can then prioritize those use cases by comparing and choosing the ones that deliver the greatest enterprise impact. Deepak continued that “this is where solve and scale becomes real” as it ensures solving the right business problem first and establishing reusable data foundations that can enable scaling across the enterprise. 

In this episode of Deloitte's CDO Perspectives series, Deepak described frameworks, playbooks, and runbooks as core strategy pillars that make delivery repeatable and scalable across the enterprise:

• Frameworks define how work is initiated, governed, and how success is measured.
• Playbooks capture and document best practices and turn proven approaches into reusable patterns, so teams do not have to start from scratch.
• Runbooks provide step-by-step instructions so trained team members can reliably execute critical tasks without relying on individual expertise.

Trustworthy and Strong Data Foundations

Deepak anchors data trust in three areas:

1. Establishing a Baseline Understanding: Many organizations assume they understand their data landscape, only to discover fragmentation, shadow systems, and blind spots when scaling AI initiatives. Metadata tagging, classification, and labelling clarify what data exists, where it resides, and how it can be used. “I’ve seen organizations invest heavily in AI platforms only to discover later that they lack even a basic inventory of their data.” For defense operations, this baseline understanding is mission-critical: unclear data provenance undermines decision confidence.
2. Quality and consistency: Operational commanders depend on consistent, reliable intelligence and analytics. “AI often looks promising in a lab or on paper but fails in production.” Curation, standardization, reference alignment, and continuous monitoring prevent fragmented inputs and failures in production. 
3. Accountability and traceability: Lineage, ownership, and clear controls answer two questions: Where did the data come from? Who is accountable? Many data platforms now embed lineage capabilities, but trust is not created by tooling alone, it’s engineered through a strong data foundation. For defense leaders, trusted data becomes a decisive operational advantage.

Final Thoughts

“AI success is not about chasing the next model or the latest AI platform. It is about building the data foundations, governance, and execution disciplines that create trust and enable scale over time.” Organizations that invest in robust foundations and operationalize execution through repeatable frameworks will adopt AI faster and more responsibly. 

Adita Karkera recently sat down with Karthik Yajurvedi to discuss his experiences in his role as the Chief Data Officer (CDO) for the Commonwealth of Massachusetts. Karthik’s educational background involved electronics and communications engineering, but many of his professional projects have involved building software. His career started in consulting for both the public and private sectors developing software. There, he found it important that he and his team were not just building technology but also focusing on answering the business questions being asked and learning to understand the desires of the stakeholders. After consulting, he transitioned to a position in financial services, where he learned new skills such as risk management, compliance, and data management at a scale. Here, the timeliness of data was integral to his job. Finally, he headed back to the government in a full-circle moment, where all his successes and learnings prepared him for his current position as CDO. Karthik has found his most useful skills are talking to the business, trust & risk management, and enterprise transformation, which translates to developing people skills to manage huge programs. Karthik noted that the public and private sectors have many things in common, including and alignment with the business, collaboration, and communication to ensure teams solve the right problems, and without this, solutions would not be scalable or sustainable. He mentions, however, that the differences lie in their priorities: in the private sector it is important to align with shareholders and revenue streams, but in government, it is important to meet the constituents’ needs. Additionally, within the private sector, companies could collaborate with marketing or sales, while in the government, collaboration happens with secretaries, agencies, the governor’s office, the quasi-governments, and higher education, but clear, concise communication is important for both sectors. Another commonality is the importance of data storytelling; this rings true for the CDO role, as one must be able to articulate the narrative of what the data is depicting.  Expanding on the CDO role in the Commonwealth of Massachusetts, Karthik explains he reports to the CIO Secretariat, Secretary Snyder, who is the secretary of Executive Office of Technology Services and Security (EOTSS) and reports to the governor, who runs the Secretariat, which is one of 11 Secretariats in the Commonwealth, which is unique, as many other states do not share this structure. Karthik articulates the CDO role is about creation and management of the Commonwealth’s analytics strategy, and the written policy standards, guardrails, and guidelines that go along with that. Additionally, the storage, collection, dissemination of data, and data retention around data assets & analytical assets are also extremely critical. Lastly and most importantly, the CDO must facilitate maturing the data and analytics culture within the Commonwealth by partnering with the secretariats and agencies as well as partnering outside of the executive offices with entities such as quasi-governments, higher education institutions, and the private sector, to ensure they are creating a vibrant data and analytics ecosystem for the state.  

Adita Karkera recently sat down with Karthik Yajurvedi to discuss his experiences in his role as the Chief Data Officer (CDO) for the Commonwealth of Massachusetts.

When asked about his thoughts on secure data sharing, Karthik shared that it is a cornerstone for cross-agency analytics. He explained that if you look at any program and services that need to be offered to the residents of the state, most of the time they are not restricted to one agency as each agency has a piece of the puzzle, but not the full picture. Because of this, he highlighted that it is essential to be able to bring various parts of the data together for constituents in such a way for analytics, that the policy makers can have the insights they need for actional, operational decision-making. Karthik went on to say that compliance risks and lack of trust occur if data is not shared securely, and the Commonwealth has invested in modern platforms where agencies do not need to move data to prevent this. It has become simple for agencies to distribute data since the modern platform allows data sharing with role-based access.

The Commonwealth has also populated an enterprise data catalog that helps data discovery as each agency can contribute their data attributes, add metadata, and build a semantic layer on top of that metadata to build responsible AI. The Commonwealth is working on many cross-agency use cases, and all of them are only possible through secure data sharing. Karthik highlighted two examples: the first is the Early Childhood Integrated Data System (ECIDS), which brought the policy leaders from seven secretariats together to help English learners from birth to age five. The goal was to ensure the programs the agencies had were truly meeting the needs of the young learners and positioning them for school effectively. The second example Karthik shared is Learn to Earn, or LTE, where five secretariats came together to create a comprehensive approach to provide unemployed or underemployed residents of MA with the ability to find and retain jobs. The program will provide individuals who receive public assistance, the support, skills, and credentials they need to gain employment and achieve economic sustainability for themselves and their families. Karthik added that the Commonwealth is working on many other use cases.

Adita Karkera recently sat down with Karthik Yajurvedi to discuss his experiences in his role as the Chief Data Officer (CDO) for the Commonwealth of Massachusetts.

Karthik was asked about unique challenges he faced and how he was able to overcome them to get data sharing into place. The CDO explained that one challenge the Commonwealth is currently facing is to bring programs out of silos, while preparing data for advanced analytics and AI. He elaborated by saying: “it’s not necessarily that the agencies don’t want to share data”, but that it is important to articulate to the agencies what the value and shared purpose is, since, as he said, the end customer for government is the citizen, and the shared purpose for everyone is to serve the citizen. Karthik stated that educating people on that shared purpose is one way to break the silos and rally the troops to work together. He also added that compliance, ethics, and equity are big concerns when preparing data for advanced analytics.

The CDO was then asked if a governance effort was required to manage initiatives. He stated that governance is important, and that it is essential to bring everyone to consistent standards for data sharing, but it could be construed as a compliance function, and the Commonwealth wants to be an enabler of data rather than a bottleneck. Because of this, they adapted a data strategy where guidelines come from a central office, but they still encourage the agencies to innovate and be flexible. Karthik explained that they collaborate with the agencies and meet them where they are. He also added that everyone is aware that it is important to follow best practices, which are based on the use case and thus can differ, meaning that lots of data literacy and education are required in the space. The Commonwealth meets regularly to discuss common pain points and how to solve them.Speaking about his own journey, Karthik shared the following areas that have helped him succeed in the role. First, he stressed the importance of strategic thinking to handle the fluidity of the role, adding that being able to anticipate change, manage risk, and do scenario planning is important because every use case is different, and data is everywhere. Second, he stressed the importance of learning the business and understanding the stakeholders to identify what problems need to be solved, and how data can solve those problems and be extracted as an asset, explaining that “learning the business up to some level is super important. This is not a technology job. It is truly a business, organization, and cross-cutting function.

Adita Karkera recently sat down with Dr. Matthew McCarville to discuss his experiences in his role as the Chief Information Officer (CIO) for the State of Nebraska. Dr. McCarville’s career began in finance, where he earned his MBA and worked at a hedge fund until it dissolved during the 2008 financial crisis. He pursued further education to deepen his existing data analysis skills, earning an MS in Business Intelligence, an MS in Software Development, and a PhD in applied AI and business intelligence. After that, Dr. McCarville joined the Union Pacific Railroad, where he modernized IT infrastructure and advanced data analytics on fuel. As AI became more prevalent in the transportation industry, he began using predictive analytics to estimate locomotive spending. He then moved to PwC, becoming their first Chief Data Scientist for Human Capital, Metrics & Analytics at their international office. After his time at PwC, Dr. McCarville relocated to the University of Florida, where his wife was also based. There, he was appointed as the university’s first Chief Data Officer (CDO) and concurrently served as a professor in data analytics and artificial intelligence. In December 2019, he filled the newly created Chief Data Officer position for the State of Florida, again collaborating with his wife, who became the Deputy Director of HR. Dr. McCarville experienced a “massive pivot” when COVID-19 hit in March 2020. With Florida’s CIO and CISO roles vacated, he was appointed interim for both positions. During this period, he employed robotic process automation (RPA), conducted data analytics, and used automation to streamline the unemployment process for constituents and support those delivering the services. During this period, he developed dashboards to measure key metrics in health, education, and unemployment, describing the experience as “a deep immersion in a lot of broad areas” that he would not have otherwise encountered. He also developed an enterprise data catalog, which enabled metadata management and consolidated siloed agency data into a repository for analytics, AI, and dashboarding KPIs for reports flowing into the governor’s office. After his tenure in Florida, Dr. McCarville went on to become the CIO for the University of Colorado. He then became the CIO for the state of Nebraska, where he began building AI and BI platforms and cybersecurity capabilities, all from scratch, while maintaining the state’s existing infrastructure.

Strategies for Collaboration between the CDO and CIO
When asked about improving collaboration between the CDO and CIO, Dr. McCarville stressed the importance of measurement, citing Nebraska's governor: “You can’t analyze and improve things that you’re not measuring and in order to measure things, you need data.” He explained that data is vital for decision-making, and he cannot make decisions unless he fully understands the impacts of initiatives as shown through government data—specifically, the people affected, and the outcomes achieved. Dr. McCarville noted that Nebraska has never paid for technologies like AI and cybersecurity, so the state must use data as an asset to scale up in the most impactful way. As Nebraska migrates to the cloud, the state’s IT leaders have recognized the critical need for comprehensive data cleansing to reduce operational costs. This initiative underscores the importance of data in the decision-making process.

How the CIO Role Fits into Nebraska’s Hierarchy
Dr. Matthew McCarville explained that his office operates under the shared Department of Administrative or Management Services and that he serves as a cabinet member for the governor. The Cabinet meets monthly, with each member treated as an independent leader of their agency. All agencies must use the CIO’s office for IT needs unless there is a specific exception.

Challenges Around Data Sharing and Department Collaboration
When asked about some of the challenges surrounding data sharing, Dr. McCarville highlighted the importance of cross-agency collaboration. He pointed out that in Nebraska, services like unemployment, Medicare, and SNAP are managed by separate agencies, requiring citizens to apply to each individually. This fragmentation displayed how data sharing would simplify constituent interactions and improve agency efficiency. He pointed out that departmental silos and shadow IT systems are ongoing challenges, with agencies often duplicating efforts. Dr. McCarville also noted that Nebraska’s budget deficit has inadvertently promoted centralization and efficiency: “It kind of forced the view of centralization and efficiency because they can’t all afford to go rogue and do things on their own anymore…they’re being forced to cut things and be more efficient and centralized, it’s not a mandate on centralization”. Because of this situation, Nebraska began to modernize its systems, which led to improved reporting and greater data quality. Silos previously hindered updates and allowed vendors to sell the same products to multiple agencies, but the push for efficiency drove a modernization mindset without strict mandates.

Cybersecurity in the State of Nebraska
On the topic of cybersecurity, Dr. McCarville shared that Nebraska previously relied on a single policy analyst to manage infrastructure security with firewalls. In the past 18 months, the state ramped up its efforts, becoming one of the first to adopt a whole-of-state model: “We have 93 counties and 90% of our state is rural small towns. The city councils are farmers, shop owners, and lawyers in the town…How we move forward here in Nebraska from a cybersecurity perspective is [by asking ourselves] 'How do we take care of those that do not have funding to do it on their own?’” To address this situation, Nebraska offset cybersecurity costs by pooling resources, platforms, and data management across agencies. The state modernized its approach by developing a joint Security Operations Center (SOC) and dedicating a 5,000 sq ft space in the CIO building, using monitors to display metrics, analytics, and KPIs. Nebraska also created Network Nebraska, a shared broadband backbone connecting schools, universities, and libraries, supported by RFPs and vendors for cybersecurity implementation. Additionally, the state is rolling out a “train the trainer” program, bringing in vendors to educate and certify counties and constituents on best cybersecurity practices, fostering collaboration in the whole-of-state model.

Promoting Data Talent in Nebraska
Dr. Matthew McCarville discussed how Nebraska is cultivating young talent in data analytics. He shared the following on how his experience as a professor encouraged him to hire former students as interns: “Some of those interns have worked for me four or five different places and that really got me to understand the impact and hunger of early career talent”. He witnessed firsthand the benefits of providing interns with the right tools and opportunities, which in turn foster growth across both private and public sectors. Nebraska leverages a split intern model, where interns spend two days a week supporting K-12 schools and two days assisting universities. This approach includes engaging high school students and teaching them tools like Tableau and Power BI. Dr. McCarville emphasized the need for data expertise from interns in academia: “The state’s sitting on mountains of data that does not get analyzed because we just don’t have the staff, so how do I leverage the research capacity of our institutions… If they can show us how to improve Medicare or Medicaid services or any entitlement program by researching our data to see where we're lacking and where those gaps could be, that's a win-win for the state, for the research world, for the applied and theoretical academic research titles”.

Success as CIO of Nebraska
Reflecting on his 18 months as CIO, Dr. McCarville spoke of his commitment to understanding Nebraska’s history by reviewing contracts and vendor relationships. He found his staff eager to modernize, despite previous limitations due to budget and resources. Now, he aims to treat vendors as long-term partners and, where possible, hire interns to drive progress. Dr. McCarville shared: “We are working with the local strategic air command air force base and the feds to be able to pump our interns through cyber and AI top secret security clearance programs so that they can become that top tier talent with a basis formed in their education and applied understanding at the State of Nebraska. So, we become that avenue that they are benchmarking our standards for education and training”.

Advice for Future Data Professionals
To conclude the conversation, Dr. McCarville encouraged future professionals to consider public sector roles, describing them as “government startups” where individuals can make an immediate impact using data, AI, and analytics. He stressed the importance of AI governance and ethics: “We're working with vendors that help us with the governance as much as we are working with the data cleansing and the enablement because they all want to enable it, but they don't want to think about the ramifications if you do things outside of ethical or legal boundaries. Not every amazing use case for AI is legally able to be implemented in the public sector.” The CIO finished the interview with advice for new practitioners: “We can do a lot of those things that students would not necessarily have the ability in their first job in corporate or consulting to be able to jump right into. They would have years of working their way up into that. Whereas in government, we can throw them right in the deep end and say, hey, we're going to swim. We're going to swim right now. You're going to help us get there.”

We are proud to spotlight Dr. Adita Karkera, Chief Data Officer at Deloitte Consulting LLP, and Michelle Boston, CIO, Head of Data Management Technology & Enterprise Architecture at Bank of America, as honorees in the CDO Magazine Global Data Power Women series.

In their recent interview, led by Amy Horowitz, Group Vice President at Informatica, Dr. Karkera and Michelle Boston offered valuable perspectives on the changing landscape of data leadership. They discussed moving from compliance-focused governance to adaptive, mission-driven strategies, with a focus on integrating AI-specific considerations such as explainability and bias. Their conversation addressed the challenges of advancing AI initiatives, balancing innovation with effective governance, and engaging stakeholders across organizations. These insights are essential for navigating the complexities of modern data and AI leadership.

We are pleased to present Part Two of our series featuring Dr. Adita Karkera, Chief Data Officer at Deloitte Consulting LLP, and Michelle Boston, CIO and Head of Data Management Technology and Enterprise Architecture at Bank of America, both recognized as honorees in the CDO Magazine Global Data Power Women series. Building on their initial conversation with Amy Horowitz, Group Vice President at Informatica, this installment delves further into how data leadership is evolving, highlighting a shift from compliance-based governance to agile, mission-driven strategies. Dr. Karkera and Michelle Boston share fresh insights on integrating AI, addressing important aspects such as explainability, ethical considerations, and mitigating bias, and discuss the ongoing challenges of implementing AI initiatives, balancing innovation with robust governance, and engaging stakeholders across organizations. Their perspectives continue to offer essential guidance for navigating the complexities of modern data and AI leadership.

Adita Karkera recently interviewed Chris Kraft to discuss his experiences in his role as the Deputy Chief Technology Officer (CTO) for AI/Emerging Technology at DHS. Chris started his career in the private sector in consulting, helping organizations implement and modernize large-scale systems. He then transitioned to Vice President of Corporate Solutions at SiriusXM, where he worked during the merger of Sirius Satellite Radio and XM Satellite Radio. During this period, Chris also attended law school, which ultimately led him to the federal space. As a public servant, Chris worked to modernize systems for the Executive Office for United States Attorneys (EOUSA) and subsequently the Federal Emergency Management Agency (FEMA). Upon joining the federal government, Chris was struck by the opportunities that technology provided to make the way individuals work more efficient and to free up time for people to do higher-value work. For example, at FEMA, Chris developed a Generative AI solution called FEMA OCFO GPT to help address budget questions that can take thousands of hours per year to answer and pull resources away from other mission-critical areas. The solution was built upon preloaded budget information and generated draft responses to budget questions, which the organization could validate and iteratively build upon outside the system, saving time and resources. In March 2024, Chris assumed his current role as the Deputy CTO for AI/Emerging Technology at DHS, where he pioneers the usage of artificial intelligence across DHS.

Differences between the public and private sectors

On the topic of the public sector versus the private sector, Chris emphasizes that the private sector offers more adaptability, but the decisions there do not nearly have as wide an impact as they do in the public sector: “At an organization like DHS or FEMA, when you make decisions on systems and approaches to implementing technology, they can have a significant wide-scale impact to the general public. I went back to the private sector for a little while, but the reason I stay in the public sector is being able to be part of the mission and being able to help people and do things that I think are really important for government to be doing, [like leveraging] technology responsibly. So, it’s really the mission and being able to help support the mission that keeps me very excited about the work that I do and the work that our teams do.”

How the CTO organization fits into the broader operations model at DHS

Expanding on how his role fits into the operational structure of DHS, Chris notes that the CTO organization works directly with the Chief AI Officer to lead AI efforts across DHS. His organization consists of three divisions: the AI Corps, the AI Governance and Engineering Division, and the Emerging Technology Division. The AI Corps was announced in February 2024, and Chris highlights how the government quickly hired and onboarded 25 AI experts who support cross-functional teams to take DHS’s use of AI to the next level. The AI Governance and Engineering Division helps with department-wide governance initiatives including use case inventories, safety and rights requirements, and complying with guidance under Office of Management and Budget (OMB) Memorandum M-24-10, which pertains to the use of AI within the federal government. The final division is the Emerging Technology Division, which focuses on the future of artificial intelligence and technology: “The idea there is that we need to continue to look forward to what’s coming next, and maybe it’s quantum or some other capability, or maybe it’s just another different application of AI; but having that group, really helps us continue to stay at the forefront of technologies.”

Generative AI initiatives at DHS

Discussing upcoming Generative AI initiatives at DHS, Chris highlights three key pilot use cases, which the DHS Artificial Intelligence Task Force (AITF) selected from a pool of around 30 submissions. The first pilot is with US Citizenship and Immigration Services (USCIS), which leverages a Generative AI chatbot to train officers to engage with refugees and asylum seekers using prompt engineering. The tool allows USCIS officers to ask questions and get human-like responses, which enhances their training as they become more comfortable with the types of answers they may receive in the field. The current training is provided by actors, which is limited by cost and requirements. The chatbot helps serve as a complement to existing training by providing more effective on-demand assistance to the officers without the cost of actors. The second pilot is with Homeland Security Investigations (HSI) and involves loading millions of investigations records into a vector database that generates responses to users’ questions. Chris states: “I sat in the room when the agents were testing this, and it was just so encouraging to hear the agents say, ‘It’s like the needle in the haystack. I’m finding things that I didn’t even know existed.’ Agents are searching for monikers for dark web cases and they’re identifying some that they didn’t even know were connected. I think that’s really the power of natural language processing and embedding information: using artificial intelligence to find similarities, surfacing the similarities, and then using Generative AI to summarize that information.” The third pilot initiative is housed under FEMA, and its purpose is to create hazard mitigation plans to support underserved communities. Chris explains that extensive and specific plans are required to qualify for funding, but some smaller communities don’t have all the necessary resources to create those plans: “The idea is to use Generative AI to capture information and then generate the plan, so you reduce the burden and make it easier for those underserved communities to get the resources that they really need.”

Challenges DHS faces in its AI deployment

Speaking on the challenges in deploying AI at DHS, Chris sheds light on the obstacles surrounding talent acquisition and bringing in the right people to implement and leverage AI technology, including contractors as well as federal employees. He notes that the DHS AI Corps (what will be the largest dedicated AI team in the federal government) is extremely helpful since it utilizes a vast breadth of knowledge to ensure things flow smoothly with internal talent and external partnerships. Another challenge Chris has seen is workforce readiness in the face of Generative AI and making sure his colleagues within DHS have the tools they need to understand and utilize this technology. Chris notes that training sessions are being held across DHS to address this problem. He emphasizes how DHS has already approved the use of commercially available tools like ChatGPT, Copilot, and Grammarly, but the challenge is making it clear to others how and why they should use these tools: “I think for those of us that are in technical positions, it’s easy to lose sight that most of our organization is just not as familiar or comfortable with the technology. So we’ve been holding a number of sessions where—and for a lot of them, I’ve actually been running these sessions—we’ll walk through these use cases of how you can leverage the tools: let’s walk through example one, here’s the data, here’s the prompt, let’s put it into the tool, let’s see what comes out.” These sessions build the workforce’s understanding of the full cornucopia of AI tools currently available to them in hopes that they will use these tools in their day-to-day tasks to make the way they work more efficient.

What the governance landscape looks like at DHS

Discussing the governance landscape for emerging technologies, Chris describes the protocols in place at DHS, including policies on the use of AI, facial recognition, and commercially approved Generative AI tools. He underscores that DHS is required to have an AI Governance Board to comply with OMB Memorandum M-24-10. Chris expands on the purpose of the AI Governance Board: “That’s a critical governing body that is actively engaged in how we’re leveraging artificial intelligence and how we’re using the capability responsibly.” He also notes that the DHS Office for Civil Rights and Civil Liberties (CRCL) leads DHS AITF’s Responsible Use Group (RUG) to actively govern DHS’s AI pilots (mentioned in part 2). The RUG and CRCL help to ensure that DHS uses AI technology responsibly, remains transparent in its use of AI, and follows external guidance such as the requirements outlined in M-24-10; for example, maintaining an AI use case inventory.

Advice for those looking to grow in the data and AI field

To wrap up the conversation, Chris shares his perspective on the importance of organizations protecting themselves as they leverage AI to achieve their missions. He states that government organizations should be well-educated on the technology they use, understand how technology constantly evolves, and have adequate resources to support these efforts. Chris acknowledges the importance of the government being at the leading edge of AI, emphasizing that public organizations need to work together in this endeavor: “I also think in the public sector—primarily at the federal level, but I also think it applies to state and local—we need to work together and we need to share information, because there’s no point in every federal agency or department figuring this out on their own and having to just blindly try to navigate. I know there are a lot of efforts underway to collaborate through the Chief AI Officer Council and other activities across the federal space, but I think it’s important that we work together, share information, help each other out, and continue to leverage the technology responsibly.”

Adita Karkera recently spoke with Fredy Diaz to discuss his role as the Deputy CDO of the US Department of Agriculture (USDA). Fredy has an extensive background in data governance and analytics and mentions the nontraditional path to his current role, which started in the US Marine Corps. That's where he says he 'fell in love' with data analytics while working in aviation and logistics and then proceeded to move through other government agencies including the Department of Defense, Department of Labor, and USPS Office of the Inspector General before he arrived at the USDA.

Where the role of Deputy CDO fits in operationally within the USDA

When asked about his current role within the USDA, Fredy reports to the Office of the Chief Information Officer (CIO), where his team has been granted technology capabilities and benefits that they're able to leverage. The Deputy CDO describes his role as 'very fun and exciting,' where he sits on the USDA CDO Council with the CDO and eight other mission-area assistant CDOs. He emphasizes his role as an adviser to CDO and Chief AI Officer Chris Alvarez. Fredy says his day-to-day is building relationships and trust across the organization: 'We value trust with our different organizations, and that’s how we’re able to leverage different programs.'

Artificial intelligence within the USDA and the Department’s core data strategy

On the topic of artificial intelligence AI, Fredy notes that the CDO is also functioning as the Chief AI Officer, and USDA is still having conversations about where the AI governance functions live. He says the USDA is ensuring it has 'AI-ready data and processes in place for AI that could have some alignment with our IT functions but may also have a lot more resonance with what we're doing currently in the data space.'

The Deputy CDO then discusses how the USDA is currently working on its second data strategy, but its five main strategic goals are focused on:

1. Strengthening data governance and leadership
2. Recruiting and upskilling the workforce
3. Maturing the enterprise data platform
4. Data sharing within the USDA, external agencies, and the public
5. Improving analytics techniques and ensuring the Department bakes in ethics from a data analytics standpoint

Fredy concludes the conversation by discussing the challenges the USDA faces, which includes understanding the existing strategy, distinguishing where it executed well, and finding communication gaps. The CDO states, 'We spent a lot of time from listening sessions, surveys, draft revisions—getting a pretty good representative group of folks across the Department to help us write the strategy. What we wanted to ensure was that it wasn't a shock when it went through the clearance process or when it finally came out.' When the strategy came out and a diversity of stakeholders were proud of it, it showed the Deputy CDO, the CDO, and others that the work was a success and not just a 'paper exercise.'

Focus on AI strategy initiative

Fredy Diaz continues the conversation with a discussion about an artificial intelligence (AI) initiative he is excited about—the USDA is currently working on developing its AI strategy in accordance with the Office of Management and Budget. The Deputy CDO reveals how the credibility his organization has built from developing and starting to execute its data strategy in a more holistic view across multiple stakeholders is ‘paying dividends,’ and now it’s just a matter of doing that exercise again. Fredy says this has been helpful for crafting a vision for AI in the USDA as well as aiding in the Department’s ability to execute on newer technologies like Generative AI.

An AI use case within the USDA and the Department’s open data work

When asked to provide a use case of AI within the USDA, Fredy elaborates on two items he’s most proud of. The first is that the Department has an AI Center for Excellence. Fredy expresses, ‘I love the fact that this started as an initial group and has scaled up, and now we have a bunch of practitioners that we can call on for various projects, big hairy problems that come up with AI, and we’re not just solving it at the Department or in small pockets—we can call in this group where needed.’

The second is the USDA’s work with its chief technology officer and working to spin up a technology quickly, evaluate it, ‘kick the tires,’ and see if it’s good for the agency’s purposes. Fredy notes, ‘That partnership has really paid off, and we’re expecting that when it fully matures it will reduce our time to evaluate technologies and software.’ The conversation then shifts to the USDA’s open data strategy, which includes public data, such as the Census of Agriculture, as well as its own open data platform in which the public can see how dollars are being spent in different areas. The CDO also mentions that the USDA is publishing data in university classrooms nationwide, which not only gives students the opportunity to view real-world data and problems, but also may pique students’ interest in working in public service by showing them what organizations like the USDA have to offer.

Successful data literacy programs within the USDA

On the topic of data literacy programs within the USDA, Fredy shares three programs that he thinks have done exceptionally well: the USDA’s data science training program, which has grown from two dozen participants to nearly 200 participants within the organization, focused on upskilling and training members through a capstone project as well as networking; increasing the amount of rotational opportunities that employees are given; and rapid prototyping and hackathon events so individuals can apply learned skills and have even more chances to participate.

Advice for aspiring data professionals

To conclude the conversation, for those looking for a career in data and analytics in the future Fredy provides some words of advice that include embracing data challenges. He says, “Don’t be afraid to volunteer. I’ve had a lot of folks that have expressed interest in rotating to the Office of CDO somewhere in the organization, and it just started with the question of how they could be of help. It starts with making your intentions known and just letting someone know—either a peer or even just a cold email to someone—and say, ‘I would love to explore an opportunity.’”

Adita Karkera recently sat down with Robert (Rob)  King to discuss his role as the CDO of the US Department of Energy (DOE). A systems engineer by trade, Rob has a wide-ranging background in IT, which includes working in the private sector before joining the Department of Homeland Security (DHS) in 2011. From 2012 to 2020, he 'took the reins on driving a data and analytic strategy' for the DHS, during which time he became its CDO for mission support in 2016. Rob then became the inaugural CDO for the Social Security Administration (SSA) in 2020 where he stood up a data governance board; introduced data literacy, governance, and analytics; and drove consistency across the SSA’s 1,500 field sites. By 2023, both of those leadership experiences led Rob to his current role as the DOE’s CDO where he and his team are building out an enterprise data management program.

Lessons learned and how the CDO fits within the DOE

When asked about what he’s learned through his years as the DOE’s CDO, Rob highlights that the role is still new (only mandated five years ago by the Evidence Act), and it’s 'one that needs to be able to orchestrate across many other functions of the organization.' The CDO discusses the importance of corporate psychology; even though he and his peers are mostly data and systems people, they still need to 'win the hearts and minds' of others in the organization and build coalitions to enhance their mission through data and advanced analytics products. Regarding how the role of CDO fits within the DOE, Rob states the position falls under the chief information officer  (CIO), but it was initially under the chief financial officer (CFO) in FY2022 before he joined in FY2023. Under the CFO, it was a collateral duty role, but the DOE shifted it to a full-time role under the CIO, allowing the CDO to manage an enterprise data management program without other distractions.

How does AI play into the CDO’s role, and what is the current data strategy at the DOE?

On the topic of artificial intelligence (AI), Rob shares that the DOE appointed a new chief AI officer and formed an AI Executive Order working group that is tri-chaired by the chief AI officer (Director of the new Critical and Emerging Technology Office), a responsible AI officer (Deputy CIO), and the CDO (Rob). Rob explains, 'We need to be working in that coordinated, collaborative manner to develop a cohesive strategy … and recognizing that we cannot separate data and its critical role to AI.' The conversation then pivots to the DOE’s data strategy, which the CDO prefaces with the fact that more than 50 national lab security sites and program areas comprise the DOE, many of which are government-owned and contract operated—a culture that poses a unique challenge to the data strategy work. 'When I work with our legal and oversight bodies,' Rob notes, 'it’s really trying to find the blend of how we balance the federal, legislative, and statutory drivers against the fact that there aren’t necessarily government organizations to which our programs are being operated, so we have to look at this through a multidimensional framework.' With that in mind, and referencing other federal data strategies, the DOE has come up with five strategic goals that have commonality with other strategies. These five goals include establishing data governance and stewardship, building data architecture and infrastructure, focusing on data quality and discoverability, determining what data literacy and a data literate workforce look like, and ensuring responsible and sustainable use of data for artificial intelligence.

Adita Karkera recently sat down with US Immigration and Customs Enforcement (ICE) CDO Carin Quiroga to discuss some of her day-to-day responsibilities as well as some of ICE’s current data initiatives and strategies. Carin spent 11 years at US Customs and Border Protection before joining ICE and becoming the CDO—a position she has held for the past two years.

The role of CDO within ICE and current accomplishments

Carin’s role is currently positioned under the chief information officer. She highlights how her role as CDO includes getting others excited about data. When Carin began her role as CDO, the goal was to become an 'action-oriented' office and tackle the issues and gaps within ICE’s data. Her team conducted various interviews with different stakeholders and leaders to pinpoint exactly what their issues were. Then the team created the IT data modernization roadmap whose framework is 'collect, automate, and use.' Carin states, 'We needed to ensure we were collecting the right data at the right time, stop collecting data that we didn’t need or didn’t use, and then really drive automation and use that data to increase transparency and make decisions.' Based on this roadmap, ICE’s data office was able to create an accountable, measurable data strategy.

An ICE data initiative that excites the CDO

One data exchange and sharing initiative that Carin is very excited about is the 'data bodega,' which she explains is 'a centralized location where everybody can go to know how to get the data and where to get the data from.'

ICE’s Data Empowerment Program

Continuing the conversation, Carin Quiroga discussed creating the Data Empowerment Program after she became ICE’s CDO. The program focuses on data literacy because “everybody has a role in the data life cycle.” She wants to ensure everyone has an opportunity to gain experience and feel “empowered” in their data journey.

Passions that currently motivate the CDO every day

When asked about what drives her every day, Carin reinforced the idea that “data is knowledge, and knowledge is power.” The CDO says she’s very excited about bringing others together and using data to make better decisions and become more efficient.

How ICE addresses data governance and quality

Carin described data governance in terms of “changing the culture to understand you’ve got to let go of your data. You’re going to share it, and someone else is going to use it.” By understanding this, individuals within the organization are more likely to implement data standards and ensure the quality of their data is accurate. The CDO also emphasized the importance of collaboration between the different agencies, particularly the Department of Homeland Security (DHS), which includes monthly CDO councils between the different agencies to discuss their challenges and the work they’re doing.

Advice for aspiring data professionals

Carin closed the interview by talking about her enthusiasm with working with women in data and creating diversity in a male-dominated technology industry. The CDO said, “For me, it’s just so important to have everybody’s point of view, everybody’s history, everybody’s perceptions while we’re going through this journey.” The CDO’s advice is to never give up and to always be your authentic self.

Adita Karkera recently spoke with Ken Pfeil to discuss his day-to-day activities as the CDO for the Commonwealth of Virginia. Ken has an extensive background in the private sector after his military service, which includes working for Dell and being a technical architect for Merrill Lynch, as well as working more than 20 years in the information security sector before taking on his current role.

Where Ken Pfeil’s office is located and the work he drives with other agencies

Ken's office is codified in Virginia law, and it officially became an agency in July 2021. The CDO reports directly to Secretary of Administration Margaret 'Lyn' McDermid, and his office works with more than 100 agencies to provide data services. Ken states his office is mandated by Virginia law to provide the data governance standards across the Commonwealth and does so through a three-tiered hierarchy, which consists of the Virginia Data Advisory Commission, an executive data board, and a data governance council. Underneath this lies an informal data stewardship group, which Ken explains is 'the meat and potatoes part, where everyone gets to learn their collective responsibilities when handling data, using data, and owning data.' The CDO also highlights a Data Maturity Management Assessment (DMMA) conducted by his office to other agencies, which surveys their culture, support, and how they’re handling data. The survey, which achieved a roughly 84% response rate, assisted Ken’s office in finding agencies that were more mature in their data governance and used some of their practices to help other agencies that are not yet operating at that level.

How Ken’s team is breaking down and prioritizing key issues throughout Virginia

When discussing how the CDO's office is addressing significant issues in Virginia, Ken states that certain objectives must be met within a time frame, such as engaging in Operation Bold Blue Line, an objective to lower crime within 13 localities in Virginia. Ken affirms, 'We were instrumental in facilitating all of the data collection transfer, dashboard analytics,' which helped the governor of Virginia make decisions regarding where a police surge may be needed. He also emphasizes how much support his office has had from the governor and first lady and Virginia cabinet members: 'Most of the time, if there is an issue, I can generally raise it with them, and they’ll go right back to that agency head and say you need to sit down … to figure this out.'

Differences between private and public sectors in terms of data challenges

Ken discusses some of the differences he’s noticed between dealing with data challenges in the public versus private sector. For instance, the data maturity level in the private sector is usually higher than in the public space, according to Ken; and other agencies are typically resistant to change, so he is wary of throwing too much information at others just because it works in the private sector. When discussing some of the challenges he faces, Ken states, “One thing that resonates everywhere is the ‘silo-ization’ of data.” Many agencies are very cautious of sharing their data with each other, and the CDO helps them to see the positive outcomes of sharing that information.

Further insight into data trust

Ken describes how data trust between agencies consists of two major parts: the Commonwealth Data Trust (CDT) and the Virginia Open Data Portal. The CDT is mainly for agency-to-agency sharing, or locality-to-locality sharing, and consists of a universal entity index of 8 million to 8.7 million Virginians. The index is stripped of personally identifiable information and allows the CDT to use the data successfully and securely on an anonymized basis. The CDT also consists of a common memorandum of understanding, for which Ken is the trustee. This document is 27 pages long. The CDO notes, “We decided early on that we were going to list the first 22 pages as what our obligations are to each agency”—for example, keeping data private and secure, classifications and roles, among other topics.

The Open Data Portal was set up for transparency per Virginia code and allows localities to federate their public data to another locality through Ken’s agency. The CDO also presents his agency’s data trust automation process, which helps facilitate data transfers between other offices. The process would usually take one to two days previously, but Ken’s team has cut that down to a matter of minutes using this automation initiative.

Ken’s efforts around data literacy

When discussing the topic of data literacy among agencies, Ken underscores that every member of the Commonwealth Data Trust is eligible to participate in data literacy and upskilling programs. The agency has partnered with Dataversity so that “any agency has the ability to take courses—from the executive level, where you know nothing about data, to the DMBoK [Data Management Body of Knowledge] certification side of the house.”

Upcoming projects Ken is excited about

Ken highlights one initiative he is especially excited about, which involves working with the Virginia Department of Veterans Services to identify every veteran in Virginia and “make sure they understand what their entitlements are, how to get access to said entitlements, and just facilitate taking care of veterans.” Another initiative that the CDO introduces is the Office of Workforce Development and Advancement where his agency provides oversight in the workforce development ecosystem. Ken’s agency has helped provide pathways from education to the workforce and is playing a major role in putting together a singular job portal to place citizens in the workforce, utilizing artificial intelligence (AI) to identify skills that align to existing jobs. The CDO emphasizes that “the governor had an initiative to bring 400,000 jobs into the state of Virginia when he was inaugurated, and I think we’ve hit over 230,000 already.”

Leveraging and governing the use of AI at the state level

On the topic of AI within his agency, the CDO explains how they are using predictive analytics for cyber security to supplement other localities not having sufficient staff to properly address their own security issues. In terms of governing AI, the CDO addresses four main focus areas to help establish guardrails and safety mechanisms: “One of them is around privacy and intellectual property, we also have policy standards for effective and safe usage around AI, we have IT-related safeguards … and lastly, we want to ensure all of our Virginia students are properly trained within those guardrails and they’re ready to compete for jobs in the workforce.”

Advice for future data professionals

To conclude the interview, Ken shares his advice for individuals who are looking to prepare for a role within the data and analytics field—it’s to not focus on a singular framework or system. “Immerse yourself in different areas of technology … get out there and understand how the whole system works together,” he suggests. The CDO also reveals that leaders must learn to work better with the newer generations and that trust always goes both ways.

Dr. Bryan Pendleton discusses his role as CDO in the DHS Office of Intelligence & Analysis (I&A), his journey to the position, and DHS efforts to make discoverable the vital intel used to aid in the protection of US citizens.

Dr. Pendleton partners ‘humans with machines’

Dr. Pendleton came to the DHS in 2011 as an intelligence officer after a career in the US Army and the Defense Intelligence Agency. Since then, he has had multiple roles within DHS I&A and, in 2016, was asked to lead a technical intelligence group leveraging computer and digital forensics that became foundational to his journey to CDO.

'We had a very large backlog of data that needed to be made sense of, but we had a very small staff, and all our processes were very manual in nature,' Dr. Pendleton said. He applied his background in Lean Six Sigma to bring in lean approaches and automate those processes using machine learning. These approaches worked to quickly derive value from the data and disseminate it to those who needed it. His efforts partnering humans with machines informed the basis for his appointment to CDO.

Connecting the dots for the mission

Dr. Pendleton said, 'In the aftermath of 9/11 and the failures of the US government broadly to connect the dots, it became really clear to me that my job was all about the dots.' He explained his primary responsibility is to get intelligence professionals the data they need to identify threat actors. He is constantly working to identify which 'dots' are relevant, negotiating appropriate access to that information, and providing the DHS workforce with advanced analytics to support them connecting the 'dots.'

For more on how Dr. Pendleton engages his stakeholders and how 'everything is about partnerships and relationships,' watch part 1 of the full interview above!

Key pillars of Dr. Bryan Pendleton’s data strategy

DHS’s data strategy is built upon four pillars, the first being maturing its data management process: 'We actually did have really good data management practices and procedures, but what we didn’t really have was a cohesive end-to-end process.' The second pillar is all about improving information-sharing within the DHS to 'help get ahead of some of these issues that resulted in the catastrophes of 9/11,' as Dr. Pendleton explains. The final two pillars include 'adopting emerging technologies to include advancement of our accelerated adoption of artificial intelligence' and improving data literacy and acumen of non-data professionals.

Dr. Pendleton’s AI responsibilities and priorities

Speaking on AI innovations, Dr. Pendleton wears multiple hats: To the intelligence community, he is the 'senior official appointed to accelerate I&A’s adoption of AI,' and he represents I&A and DHS on the Augmenting Intelligence Using Machines Executive Steering Group. For DHS, he is I&A’s representative to a consortium of senior officials responsible for AI within DHS components. They meet to ensure strategies and initiatives are 'in synchronization' with secretary  guidance and priorities and are led by the DHS headquarters.

Managing data governance across the DHS

'When it comes to data governance broadly, our team facilitates something we have—which I view to be a best practice across the US government—called the Data Access Review Council (DARC).' DARC’s core focus is the sharing of data agreements to ensure that any distribution of bulk departmental data comports with all the applicable privacy impact assessments and is appropriate and lawful. This works by 'being very use case case-driven. It’s an extraordinary, complex landscape to facilitate information- sharing agreements, but the DARC, in my mind, really accelerates those things by bringing all the stakeholders together from the very beginning.'

Dr. Pendleton’s key accomplishments as the DHS I&A CDO and future key initiatives

The CDO celebrated as “monumental” publishing, implementing, and measuring the success of his first data strategy. Dr. Pendleton also highlighted how getting the DHS’s data scientists to work alongside the intelligence team enabled mission work: “The more they work side by side, the more they start to speak each other’s language and understand what [the others] are capable of.” On data acumen and literacy and through a partnership with the intelligence training academy, they have launched a Data 101 module that is now part of the basic intelligence training course for all DHS analysts.

Among other initiatives, the CDO celebrated the kickoff of the DHS AI Taskforce: “I think that the intelligence community has a lot of capability when it comes to AI, and to the extent that some of those capabilities actually are unclassified in nature, I want to see how I can be the bridge to bring intelligence community capabilities to the department.”

Dr. Pendleton’s advice to aspiring data professionals

Dr. Pendleton provided some advice for aspiring data professionals: “Figure out what you’re passionate about, what drives you, what mission you want to support, and let that be the thing that guides you.” He emphasized the importance of soft skills like teamwork, relationships, and negotiation in helping individuals advance their careers. Dr. Pendleton explained that diversity, equity, and inclusion must work together in the workforce: “The problem with having a simply diverse workforce is if they’re not included in actually solving problems and bringing their very best to bear, we get suboptimal solutions to the toughest problems. … It’s not just about the diversity of our workforce; to me, it’s about actually including those people practically and bringing them to the table.”

The First CDO at FDA

Ram Iyer became the FDA’s first CDO during the COVID-19 pandemic after 25 years working across private industry, non-profits, consulting and international agencies.

At FDA, “our focus was very heavily on just getting things done in the first few months for the immediate needs”. Mr. Iyer believes that having the CDO one level removed from the commissioner has given his team the needed visibility and helped them make tremendous progress over the last two years.

Creating an Actionable Data Strategy

The FDA’s data strategy consists of three pillars: value-driven projects, building foundational capabilities, and a hyper-focused plan for talent. “With all these three we’ve been able to successfully create an ecosystem of tools, processes and data removing some friction that existed in the organization for user data.” Emphasizing the combination leadership support and these pillars, he celebrated resolving problems in the data-practices side and mission priorities.

A Few Success Stories as a CDO

Mr. Iyer highlighted creating FDA’s master data catalog for the products or ingredients they need to monitor and how they have used AI to pull data from artifacts to create digital assets. Mr. Iyer also discussed the office’s success in upskilling, laying out their “70-20-10” model: “70% of their time is spent working on real problems... 20% of the time is spent on mentoring….only 10% is spent on typical traditional classroom training so we have found this to be particularly effective”. Using this strategy, the FDA was able to solve essential problems while saving time and significant funds and produce a cohort of 36 trained data scientists with a common understanding of how to solve problems for the agency.

On the Value of Publishing a Data Strategy

Publishing the data strategy “allows us to test our implementation against the north star that we establish”. Mr. Iyer notes the critical step to collect input from key stakeholders and individual perspectives when developing their strategy: “when we built our strategy, we did not do it in isolation... it’s important that the process is collaborative and inclusive”. The strategy must be able to adapt to new things like AI, but not be so high level as to be ineffective for implementation.

FDA Initiatives Implementing AI

FDA looks at AI at multiple levels: operations, productivity, and policy. Mr. Iyer says operations provide a learning and testing ground to be more aggressive on AI. On productivity, the FDA is considering capabilities for office suites or other operational areas. He identifies policy and regulation as the larger and time-consuming issue. FDA both supports centers on policy and is working with multiple agencies and public private partnerships. “We have formed an AI governance and guidance board that is meeting to help identify the right use cases and also implement them and through those learn what is the right use of AI within the agency”. They use a guiding principle of “faster is not always better and easier is not always the best.”

Career Lessons Learned

Mr. Iyer expressed appreciation for how quickly and effectively the FDA responds to challenges and the remarkable talent that exists at FDA: “because I started in the virtual environment, I did not have the interactions that I would have liked to have with so many of the great talented folks across the agency and if I had known all the great talent that was available, I think I would have built my network faster to solve some of the bigger problems”.

Adita Karkera recently sat down with Dr. Alan Sim to discuss his role at the CDC and their innovative project initiatives as well as his insights into public health data. With a background in epidemiology and public health, Dr. Sim’s career has spanned 25 years at the CDC “looking at the application of informatics IT in public health.”

Dr. Alan Sim’s role within the CDC

Dr. Sim serves the CDC’s new Office of Public Health Surveillance, Data, & Technology (OPHDST).  Expanding on his role, the CDC’s CDO calls himself a “data diplomat” to his peers: “the term data diplomat best suggested that I would take a role where I would represent them…highlighting often the work that’s being done that unfortunately sometimes gets buried under layers of bureaucracy”. To accomplish this Dr. Sim created the Enterprise Data Office based on three priority areas: Enterprise Analytics and Data Science, Strategy and Enterprise Data Management, and Data Governance and Policy.

How Dr. Sim is operationalizing data governance for the CDC

Dr. Sim noted that: “To try to implement data governance is extremely challenging and you generally don’t have consensus, and the progress you make is very minimal “. The CDO noted his approach is about breaking things into domains of work by communities that are particularly invested in the data. Dr. Sims suggests that “when you have those communities that are really invested in the data products that are being generated there, you have a higher chance of finding commonality and consensus”. To achieve this unanimity, the CDC stood up a data standardization management and harmonization working group along with an IT and Data Governance initiative focused on investments and data architecture implementation.

Initiatives at the CDC

Dr. Sim highlighted CDC’s where CDC is making public health systems more responsive to public health disasters. Dr. Sim discussed strengthening the public health data by creating a front door for bringing data to the CDC. To accomplish this, the agency is looking to standardize the funnel of where data comes in and use “best of breed” processes to ensure the data then gets disseminated across CDC. Dr. Sim is also excited about partnering with program data scientists who leverage computer vision technology to identify satellite images and locations of cooling towers during a Legionnaires’ disease outbreak, as well as developing machine learning and NLP techniques to mine social media posts to identify trends in suicide incidents and aide in suicide prevention.

Improving internal data literacy at the CDC

Dr. Sim discussed how the CDC HR office “developed a career ready program that creates a career community where we can identify what are those necessary skills, necessary milestones to help existing staff move up the career ladder”. The CDC has an online data academy where they provide self-guided training on data science languages. Dr. Sim spoke about helping individuals learn about the diversity of public health data and the resources they would need: “if I myself was a junior data scientist or someone who is interested in improving data literacy, what are those resources that would be available to help me to do those kinds of things and advance my career”.

Advice for aspiring data leaders

To conclude the discussion, Dr. Sim provided advice he would share with aspiring data leaders. “Be bold and innovative, attack problems creatively… leverage the data, the tools, the services that are available and while you’re attacking the problem connect with your peers”.