Everybody is talking about this new tech. But how do we use it?
The situation
Dutch company PostNL has been in business for over 225 years, which reflects their ability to transform, adapt to the ever-changing needs and preferences of their customers and take courageous steps to innovate. The company strives to be the 'favourite deliverer in, from and to the Benelux' by delivering a distinctive customer and consumer experience. Bram Tilmanns, Program Director Generative AI at PostNL: "Generative AI is a promising technology and we felt that GenAI could help us realise our ambitions in growth and innovation. This technology will help us optimise processes, make our colleagues' work more efficient and enjoyable and improve our services towards customers. We looked for a partner who could help us structure and organise the knowledge and insights we needed." PostNL approached their long-time partner Deloitte to develop a comprehensive plan to dive into the opportunities and possibilities of using the new technology responsibly.
Morris Boermann, Deloitte Senior Manager: "Our approach focused on delivering value fast and it comprised of three streams. Value Delivery: we wanted to create buy-in by showcasing potential value of GenAI through the discovery, prioritization and development of a number of GenAI POCs. Accelerate Organisational Change: we developed a vision of the organisational structure, governance and change needed to accelerate GenAI at PostNL. And finally, we used a stream called Guide Future Technology: providing advice on technology partnerships, including with AWS, who are PostNL's cloud provider, to accelerate, orchestrate and integrate GenAI into their technical landscape."
PostNL already had its ethical framework to work with AI — introducing GenAI required a renewed take on this. Naser Bakhshi, Deloitte Partner PostNL: "This new technology needed guardrails to make sure PostNL employees could use it in a reliable, trustworthy and transparent manner. We drew up risk checklists and we enriched existing lists. For instance, consumers need to be aware when they interact with an AI system. In itself this is a simple check, but it needs to be embedded in the risk frameworks to ensure that the company complies with regulations."
Deloitte & PostNL: Discovering and Scaling GenAI together
Prototyping, proofs of concept, simulations games: just like the real thing
The solve
Deloitte hosted one of its GenAI labs for the PostNL leadership. This is a lab that consists of a ‘birth place’ and site for customer events and collaboration, offering a whole tool kit of collaborative hackathons, labs and design sprints. The objective: helping our clients identify GenAI possibilities and co-create tangible concepts to integrate the technology. This approach triggered interest in assessing the potential of GenAI for enabling the company’s goals. Through fast prototyping it showcased the discovery, prioritization and development of multiple proofs of concept (POC). Answering questions like: how can we leverage GenAI as a competitive advantage in delivering our ambitions? Are there new ways to make operations more efficient? And how can we reduce costs?
These POCs resulted in 60 to 80 percent of resolution within six weeks. They were produced together with the PostNL developers, within the company, and then scaled up and delivered by PostNL themselves. Simulation games were organised to determine if a proof of concept would work in real life. Based on a specific use case, stakeholders were identified and invited to join. Stakeholders would interact with the simulated processes as if these were already developed and deployed. In this way we could test if the risk elements were addressed properly and if there were any gaps left to cover.
Bakhshi: "A lesson learned: data quality is paramount in designing reliable GenAI solutions. You want to avoid the classic 'garbage in, garbage out' pitfall. Also, we see a direct correlation between the quality of the data and the level of GenAI adoption by colleagues." Another important factor in creating a successful use case is paying attention to the human factor. Listening to people and asking: what do you do daily? Which elements of your work are repetitive? Which parts would you like to be different? Tilmanns: "A good solution is a human solution. GenAI should not be about replacing people by technology, it should be about people benefiting from technology. Then the sales managers or the service teams can add value by focusing on things that demand human attention."
It is of utmost importance to implement AI solutions that are reliable, trustworthy, and transparent
At Deloitte we use the three lenses approach to evaluate initiatives like these POCs. We ask ourselves: is it desirable? Is it feasible? Is it viable? Boermann: "This is an easy way to prioritise what is right and what you need to do tomorrow. When thinking of a GenAI solution, start with the people who should use it. Always keep them in mind, whether they are your employees or your customers, and take it from there."
More tips for companies that consider starting with GenAI:
- make sure you have executive sponsorship
- start small, fail fast
- talk to teams and colleagues to find out which parts of their work are repetitive or boring
- develop your solutions around these issues
- learn, improve and then scale your successes
The impact
As a result of this approach PostNL now shows a clear executive sponsorship for GenAI as a key topic for the organisation. A taskforce was set up to accelerate the value of GenAI in the short term, which will be evaluated after twelve months. The taskforce is instructed to create continued value through delivery of additional POCs, as well as establishing the expertise and ways of working with GenAI. The lab has opened a new perspective to future-proof PostNL.
A promising journey has started
