Skip to main content
Welcome to Deloitte
If we have selected the wrong experience for you, please change it above.

From concept to care: accelerating health care innovation with AI

Introduction

When new ideas have the potential to improve patient care, turning them into working solutions quickly and safely matters. For Health New Zealand | Te Whatu Ora, the challenge was not a lack of ideas but finding ways to accelerate innovation while maintaining the high standards required in health care. Together, this government agency and Deloitte New Zealand helped create an environment where clinicians, engineers, and digital specialists could rapidly explore how AI could support health care delivery, transforming concepts into working prototypes in just two days.

It is fantastic to see solutions I’ve been dreaming of for 20 years come to fruition in just a couple of days.

─Bryan Simpson, Clinical Informatics Director, Health New Zealand

The situation

Health New Zealand | Te Whatu Ora is building a future-focused public health care system that supports all Kiwis (New Zealanders) to live better and longer lives. As health care becomes increasingly digital, clinicians and technology teams are exploring how AI can help improve patient care and speed up software development.

However, innovation in health care comes with unique challenges. Digital solutions must be safe, reliable, and fit for purpose because people’s well-being is at stake. These high standards often lead to long and resource-intensive development cycles for digital teams, slowing down engineers, clinicians, and, ultimately, the delivery of benefits to the public.

Recognizing both the opportunity and the challenge, Health NZ sought a way to explore emerging AI capabilities in a controlled environment while ensuring security, privacy, and responsible development remaining front and center.

The solution

To encourage innovation and safely accelerate the prototyping of ideas, Deloitte and Amazon Web Services (AWS) hosted the Code for Care hackathon in Wellington, New Zealand. The two-day event brought together software engineers, developers, testers, and clinical specialists from across the organizations to solve real-world health care challenges in a structured setting.

Teams followed a clear three-phase workflow to move ideas to software prototype:

  1. Requirements: start with real clinical use cases and strict problem framing
  2. Design: craft a clear and structured approach that defines behavior before any coding takes place
  3. Implementation: using AI, translate specifications into working prototype components efficiently

To help attendees bring their ideas to life efficiently, AWS introduced proven specification-driven workflows and AI-assisted development practices, leveraging Kiro, AWS’s agentic coding environment. Kiro uses autonomous AI systems to handle complex coding tasks, and is built on Amazon Bedrock, a platform for building secure AI applications. Using Kiro, attendees could easily translate requirements in natural language into design, generate tasks, generate code, run tests automatically, and use steering files (files to inform Kiro standards that apply to a workspace) to maintain consistency. To protect security and privacy, all data was mocked or sandboxed, meaning software testing took place in isolated, controlled environments.

The impact

By following the structured three-phase workflow and using AWS Kiro, teams created three working prototypes over two days. Each solution addressed real pain points for clinicians and provided a solid foundation for future scaling:

  • Medication checker: brought mocked medication and lab information into one place, helping clinicians quickly spot any risks before prescribing
  • Waiting-room triage: used a digital form to collect key patient information before triage, helping clinicians prioritize people more easily
  • ED notes encoder: applied AI to analyze emergency department notes and match them to standardized clinical codes, reducing the need for manual coding

The speed of development was particularly significant. Work that would traditionally take weeks was completed in just two days, demonstrating the positive impact this technology can have on development cycles for digital teams and, ultimately, on frontline health care.

Did you find this useful?

Thanks for your feedback