Skip to main content

Generative AI as a design partner for process improvement

User Case Study

Rick Hanley, Senior Grants Officer, Nuffield Department of Medicine

Before using GenAI tools, my process for developing internal improvements was limited by time and the steep learning curve of software design. I often had ideas for streamlining grant admin workflows or filing processes, but progressing them alongside day-to-day responsibilities was slow and required a lot of trial and error.

GenAI has changed that. It has allowed me to prototype, refine and stress-test ideas far more easily, particularly when developing tools like JournalWings (Excel/Oracle journal automation) and FileAssist (a configurable document-naming and filing tool). The work itself remains substantial, but progress is steadier. Tasks that would previously have stalled – usually because of architectural uncertainty or getting stuck on a technical detail – now move forward more consistently. AI has not made development instant, but it has removed many of the dead stops that used to slow things down.

I primarily use ChatGPT as a design partner. I provide it with Python, JavaScript and front-end code, draft UI behaviour, or outline a workflow problem, and it helps me shape solutions. I often iterate by discussing small, focused improvements or by validating architectural decisions before I commit. I also use it to explain unfamiliar concepts and to check my understanding.

Using GenAI has reduced development time dramatically; I would estimate by 70–80%. My workload has shifted: instead of wrestling with low-level debugging for days, I can focus on designing better workflows and creating tools that are tailor-made to solve targeted pain-points in my department’s processes. It has effectively enabled me to become a part-time developer within a professional services role, creating innovations that previously would have required dedicated IT resource.

My advice is to treat AI like a collaborator: question it, challenge it, guide it and make it explain things until you understand its decisions. The real power for me has been to treat it as a very enthusiastic teacher that needs frequent reminders of context and direction. You still need to understand what you are trying to achieve. AI can help you assess the steps to get there, but you are driving the process.