This project is intended to determine a proof of concept foundation for the use of an AI Large Language Model tool (‘RegNav’) to minimise the density of regulatory review and response by reducing the time required to metabolise and disseminate new regulatory proposals. The current process is entirely manual and this project will help us understand if AI LLM (Large Language Model) tools can be used to improve the associated times and costs through the introduction of automation usage and generation of long format responses, sometimes referred to as reflection.
This tool will assess published regulatory changes in the form of consultations, calls for input, calls for evidence, and create prospective responses, and incorporate a feedback loop to improve responses over time. The tool is aimed to help minimise the density of regulatory early-stage response by reducing the time required to consume new regulation and quickly understand its impact.
Benefits
This project will help improve the efficiency and effectiveness of regulatory processes, ensuring that NESO can respond to regulatory changes with faster, more precise, more targeted and scalable outcomes. The tool this project will provide test will help to develop a repository of knowledge and ‘information memory’ from business areas across NESO. In addition, the tool will be able to generate the information concisely and provide efficient summaries of complex consultation documents.
With digitalisation being pivotal, we will be working more closely towards NESO’s strategic priority to develop a digital mindset to unlock the potential of technology and teamwork. This also fits in with NESO’s four key values: to accelerate progress, be curious, build trust and create belonging.
Learnings
Outcomes
Delivered the proof-of-concept prototype.
Delivered recommendations for progression to a minimum viable product.
Positive engagement with key stakeholders and with wider business.
Lessons Learnt
1. Importance of business engagement
Active engagement from the business was critical to the success of the project. Establishing a clear vision for how stakeholders will interact with and benefit from the prototype was essential. This was achieved through developing strong working practices and providing clear focus and direction throughout the project period. Early and consistent involvement ensures alignment with organisational priorities and maximises adoption potential.
2. Value of iterative development
The use of iterative cycles – build, review, refine – was highly effective. This allowed for rapid identification and resolution of issues and continuous improvement based on stakeholder feedback. This resulted in delivery of a solution that better meets needs and expectations.
3. Wider business engagement
While engagement with the core team was strong, broader involvement across wider areas would have been even more beneficial. Demonstrating the use case to a wider audience earlier and more frequently could have increased awareness and understanding of the project’s potential, generated additional insights to strengthen the prototype and supported early buy-in for any future scaling and implementation.
4. Seizing emerging opportunities
A key lesson was the importance of recognising and acting on opportunities that were raised during the project that was not originally planned for. For example, one of the artifacts of the prototype originally intended to operate in the background had the potential to significantly expand the use case, demonstrating the value of flexibility and adaptive thinking during the project.
