Designing for Citizens needs


This project will deliver a next generation chatbot, built around Large Language Model (LLM) technology, to deliver new advice service gateways but built to be accessible based upon all user needs. We aim to:

1\. Deliver a smooth, multi-channel experience for citizens accessing advice services

2\. Use co-design with users to fully consider the requirements of marginalised and vulnerable citizens

3\. Harness the potential of new AI technologies to deliver personalised advice at scale.

Whilst face-to-face advice is a unique selling point for Citizens Advice Network (CAN), we want to offer multi-channel, 24/7 access to our services. This project will allow us to prototype and implement a chatbot/voice assistant and other approaches to augmented self-service. Advances in LLMs (as witnessed by the stunning success of ChatGPT) means there are great opportunities to deliver improved services. We don't intend to replace our volunteer advisers with technology: human support is vital, especially for our more vulnerable clients with complex needs. However, if technology can allow some proportion of clients to self-serve, this frees up capacity for those who really need help and also helps avoid adviser burnout. During the pandemic, CAN saw that customers want to access advice through a variety of channels to suit their needs, time of availability and other factors. Given the scale of our network and our limited funding we have been harnessing 'no-code' technologies to improve service design. This has proven an effective way to improve our service, at a fraction of the usual cost.

The another key innovation of this project is to use the latest thinking in person-centred design, service co-design, iterative methods for service design development and frameworks for continuous improvement of service capacity. By co-designing and co-developing a bespoke demand management tool for CAN members, which remains rigid enough for collective agreement, but flexible enough to allow each CAB to customise and interpret flexibly, locally and generate specific service projects. Initially based on use of no-code technology but with the potential to be expanded to other non-tech areas, upskilling the CAN network and increasing its design capacity on how better to meet the needs of service users will enable CAN to explore creative ways to engage service users of all ages and backgrounds to be involved to help us to develop, shape and evaluate our current services for the benefit of both CAN members and our service users throughout Scotland.

Lead Participant

Project Cost

Grant Offer



HELPFIRST LIMITED £49,722 £ 49,722


10 25 50