Can chatbots and AI help solve service design problems?

Contents:

  1. Project outputs
  2. Project timeline
  3. Feedback

This discovery has explored the feasibility of adopting or developing a common, open source platform for local government to use to develop chatbot/AI solutions.

Local authorities face a variety of issues when looking to adopt chatbot and AI solutions. As there is no shared understanding of the pros and cons of the tech, the range of products available is broad with no sound evidence base for their use.

This discovery project aimed to summarise the available options, create a methodology for evaluating available solutions, and produce research and case studies to help councils develop business cases.

Project outputs

All Local Digital Fund discovery projects were asked to provide the following information at completion:

  • User research report
  • Benefits case
  • Recommendations for next steps

Downloads

Contact [email protected] if you are having problems accessing the outputs.

Project timeline

April 2019

‘Can chatbots and AI help solve service design problems?’ discovery delivers project outputs which are published on the Local Digital website.

Feedback

Each project was assessed using these lenses by the Local Digital Collaboration Unit. We have provided feedback directly to the project teams and this is a summary of what we shared with them.

It aims to be constructive for both the project team and any other organisation wishing to learn about the project or make use of the work done.

  • The team used Slack, Twitter and a project website localdigitalchatbots.github.io, where they blogged regularly and posted videos of their show and tells. This made it easy to understand how the project progressed and helped to build a community of interest.
  • With 13 collaborating local authorities, the project team were able to split up and separately examine 4 different service areas (planning, waste and recycling, revenue and benefits, highways) coming back together to learn from one another’s user research findings. This looked like a good model for other local authorities planning to undertake a collaborative project.
  • The project team selected waste and recycling to pursue at Alpha stage, a service area that was looked into by 4 of the 13 local authorities involved. The team should consider validating this user research with additional local authorities to provide a stronger research base. Gaining a deeper understanding of the level and scale of automation around waste service enquiries will help build the case for chatbots and whether they can resolve queries first time.
  • The project team had a focus on skills and worked closely with their supplier to enable user research training to their partners to help ensure that knowledge and skills were retained within these organisations.
  • The project team openly shared resources including collaboration commitments, user recruitment guides, consent forms, interview scripts and user experience map templates which are likely to be of interest to other local authorities.
  • Project outputs were presented with a strong narrative around the methodologies used and awareness of limitations, with helpful referencing between the output documents to provide further evidence in support of the recommendations made.
  • The benefits case describes a methodology for calculating a ‘Chatbot Target Value’ based on call volumes and first-time resolution rate. It would have been helpful to have more supporting narrative about the methodology and the assumptions made in settling on this model.