DVLA Deploying Agentic AI and Voice Technology across Customer Services

TL;DR:

  • The DVLA is rolling out a large language model system that lets advisers search 2,500 knowledge articles using natural language queries, with full deployment expected within months.
  • The agency is collaborating with GDS and DSIT on Gov Voice, a project exploring AI-powered voice responses for simpler queries across government call centres.
  • CEO Tim Moss framed the AI push as customer-experience-first, with efficiency savings secondary — a notable stance as government agencies face 15% spending review cuts.

The Driver and Vehicle Licensing Agency is expanding its use of AI across a contact centre operation that fields 800,000 phone calls and 300,000 webchat interactions every month, according to CEO Tim Moss.

The agency’s Swansea-based centre employs 150 multitasking advisers handling web chat, email, social media, and telephone enquiries. Moss described the team’s work with agentic AI as “almost black magic” in an interview with Civil Service World.

One immediate deployment is a large language model system that allows advisers to run natural-language searches across the DVLA’s library of 2,500 knowledge articles — replacing the keyword-based lookups they currently use. The tool is due to roll out to all advisers over the coming months, aiming to reduce the time spent finding answers during live customer interactions.

Gov Voice: AI for Government Call Centres

The DVLA is also working with the Government Digital Service and the Department for Science, Innovation and Technology on Gov Voice, a project testing whether AI voice technology can handle simpler queries without routing them to a human adviser.

Moss described the vision as customers being able to interact with an AI system “as if you were talking to a human being and it’s answering back to you.” If successful, the model could extend beyond the DVLA to the roughly 90,000 people staffing government contact centres across the UK.

Efficiency vs Experience

The AI investment arrives as the DVLA faces a mandate to find 15% efficiency savings over the Spending Review 2025 period. But Moss was explicit that customer experience comes first: “It’s not about saying: how can we replace people or do things cheaper? It’s about saying: how can we use AI to deliver a better customer experience?”

This framing positions the DVLA alongside a growing number of public sector bodies — including HMRC and DWP — exploring AI to manage demand rather than simply cut headcount. Whether that distinction holds under sustained budget pressure will be worth watching.

Looking Forward

If Gov Voice proves viable, it could set a template for AI-assisted telephone services across central government — a prospect that would represent one of the largest deployments of conversational AI in UK public services. For now, the DVLA’s phased approach — enhancing adviser tools first, then testing direct customer-facing AI — offers a pragmatic model for other agencies.

Share this article