The UK's Driver and Vehicle Licensing Agency now handles roughly 900,000 calls per month through a natural language processing (NLP)-powered interactive voice response system that has slashed average caller navigation time from three minutes to 90 seconds. The deployment, reported in June 2026, offers a measurable benchmark for high-volume contact centres looking to speed up routing and reduce reliance on touchtone menus.
The problem with touchtone menus
The DVLA's legacy phone system relied on layered keypad menus that forced callers through multiple tiers of options before they could reach the right adviser. Published DVLA transparency documents described the setup as delivering a "clunky customer journey." Handling around 900,000 calls each month meant those extra minutes per call compounded into significant operational drag. The agency needed a way to simplify navigation, improve routing accuracy, and keep pace with rising demand for digital services without cutting off citizens who still preferred calling.
How NLP replaced keypad selections
The new IVR uses Google AI technology to let callers describe their enquiry in their own words instead of tapping through menus. The platform analyses intent and then routes accordingly. Depending on what someone says, they may receive automated answers, get an SMS with a gov.uk service link, land with the right adviser, or complete the transaction through self-service - no agent needed. The system sits inside a wider contact centre transformation that also uses cloud-based infrastructure for service delivery and operational resilience.
Measurable gains from the new system
Since the switch, average call-navigation times have dropped from roughly three minutes to 90 seconds. Around 20,000 calls per month now reach the correct adviser automatically, with routing accuracy improved through better identification of caller intent. The DVLA also reports gaining clearer visibility into the types of enquiries coming in, which helps longer-term planning.
For professionals working with AI for Customer Support, the numbers provide a concrete reference point. A 50 percent reduction in navigation time and thousands of accurately routed calls each month show that NLP-based IVR can deliver hard operational returns in live government contact centres, not just smaller pilots.
Why this matters for customer support
The DVLA's shift strips away one of the most persistent friction points in high-volume support: the time callers waste just trying to be understood. When customers describe a problem in their own words and get routed correctly the first time, handle times drop and resolution rates tend to rise. For support teams managing thousands of daily interactions, the case suggests that moving beyond traditional menu trees to intent-based routing can pay off in both speed and queue management, even with strict public-sector constraints on cost and compliance.
Your membership also unlocks:
