Award-Winning AI at the Front Door: How Renfrewshire Council Transformed Customer Services

From pressure to progress at the front door

Councils across the UK are facing sustained pressure on customer services. Demand continues to rise while resources do not. For many authorities, the challenge is no longer whether to explore AI, but how to do so in a way that delivers real value, protects trust and enhances services.

That was the backdrop for our recent webinar on AI at the front door for council services. At its heart was Renfrewshire Council’s story - a practical, evidence-led journey that shows what is possible when AI is introduced with clarity of purpose and strong leadership.

Starting with evidence, not assumptions

Renfrewshire’s journey did not begin with technology for its own sake. By 2023, the council had already reduced telephone demand significantly by shifting over 100,000 contacts to online services. Customer behaviour data showed residents were increasingly comfortable using digital channels, and the council’s five-year customer strategy set out a clear ambition to expand choice and accessibility further.

AI became the next logical step - but only if it could be proven to work in the real world.

Rather than committing immediately to a full deployment, Renfrewshire ran a controlled phone-based trial using an AI digital advisor. The goal was simple. Test the experience properly, understand how residents and staff responded, and use real data to inform the business case.

That trial proved to be pivotal. It moved conversations from theoretical to practical, built confidence internally and created a clear evidence base for decision-making.

Exceeding expectations at scale

When Renfrewshire’s AI digital advisor, Millie, went live, expectations were deliberately conservative. The original business case assumed that between 5 and 10 percent of calls might be handled end-to-end by AI. The reality has been very different.

Since launch, Millie has handled nearly 400,000 calls and answered more than 700,000 questions. Around 40 percent of customer queries are now resolved without human intervention, far exceeding the original projections.

The service operates 24 hours a day, seven days a week, giving residents access to help at times that suit them.

Crucially, this has not come at the expense of service quality. Customer satisfaction has increased, rising from 43 percent positive to around 70 percent. Complaints related to the AI service remain extremely low when set against the volume of interactions.

Designing for trust, inclusion and Scottish voices

One of the most common concerns raised about voice AI in public services is whether it can genuinely understand local accents and everyday language. This concern is particularly strong in Scotland, where regional dialects are an important part of how people communicate.

Renfrewshire took this seriously from the outset. The AI solution was tested, trained and refined using real-world language and local phrasing. It was also designed with clear human fallback routes so residents could always reach an advisor if needed.

The results speak for themselves. Over 98 percent of calls achieve successful speech recognition, directly addressing one of the biggest public concerns around AI on the phone.

This focus on inclusivity closely reflects what we are seeing nationally. Early findings from our recent national AI survey show that while residents are open to AI in council services, that openness is conditional. People want AI that is trustworthy, well governed and designed to work for everyone. Renfrewshire’s approach aligns strongly with that expectation.

Supporting staff, not replacing them

Another key theme from the webinar was the impact on staff. From the beginning, Renfrewshire involved its customer service teams in shaping the AI service. Advisors helped identify the most common questions, tested responses and fed back on what worked and what did not.

Rather than removing roles, the introduction of AI has allowed advisors to focus on more complex and sensitive enquiries, while being upskilled to work alongside new technology. Millie is now talked about as part of the team, not as a threat to it.

Recognition and lessons for others

Renfrewshire’s work has received national recognition, including COSLA award success and wider industry awards for service deliver and innovation. For the team involved, this recognition reflected not just the technology, but the way the change was delivered.

Several clear lessons emerged for other councils considering a similar path:

• Start with a trial before scaling

• Build the business case on real data

• Communicate early and often with staff, members and stakeholders

• Design for trust, accessibility and local needs

• Treat AI as a service transformation, not a technology project

From proven results to confident next steps

Renfrewshire’s experience shows what’s possible when AI is introduced with clarity, strong governance and a relentless focus on real service outcomes. This isn’t experimentation for experimentation’s sake. It’s AI operating at scale, on the phone, delivering measurable impact for residents, staff and the organisation.

But the webinar went beyond Renfrewshire’s story alone. Attendees also explored how lessons translate into a smarter, more unified approach to the customer service front door - where AI acts as a consistent first point of contact across channels, guiding residents quickly to the right answer, service or human support.

Live demonstrations showed what this looks like in practice, including:

• Adult social care and children’s services journeys

• A multilingual demonstration, highlighting accessibility and inclusion

• Real examples of end-to-end conversations, including SMS follow-ups

• Direct feedback from end users on usability, inclusivity and trust

The session also unpacked how councils can build a robust business case for AI using real operational data rather than assumptions - and why customer services is often the most effective place to start.

Importantly, the webinar also introduced how proven success at the front door creates confidence to think more broadly - supporting the development of customer service target operating models where AI and human teams are designed to work together by default.

As part of this wider discussion, ICS.AI outlined how these foundations support the next phases of front-door AI development - extending beyond voice and digital into physical environments, and underpinned by a newly announced guarantee that reflects platform maturity, strength of evidence and consistency of outcomes across deployments.

Watch the full webinar

Hear directly from Renfrewshire Council’s leaders, see the live demonstrations, and explore the data, decisions and lessons behind one of the UK’s most advanced AI customer service deployments. Watch the full webinar.

Ready to take the next step?

If you’re exploring how AI could improve access, reduce avoidable demand and protect staff time in your own organisation, the next step doesn’t have to be a leap.

A Customer Service AI Trial allows councils to test AI safely, gather real data and build a credible business case before scaling - exactly the approach Renfrewshire used to move from curiosity to confidence.

For organisations ready to look beyond the front door, a wider AI assessment can help identify where AI can deliver the greatest impact across services, operating models and resident experience.

---