760,000 Calls. 120 Advisors. How Does Any Council Handle This?

Bristol City Council’s experience with AI in customer services is one of the most honest accounts we’ve heard.

Most council customer service teams are operating under the same arithmetic.

Demand keeps rising. Budgets keep shrinking. Advisors are highly trained but stretched thin. Residents increasingly expect 24/7 access to services that historically operated inside office hours.

Bristol City Council was no different. Before introducing its AI digital assistant, Briz, Bristol’s Citizen Services team was handling nearly 760,000 calls a year through a workforce of around 120 advisors. Average wait times sat at roughly eight minutes. The council answered around 75% of calls. Teams had already done the things local government always does first: process improvement, channel shift, careful triage, service redesign. It still was not enough.

“We were operating under severe cuts. Teams were getting smaller. The volume of contact was going up. And regardless of what we did, our performance was actually quite low. We had to do something different.”

Rizwan Tariq, Head of Citizen Services, Bristol City Council

That challenge is now sector-wide.

Local government has spent more than a decade operating under sustained financial pressure. At the same time, citizen expectations have shifted dramatically. Residents compare council services not against other councils, but against the digital experiences they receive from banks, retailers, insurers and utilities.

The question therefore stops being:

“How do we fund the contact centre we have?”

And becomes:

“How do we redesign the front door so the workforce we can afford can still deliver the service residents need?”

AI Was Not Introduced to Replace Staff

One of the most important aspects of Bristol’s approach is what the programme was not trying to do.

This was not a redundancy programme disguised as innovation.

The objective was to absorb high-volume, repeatable contact that did not require a human advisor in the first place, allowing experienced staff to focus on complex, sensitive and high-value cases where empathy and judgement matter.

• Missed bin reports

• Council tax information

• Clean Air Zone queries

• Routine transactional demand.

  
  

These are precisely the kinds of interactions AI can now manage effectively at scale.

Meanwhile, advisors gain time to focus on safeguarding concerns, housing complexity, complaints resolution and vulnerable residents.

“Advisors now have breathing space. I want them focusing on complex stuff. I don’t want them answering the simple things all day.”

Rizwan Tariq

That distinction matters because too many conversations about AI in customer services still reduce to a single question:

“How much can it cut?”

Bristol approached it differently.

The starting point was service quality.

Residents were waiting too long. Too many calls were abandoned. Staff were under pressure. The AI deployment was designed to improve the citizen experience first. The efficiencies followed from that.

The Trial Changed Minds

One of the clearest lessons from Bristol’s experience is that AI adoption accelerates when people can experience it directly.

Bristol first encountered ICS.AI through a webinar similar to the one hosted this week. Like many councils, they were interested but cautious.

The turning point was the trial.

ICS.AI configured a tailored phone AI trial using Bristol-specific services and knowledge areas. Senior leaders, elected members, service teams and customer advisors were all given a live number to call and test themselves.

That changed the conversation entirely.

“If you can give decision-makers a phone number to call and try themselves, it’s incredibly powerful.”

Rizwan Tariq

The discussion moved from: “Should we do this?”

To: “How do we implement this safely?”

And in local government, that second question matters enormously.

Governance Is Not a Barrier to AI. It Is What Makes AI Possible.

Bristol’s deployment took longer than a private-sector rollout might have done.

That was deliberate.

The programme included:

• DPIAs

• Equalities assessments

• Information security reviews

• Member engagement

• Staff consultation

• Accessibility testing

• Diverse language and accent testing

• Resident stakeholder involvement

The council worked with housing tenants, disability groups and residents where English was not a first language. Advisors were deeply involved in shaping intents, refining responses and improving journeys.

The result?

According to Rizwan Tariq, the launch itself was “a non-event”.

In public services, that is arguably the highest compliment possible.

Done properly, governance does not slow AI adoption down. Governance is what allows AI to operate safely, ethically and credibly in the first place.

The Early Results Are Significant

Only a few months after launch, Bristol is already seeing measurable operational impact.

Briz has now:

• Handled more than 176,000 resident calls

• Answered over 436,000 questions

• Supported 31,000+ out-of-hours queries

• Achieved around 90% positive web satisfaction

• Improved call answer rates by approximately five percentage points

• Reduced average wait times

• Expanded service coverage across more than 20 council services

Importantly, Bristol reports that service pressure now “feels calmer” internally.

That matters.

Because sustainable transformation is not just about metrics. It is about creating operational breathing space inside services that have spent years under strain.

AI Transformation Is a Journey, Not a Single Deployment

Bristol is clear that this is only the beginning.

The roadmap now includes:

• Broader GenAI capability

• ultilingual interactions

• Deeper service orchestration

• Expanded automation

• Improved contextual awareness

• Enhanced accessibility and localisation

As generative AI evolves, councils increasingly move from answering questions to orchestrating outcomes across departments.

That maturity journey is already visible across local government.

And the councils moving fastest are not necessarily the most aggressive.

They are the ones approaching AI with:

• Governance

• Pragmatism

• Operational realism

• Measurable outcomes

• A willingness to test before scaling

The Place to Start Is Not Procurement. It Is a Trial.

If your council is asking the same arithmetic question Bristol faced:

How do we serve more residents with fewer hands, without compromising quality or governance?

The place to start is not a procurement exercise. It is a controlled trial.

A live environment. Your services. Your data. Your residents. Your governance processes. Your advisors. A phone number you can ring yourself.

Book your tailored AI Front Door trial workshop

---