About British Gas

British Gas operates in one of the UK's most closely regulated industries, managing millions of customer contacts per year across voice and webchat. Customer needs can be complex and highly variable, and frontline contact centre agents must navigate those nuances consistently across every interaction. The organisation’s regulatory and quality assurance benchmark targets represented a high bar, and achieving these goals remained a challenge. Not because of a lack of talent, but rather a lack of data needed to coach and improve at scale. The organisation needed a way to move from a manual, sample-based approach to something that could match the pace and volume of its operation.

Challenges

Manual assessment could not keep up with a 20-million-contact operation

British Gas had long relied on skilled analysts to manually assess contact centre agent interactions against its quality assurance (QA) and regulatory assurance (RA) frameworks. The Ofgem Regulatory framework governing British Gas covers nine areas of interaction, all of which form part of the company's licence conditions. Alongside this, British Gas assesses quality through its Five Steps to Customer Excellence framework, which contains criteria designed to evaluate conversations from a behavioural and customer experience (CX) perspective.

Manual assessment scored only a limited number of contacts per agent per month and that volume created challenges. Coaching opportunities were narrow, performance management data at the individual level lacked scale, and the business could not always confidently generate the assurance it needed to confirm contact centre agents were consistently meeting their quality and regulatory obligations. With a desire to address this, automating QA and RA became the logical next step.

"We traditionally manually assessed quality and regulatory compliance,” said Jonathan McWilliams, Operational Root Cause Manager at British Gas. “And while we have some really talented individuals doing that role, it's limited in terms of volume, can be prone to inconsistency, and it’s very difficult to drive performance at an agent level because you're only looking at a limited volume of data points.”

Building an automated quality and regulatory engine

British Gas launched its Setting the Standard for Quality programme, a cross-functional effort to drive technical, cultural, and behavioural change across the contact centre. At its core was the automation of both the QA and RA frameworks using CallMiner, replacing manual sampling with consistent, automated scoring across every interaction.

The build reflected the full complexity of the challenge. The team constructed dozens of separate scores within CallMiner, comprising hundreds of individual components designed to assess whether contact centre agents respond to customer needs correctly. Confirming compliance requires more than reviewing the conversation alone, so the model was overlaid with multiple metadata items drawing on separate data sources. This gives the business a complete picture of whether agents are saying the right things, as well as taking the right actions within core systems.

Delivering the solution required contributions from across the business. Technology teams built the integrations needed to ingest webchat interactions into CallMiner alongside voice. Training and change management functions ensured frontline agents and team leaders understood what was changing and what was expected of them. Because the Five Steps to Customer Excellence framework takes a behavioural science approach that deliberately avoids prescriptive language, the team worked closely with their consultancy partner to research natural language variants and ensure the CallMiner components stayed true to the spirit of the framework. The frontline operation then participated in extensive testing until the automated scores reached a minimum 80% agreement threshold with human reviewers before the solution went live.

"CallMiner is straightforward and it’s easy to develop categories and scores, isolate language in relation to an emerging challenge, and then start to understand the impact that is having,” said Jonathan McWilliams. “We are able to generate coaching insights from customer interaction data and take this information to agents, to advise them appropriately."

Results

Millions of assessments, measurable improvement in quality and compliance

Since launch, British Gas has delivered millions of assessments across its QA and RA frameworks. The organisation has moved from thousands of manual assessments annually to a model operating at an entirely different order of magnitude.

Quality scores against the Five Steps to Customer Excellence framework have improved by approximately 10%, with a general upward trend. On the regulatory side, we have assurance that multiple scores now regularly achieve or exceed target with positive movement across the remainder, all as a direct result of the coaching and feedback loop.

Coaching activity through CallMiner continues to increase, reflecting strong operational buy-in from team leaders across the business. This is being driven by a regular and consistent flow of insights at scale, enabling targeted coaching that supports continuous development, highlights areas for improvement, and recognises success where colleagues are delivering great outcomes for customers.

British Gas continues to refine the model through a frontline feedback process that gives contact centre agents and team leaders a structured channel to challenge scoring and contribute to ongoing improvement. What began as a challenge to provide more robust and reliable data has become the foundation of a coaching culture that is built to last.