£4m in annual savings identified.
Contract renewal won.

Nobody could say what AI was worth, because nobody could say what operations already cost.

Leadership and operational teams had no working knowledge of agentic automation or retrieval-augmented generation (RAG), and no sense of how either could be applied to streamline day-to-day operations — incident triage, contract retrieval, customer communications, or anything else.

Worse, the organisation had no reliable baseline for its own operating costs. Without one, there was no way to size the financial benefit of automating any given process, prioritise which use cases mattered most, or produce a savings forecast credible enough to put in front of the contract owner as part of a renewal bid.

Three constraints were non-negotiable:

• No shared understanding of what agentic automation or RAG could actually do for the business
• No cost baseline to size the financial upside of any candidate use case
• The business case had to be credible enough to win a competitive renewal bid, not just plausible on paper

Workshops, a scoring framework, and rapid proofs of concept.

Workshops with operations, customer support and back-office teams surfaced roughly 80 candidate use cases, each tied to a real pain point and its time or cost driver. Every candidate was scored against three criteria: ease of implementation, expected impact, and cost to implement.

The scoring surfaced the most complex but highest-potential opportunities — automated triage and remediation of service incidents, intelligent document and contract retrieval, and contextual, automated customer communications — as the strongest candidates for agentic automation and RAG. The top 10 were developed into full business cases: savings calculations, implementation costs, risk assessments and the technology stack required, including cloud-hosted models, vector databases and orchestration layers.

Deliverables covered the full path to execution: an AI adoption framework defining governance, roles and a phased rollout; a technology roadmap and transformation path with integration, security and cloud recommendations; use-case walkthroughs and rapid PoCs to pressure-test feasibility; and a programme plan with timelines, resourcing and KPIs tied directly to the bid.

The savings that won the bid.

The savings case and the delivery plan behind it did the work: the customer won the contract renewal on the strength of the lowest bid, backed by a demonstrable, costed commitment to AI adoption — proof that a clear-eyed automation roadmap can win business, not just cut costs after the fact.