From scattered experiments to a structured AI capability across 120 staff

The situation

A further-education college group knew its staff were already using AI — unevenly, informally and largely invisibly. Some teachers had built genuinely useful routines; others avoided it entirely; students were far ahead of policy. Leadership faced the standard triple pressure: workload, academic integrity and an inspection-ready governance position, with no structured capability underneath any of it.

One-off twilight CPD sessions had been tried and had not stuck. What was missing was not enthusiasm but structure: levels, role-relevance, practice time and a governance framework people could actually follow.

What we did

We ran the six-month AI Capability Academy across the group. A baseline capability assessment placed staff into beginner, emerging, intermediate and advanced tracks. Teaching staff worked on planning, resource creation, differentiation and feedback workflows; business support teams worked on administration, communications and reporting; leaders took a dedicated strand on governance, acceptable use and assessment integrity.

Monthly workshops were paired with role-based practice projects, a shared workflow library the staff built as they went, and a champions network — twelve staff trained to coach colleagues and triage questions between sessions. CedarPro drafted the acceptable-use and assessment-integrity guidance with the leadership team in month two, so practice and policy grew together rather than colliding at the end.

The human role

Teachers remain the professional judgement in every workflow the programme touched: AI drafts and differentiates, staff review and decide. Assessment decisions, feedback ownership and safeguarding judgements stayed explicitly human throughout, and the acceptable-use policy records that boundary in plain language.

Results

One hundred and twenty staff completed the programme. At exit assessment, 86% reported confident, governed use of AI in their role, against 31% at baseline. The shared library holds around thirty staff-built workflows in active use. The champions network now runs induction for new staff without external support, and the group’s AI guidance has been through governors once rather than rewritten three times.

“Previous CPD gave people a taste and then left them alone. This gave them a level, a path and colleagues to ask. That is why it held.”
— Vice Principal, FE college group (anonymised)

What the client keeps

The levelled curriculum and materials, the staff-built workflow library, the acceptable-use and assessment-integrity policies, twelve active champions, the exit capability assessment for future cohorts, and an induction module the college now runs itself.