CultureAI
Public Sector

Governing AI use across a 6,000-person local authority

Staff were using AI to draft letters, summarise casework and answer resident queries. The authority needed adoption to continue, with resident data protected and every interaction accountable.

Industry
Local government
Size
6,000 staff
Region
United Kingdom
0%

of shadow AI use moved to approved tools

0/10

policy violations resolved by a nudge, not a block

0%

of AI interactions logged for the governance board

The challenge

Resident casework contains some of the most sensitive personal data there is, and it was appearing in prompts across dozens of AI tools, most of them personal accounts invisible to IT.

As a public body, the authority had obligations its AI policy alone could not meet: data protection, freedom of information, and a governance board asking for evidence rather than assurances.

The approach

Discover, control, detect, prove

01

Discover

Sensors gave the security team its first real picture of AI use across every directorate, from contact-centre chatbots to AI features inside existing case-management SaaS.

02

Control

Real-time nudges redirect staff from personal AI accounts to the approved Copilot deployment, with hard blocks reserved for resident-identifiable data.

03

Detect

Context-aware detection recognised resident personal data in prompts and uploads without a data-classification programme the team had no capacity to run.

04

Prove

Every interaction is logged and reportable, giving the AI governance board and information-governance team a standing evidence base.

The results

  • 78% of shadow AI usage moved to the approved tools within three months, driven by redirection rather than restriction.
  • Nine in ten policy violations were resolved by an in-the-moment nudge; blocking is now the exception, not the default posture.
  • The governance board reviews actual usage evidence quarterly, and the authority can answer "how are we managing AI risk?" with data.
app.culture.ai· In-flow guidance LIVE
Employee, in their AI tool
“Draft a renewal email for [customer list pasted]…”
CultureAI

This paste contains customer personal data. Remove it, or continue in the approved workspace.

Open approved toolEdit paste

Guidance in the moment, not an alert in a queue

We could never have justified blocking the tools our teams were using to serve residents faster. Redirecting them to the approved route, and being able to prove it, changed everything.
Head of Information Governance, local authority

Uncover hidden AI risk

Start a free 2-week AI Risk Assessment. Easy setup. Fast visibility.