An academy that puts AI to work, team by team.

Guardrails first, then the baseline, then the skills.

The Academy began by capturing a quantified before-state for the cohort: AI literacy, time on manual work, operational bottlenecks. Responsible-AI guardrails were agreed at the same time: a person in the loop, no tenant or personal data into any tool, bilingual considerations written in from the start.

Then a blended cohort of around twenty staff, learning on the organisation’s own processes rather than exercises: repairs, arrears, appointments, resident contact, in workshops with coaching in between. Alongside it, development for leaders so the rules for using AI can be set and overseen, and data ethics throughout.

Routine work handed to the tools, judgement kept with people.

Resident communications about repairs, arrears and appointments were the first territory: routine drafting, sorting and summarising handled with AI inside the association’s own Microsoft environment, with a person approving anything that changes an account or affects a tenancy. Departmental playbooks and prompt packs were written for the cohort’s own functions, with bilingual templates for common tenant communications, grounded in the sector’s own rules: language standards, tenancy legislation, and the regulatory regime for data governance and audit trail. Time saved is measured per task and per week, against the week-one baseline.

A capability that belongs to them.

A permanent digital capabilities lead now in post. Staff trained during the build to make and release changes themselves. AI champions identified and briefed so knowledge cascades without us, and train-the-trainer throughout. The pilot closed with an impact report against the baseline and a costed plan for organisation-wide rollout, on evidence the association owns. Data and AI literacy now reaching the wider organisation, with Scout’s involvement reducing as theirs grows.