National Lottery

Premier Lotteries Ireland / National Lottery

AI Agent Workshop

Microsoft Copilot Studio — two days of hands-on build, plus an optional governance half-day on Wednesday 14 October, delivered hands-on around one worked scenario.

Dates
12–14 October 2026
Where
On-site at the National Lottery offices
Group
12 attendees
Sign in to your workshopView the running order

Materials, labs and progress are for enrolled attendees. Accounts are set up by Richard Foley before the workshop.

The scenario

Internal support agent

One worked scenario built in depth: an internal support agent that uses approved example support knowledge to gather information, provide first-line guidance and, following user confirmation, prepare a structured support ticket.

The programme deliberately builds one scenario in depth rather than starting several agents superficially. Grounding, tools and flows, confirmation, escalation and testing transfer directly to other use cases across the organisation.

Structure

Pre-work

Before the room

Required: the baseline diagnostic, about 15 minutes. Full pre-work is about 90 minutes: two recommended 30-minute primers, the 15-minute diagnostic and optional 15-minute scenario reading.

4 sessionsMonday 12 Oct · Day 1

Foundations and first agent

Theory and guided discussion in the morning, hands-on labs in the afternoon. By close of play every attendee has a working, grounded support agent.

12 sessionsTuesday 13 Oct · Day 2

Integration and orchestration

The agent stops answering and starts acting: tools, flows, a real ticketing pattern, human confirmation and escalation.

12 sessionsWednesday 14 Oct · Half day, 09:30–13:00

Governance awareness

Awareness-level module covering what teams should understand before an agent moves anywhere near production.

7 sessions

Delivery model

Everything runs in the Artellis non-production training environment

Each attendee receives temporary individual workshop access, prepared and tested in advance. Workshop time is spent learning and building rather than resolving tenant or infrastructure issues.