Prompt Engineering Sprint — a hands-on AI workshop
Your real work questions, solved with AI
We map your work questions beforehand. During the sprint teams solve them with AI, share what works and build a prompt library for the organisation.
- An intake call beforehand to collect your cases
- Working with ChatGPT, Claude and Microsoft Copilot
- Your own prompt library as a tangible result
- Ideal for knowledge teams and management
What will you be doing?
The sprint is built around the real questions that cost you time every day: summarising emails, scanning tenders, answering customers, cleaning up minutes. In three hours, teams build a set of prompts and workflows they can use tomorrow — and which we leave behind in a shared prompt library.
Programme
00:00 — Intake recap (15 min)
We show the three to five cases we collected from you beforehand. Teams pick one to start with.
00:15 — Sprint 1: 'The prompt as a product' (45 min)
Teams build a first version of their prompt, test it on real input and measure: what works, what's missing?
01:00 — Cross-team review (15 min)
Teams present briefly to each other. The best insights go into the shared library.
01:15 — Break (10 min)
01:25 — Sprint 2: Building chains (55 min)
Teams build a workflow of several prompts, or pick a new case. We work with ChatGPT, Claude and Microsoft Copilot to see where each model excels.
02:20 — Documentation (25 min)
Teams write up their prompts on a shared Notion or Miro board, with notes on when to use each one and when not to.
02:45 — Wrap-up (15 min)
Everyone gets a personal 'what am I doing tomorrow' assignment, and we show the prompt library live.
What do participants need to bring?
- One laptop per person (this workshop is hands-on)
- Access to their own work files to practise with (optional)
Which kind of team is this for?
Knowledge teams, management, HR, marketing, sales and policy. Less suitable for teams who don't regularly work at a computer.
Which teams is this for?
Teams who want to use AI for their own daily work, not for a demo.