Training
Everyday AI for administrative functions
A one-day workshop for non-technical teams that do document-heavy work — pulling figures out of formats that never match, reconciling and researching numbers and figures. Built to embed in how the team already works, so the change holds the next day instead of staying a nice idea.
- Audience
- Operations, legal, ESG, and corporate-reporting teams
- Format
- One-day workshop, on-site preferred
- Duration
- 1 day
Who this is for
Teams whose week is full of careful but repetitive document work — legal, ESG, finance operations, group and corporate reporting. This is usually not a tech-forward audience, and it doesn’t need to be. Most have sat through a generic AI session, nodded along, and changed nothing on Monday morning. The tools looked impressive in the demo but bore no resemblance to the actual task in front of them.
We start from the opposite end. We work inside the team’s real, recurring tasks and build from there. Because for this kind of work, a workshop that doesn’t connect to the current context is a pleasant afternoon that produces no change.
What we cover
- Your actual workflow first: we map a real recurring task the team owns, step by step, before any tool comes out. The tool has to fit the work, not the other way around.
- The same numbers, different shapes: consolidating figures that arrive in inconsistent formats, reconciling them, and normalizing units; the kind of collation that eats hours every cycle.
- Research you can trust: how to do focused, repeatable research that draws only on the sources you already rely on, rather than whatever a model decides to surface.
- Checking the output: where these tools quietly get it wrong, and the habits that catch it. Because in this work, a confident wrong number is worse than no help at all.
- Making it stick: turning what we built together into a small, repeatable routine the team can run on its own after we leave.
What we don’t cover
- AI 101 and the hype. We’re not here to convince anyone the technology matters; we’re here to make this week’s work faster.
- Building software. Nobody needs to write code or learn anything an engineer would. We assume a ChatGPT equivalent tool the team can already use.
- Replacing judgment. The team stays accountable for the numbers and the wording; we make the path to them shorter, not automatic.
Format
One day, on-site where possible so we can sit with the team’s real material. Ahead of time we ask for a couple of the recurring tasks that take the most time, and we use those as the working material throughout the day. Limited to 8 people.