Go beyond the chat box. Set up your first AI coworker in a hands-on workshop built around your team’s actual work.
One department. Your own files. No coding.
Approved files. A task you know.
“Summarize progress, blockers, and open questions for the team. Use only these files, keep uncertainty visible, and do not assign owners or invent dates.”
Context + instructions + boundariesThe delivery notes describe progress on the current work and flag an unresolved handoff question. The updates do not name a confirmed owner or date for that handoff.
Compare each stated blocker and progress item with its source file; confirm no owner or deadline was added.
Summarize the team's weekly updates using only supplied files; separate confirmed progress from open questions and never infer owners or dates.
Approved files. A task you know.
“Group repeated feedback themes across these notes, include a short supporting excerpt for each, and separate direct evidence from interpretation.”
Context + instructions + boundariesCustomers mention confusion about where to find product guidance. Several notes ask for clearer explanations, while the material does not establish how common that concern is.
Trace each theme and excerpt to the source notes; ensure the summary does not imply frequency beyond the supplied material.
Group supplied customer feedback into themes, attach source excerpts, and label interpretation separately from direct evidence.
Approved files. A task you know.
“Draft a plain-language guide to the first week using only these documents; flag conflicting or missing instructions instead of resolving them yourself.”
Context + instructions + boundariesThe draft groups the checklist, common questions, and team tools into a day-by-day guide. One item needs confirmation because the source documents describe different ways to request access.
Verify the access-request steps and any day-specific instructions against the source documents before sharing the draft.
Turn approved onboarding documents into a plain-language guide; preserve source instructions and flag gaps or conflicts for HR review.
Approved files. A task you know.
“Create a staff-facing summary of the documented expense process. Quote policy wording for requirements and flag anything unclear for finance review.”
Context + instructions + boundariesThe notes describe which materials staff submit and where the request is sent. The source set does not clarify how exceptions are handled, so the draft marks that for finance review.
Check every requirement against the policy extract and confirm exceptions are referred to finance rather than answered by the draft.
Summarize approved expense guidance for staff, quote requirements accurately, and send policy gaps or exceptions to finance for review.
Built for people
who know their work.
Bring the files.
We’ll bring the method.
You set the direction.
You own the result.
Private AI workshops for one department at a time. Start with one recurring task, then build a way of working your team can share.
Work through one recurring task on your own files, check the result, and save the brief you used.
Choose 2 hoursTeams trying practical AI work together for the first time, with one real task per person.
Apply AI to two workflows on your own files, then write reusable instructions and test them with a colleague.
Choose 4 hoursTeams ready to develop and share repeatable ways of handling their everyday work with AI.
Think of it as onboarding a new coworker. Give it context, work through a first task together, and check before you trust the result.
Your
knowledgeContext
Judgement
Work you
review
and own
AI
assistanceDrafting
Organising
Your judgement stays at the centre.
Give it a goal, the right files, your constraints and an example of a good result. Practise improving the instructions when the first draft misses the point.
Check coverage, facts, numbers and tone. Trace an answer back to the files. Decide what is a draft, what is ready to use, and what should stay with a person.
Save the instructions that worked. In the longer workshop, have a colleague try them. Learn when to reuse context and when to start fresh.
A useful session needs somewhere to go next. We build follow-through around a task each person owns, with the manager involved.
Agree a recurring task, the company-approved tool and which files are appropriate to use. Prepare sample files where needed.
Bring your laptop. Work in pairs, try the task, check the output and save the instructions. Leave with a next task to practise.
A day-7 office hour to work through sticking points. A day-30 check-in with your manager to discuss what was used and what needs changing.
No coding is required. The starting point is knowing your own work: a report, a set of documents, a recurring question. We practise together using a tool your company approves.
We choose one platform for the session: Claude Cowork or ChatGPT on your company plan. We check the available features and permissions beforehand, so the exercises fit your setup.
Only the files your company permits in its approved AI workspace. We agree those boundaries before the session. If real files are not suitable, we work with sample material. People review outputs before using or sharing them.
A laptop, access to the agreed tool, one recurring task and relevant approved files. Your manager helps pick a useful starting point. We confirm the preparation together before the workshop.
That is a useful finding too. We teach you to recognise when checking costs more effort than the task saves, when context is missing, and when the work needs a person. Not every task needs AI.
Start with one task you would like help with. Put together a short workshop brief to share with your team.
This creates a file for you to keep.
Nothing is sent, and no booking is made.