Bought, not used
Copilot or ChatGPT seats sit idle because nobody practised on the actual memo, bid or SOP.
For organisations · 1 day · On-site
A one-day on-site workshop so employees leave knowing how to draft, summarise, search and check work with AI — inside your tools, your data rules, and a named plan for adoption. Not a webinar. A joinable cohort in your room.
The stall
Most AI programmes fail on access, evaluation and ownership — not on model quality. People try a chatbot once, then go back to the old draft. Shadow tools appear. Legal is uneasy. Nobody owns Monday morning usage.
Copilot or ChatGPT seats sit idle because nobody practised on the actual memo, bid or SOP.
Staff drop customer data into public models because the approved path was never shown.
A launch webinar is an event. Adoption is a curve: champions, measurement, and a named owner.
What employees leave able to do
The day is taught with worked examples — tenders, emails, reports, SOPs — not a product demo and not a laptop lab.
Produce a memo, bid section or report from approved sources, then spend time judging it.
Compress a long document and show where each claim came from. Fluent and wrong is not acceptable.
Use the company assistant or licensed Copilot / Gemini / ChatGPT Enterprise without leaking data.
Approvals, customer numbers, HR and legal stay with a person. The model recommends; it does not sign.
Who should join
Do not send a mixed lecture. Nominate people who will actually change a weekly artefact.
CIO, CHRO, business heads. Where AI may decide vs recommend, who owns accuracy, what to measure in 30 days.
Tender, finance, legal, HR, engineering. How to draft, retrieve and cite from approved libraries — without leaking data.
Plant, mine, branch, service desk. Short answers from SOPs, checklists, and when to escalate to a person.
1 day · on-site
One day on-site. Facilitated sessions and worked examples. No laptop labs; people leave with methods they can apply the next morning.
Access, evaluation, ownership. What “used” will mean in your organisation.
What may be pasted, what must never leave, and the human sign-off line.
How to compress a long document and check every claim. Fluent-wrong answers, and how to catch them.
How a memo, note or bid section should be prompted, grounded and reviewed before it leaves the desk.
Leadership / knowledge work / operations — what each seat should do differently from Monday.
One champion per team, a usage check, and the five workflows worth automating next.
Q&A and how the next cohort in your organisation will run.
How a cohort runs
We brief, run one day on-site, then stay for a 30-day usage curve.
Tools you already licensed, data that must stay in, who must attend, and the room on the day.
One day on-site. Taught sessions and role tracks. Written acceptable-use people can actually follow.
IT and Legal see the same rules as staff. DPDP-aware. No public paste of customer or employee data.
Champion clinic, usage check, optional LearnOps attendance pack. Then the next cohort.
Guardrails
The workshop assumes a licensed enterprise tool — Microsoft 365 Copilot, Google Gemini, ChatGPT Enterprise, or a 22D assistant on your identity. Public consumer chat is treated as out of bounds for company work.
Examples on the day are generic or sanitised. Production secrets stay out of prompts.
Every use case states: decide, recommend, or stop. A person signs anything that leaves the organisation.
Packages
Prices stay off this page until a briefing. Forward this URL; we will scope after 30 minutes.
First join · up to 25 people
The usual first expansion
Rollout with evidence
Join a cohort
Nominate a first cohort. We reply within one business day with dates, the on-site room we need, and what IT must confirm.
If your mail app opened, send the draft to complete the hand-off. We reply within one business day.
Or write Feby@22d.ai / Yogendra@22d.ai. Next step is a 30-minute briefing.
Enable AI on real work.
Join a cohort