A hands-on workshop run at your office, on your own emails, documents and processes. Your team leaves with a signed AI-use policy, a shared prompt library, working team assistants and a ranked list of what to automate next.
Taught on the tools your office already pays for
Most small offices are in one of these three places. The workshop exists to get you out of all of them in a single day.
Copilot or ChatGPT seats were switched on months ago. Two people use them for emails. Everyone else forgot the login. The subscription keeps billing.
Client details, contracts and financials are being pasted into personal chatbot accounts because nobody has said what's allowed. That's a policy problem before it's a training problem.
One person has a great prompt for proposals, another rewrites everything from scratch, a third gave up after it got something wrong. Nothing is shared, so nothing compounds.
Every exercise runs on your team's real inbox, documents and spreadsheets. Every module ends with something you keep.
What these tools are and aren't: a brilliant, fast, over-confident intern with no memory. Where they shine, where they invent things, and why "it was wrong once" is a reason to learn verification, not to give up.
What can and can't be pasted, client confidentiality, the Privacy Act basics that apply to an office, workspace vs personal accounts, turning training-data off. The team edits a pre-filled policy together and the manager signs it off before morning tea.
You keep: a signed AI Safe Use PolicyOne framework, drilled on real work: role, ask, context, example, then iterate. Asking for options, asking it to critique its own draft, asking it what it needs to know. Each person rewrites a real email and a real document from their own week.
You keep: 3 personal prompts eachBreak
Six stations, each person picks two: email and writing, meetings-to-minutes-to-actions, documents and proposals, spreadsheets, reading long PDFs and contracts (with a verification step), and customer replies in your tone. Each station ends with a reusable prompt saved to the shared library.
You keep: a shared prompt libraryLunch
Load your tone guide, templates and SOPs into a persistent assistant so nobody re-explains the business every morning. Teams build one each — proposal writer, customer reply desk, onboarding guide — and test it on today's work.
You keep: 2–3 working team assistantsYour recurring processes go up on the wall and get scored by frequency, minutes and pain. Then we sort them: keep manual, AI-assisted, or automate outright — with a live demo of a real automation running end to end.
You keep: a ranked list with hours saved per itemEach person commits to three uses. The team picks one shared assistant to adopt. The manager gets a one-page measurement sheet, and the check-in call goes in the calendar before anyone leaves.
You keep: a 30-day plan and a check-in dateBy the end of the day, and in the days after, your office has these.
Pre-filled with your tools and your data categories, edited by the team, signed off by the manager on the day.
Every prompt the team wrote during the day, in a shared document your whole office can use and add to.
Two or three persistent assistants loaded with your SOPs and tone, built by your team, tested on real work.
Your recurring processes, ranked by hours saved, sorted into AI-assisted vs fully automatable.
What was covered, who committed to what, the map, and a fixed-price quote for the top automations — if you want them.
We walk the measurement sheet together, unblock whatever stalled, and decide what's next.
We're launching with ten founding-client workshops at half the list price, in exchange for honest feedback and a testimonial.
Two hours a week per person is a conservative figure — the staff survey usually finds more. Replace it with your own numbers on the call.
No proposal cycle, no retainer, no procurement. A conversation, a date, a day.
Who's in the office, which tools you pay for, what prompted the call and what "working" would look like in 30 days. We'll tell you honestly whether the day is right for your team — and offer two dates if it is.
A ten-question anonymous survey goes to your team. The answers pick the examples we use, seed the Opportunity Map and tell us who the sceptic is — they get the first win.
Six hours at your office. Within three working days you have the written report, the map and the shared library. Thirty days later, we check in on what actually changed.
Whatever your office already pays for. Microsoft 365 shops train on Copilot alongside ChatGPT or Claude; Google shops on Gemini alongside one of the same two. The method is the same across all of them, so nothing you learn is wasted when a vendor renames a feature.
That's the second module of the day, and it's why you leave with a signed policy and the account settings fixed. Most offices already have the risk and no policy; the workshop closes it. We never retain client documents used in exercises.
Good. It will get things wrong again, which is why verification is taught as a skill, not a warning. The sceptic's most dreaded task becomes the first live example of the day. A converted sceptic is the best result we get.
No. Free tiers are enough for the day, and we'll tell you on the call which paid plan, if any, is worth it for your team — we don't sell licences.
Up to 12 in one room, at your office, anywhere in Auckland. Larger teams run as two cohorts. Bring laptops and access to your own email and documents; we bring everything else, including a backup internet connection.
Corey from Arkgaze — a practitioner who builds automations and AI systems for New Zealand businesses, not a presenter working from a slide deck. Every example on the screen is from a real office.
Within three working days you get the written report: what was covered, the shared prompt library, the team assistants and who owns each, and the Automation Opportunity Map with a fixed-price quote for the top items. The quote is there if you want it; the workshop stands on its own either way.
Book a free 20-minute call. We'll tell you in that call whether the day is right for your team, and offer two dates if it is.
