In six weeks your team puts one recurring deliverable into production — faster and more consistent — on a shared, governed knowledge base inside the tools your organisation already approves, with adoption that survives the workshop.
For a functional team and the manager who runs it. The point is to lift one repeating output: faster cycle time, consistent quality across every person, and an auditable trail that satisfies IT and security. We work entirely inside your sanctioned tools and measure against your own loaded labour cost.
Often bought to get ahead of shadow AI rather than ban it — giving the team an approved surface and a shared standard so the behaviour comes into the light.
In week one the team names one output it produces over and over, that several people touch, where inconsistency or turnaround already hurts. That output is the spine for all six weeks.
Good candidates: a weekly board pack with written commentary, a held-to-standard customer-response set, month-end variance commentary, a standardised onboarding pack, an RFP response. We build the shared prompt, the team knowledge base, the review step, and the adoption around that one thing.
One canonical asset in a shared library — not eight private versions. Governed, logged, and embedded as the default first step of the workflow.
Find where the team bleeds time and consistency. Pick the one output.
Trace the path of a candidate output together — trigger to send — marking every time leak and every drift between people, then vote to lock the spine. Why a team game differs from everyone running their own tab.
A one-page charter: the output, top three leaks, who reviews, and the success metric.
Use the approved surfaces well. Decide, in writing, what's safe to put in.
Set up a shared workspace on the sanctioned tool — Claude for Work, Copilot, or ChatGPT Enterprise — scoped to the right material, and co-write the data-handling one-pager using the team's actual inputs as test cases.
A live shared workspace + a one-page data-handling rule everyone has read.
Get the team's standards out of senior heads and into shared files.
Interview the team like journalists; structure what comes out into five shared files — checklists, rules, reference data, worked examples, and the failure log. The base is co-authored, owned by the team.
A five-file /team-knowledge base — 8–20 pages of your own standards.
Wire the standards into a prompt the whole team runs the same way.
Author the shared asset against the skill anatomy, wire in the knowledge files, set the house output format and the guardrails — then run it cold against a known example and red-pen every miss as a group.
A working shared prompt/agent in the library, plus its first cold test run.
Close the gap on real work — and get the whole team using it.
Three live hot-seat runs on this week's work: run cold, diagnose the cause, patch, re-run, confirm. In parallel, name champions and agree the embedding hook so the asset is used when the cohort ends.
Three live runs, a scoring trend, named champions and an embedding hook.
Make it stick without the facilitator. Put a real ROI number upstairs.
A champion walks the asset back as if onboarding a new starter, we compute team-level ROI from the tracker, set the maintenance cadence and a named owner, and map the next output to standardise.
A handover, a one-page ROI summary for leadership, a maintenance owner, and asset #2 mapped.
A 9-person team spends most of a day on the weekly leadership pack. The commentary reads differently depending on who drafts; two reviewers needed.
The commentary step standardised on Copilot against a shared "what we always explain" file. ~90 minutes, one consistent voice, one reviewer.
Eight reps each wrote escalation responses in their own voice — quality swung from excellent to terse, ~25 min each, heavy QA.
A shared asset holds the team's apology-and-remedy structure; two champions coach the rest. Consistent house voice, ~9 min, light QA.
A 7-person team's first-draft RFP response took ~3 days, and tenders were declined for lack of capacity.
First drafts in ~1 day, more tenders answered, and consistent win-themes drawn from the shared knowledge base.
+$900 per extra seat to a 12-participant cap. Single PO and invoice.
For teams that prefer headcount-based budgeting.
Monthly check-in + tracker review to keep the ritual alive and report ROI.
A team of 10, each recovering 3 hours/week at a loaded ~$75/hr, is roughly $115,000/year of recovered capacity — payback inside the first six to eight weeks of real use, before counting consistency and de-risked shadow AI.
Pricing in AUD · sold per team · draft, confirmed before each cohort
Book a call and we'll find the recurring deliverable worth standardising — and shape a cohort your IT and security teams will sign off on.