Half a day.
On-site.
By lunch,
three names.
Your team builds a working AI agent on a real workflow from your agency. Not a demo. Not a training session. A working system, in your tools, owned by your people. By lunch, the room knows who the first installs are.
Format
Half-day
Location
On-site
Output
Working agent
Credit window
14-day
Who books this
Agency founders, MDs, and creative directors who suspect AI is about to change the economics of their team — and want to surface the leverage before someone else does.
You run a creative agency — marketing, brand, production, film, web, specialist. 8–80 people.
You feel AI pressure — from clients, competitors, or the billable-hours model itself — and you don't want a strategy deck. You want it done.
You can put the right people in the room — not the whole agency, but the 4–6 operators whose workflows will change first.
You're willing to act on what surfaces — the workshop names the first installs. The 14-day credit window assumes you'll move.
Not for
- Procurement-heavy organisations who need three sign-offs before a pilot
- Teams shopping for an "AI strategy deck" — this is not a slide output
- Anyone whose founder won't sit in the room
- Generic AI training for the whole company
Who must be in the room
The people whose workflows actually run the agency — not the IT team, not the innovation lead.
The founder or MD — non-negotiable. If the decision-maker isn't in the room, the installs won't land.
2–4 senior operators — the people who own production, accounts, strategy, content, SEO, or whatever the agency's core workflow is. These are the install candidates.
One real workflow — bring a live brief, a proposal you're writing, a production schedule, a client deck in progress. Not a hypothetical. The build uses real material.
Typical room: 4–8 people. Smaller is sharper. The workshop scales with the room — we don't fill seats, we surface leverage.
What happens
The half-day runs in four blocks. No slides unless they're yours. No generic AI overview. Every minute is on your agency's workflows, your team, your leverage.
Agency economics of AI
The room gets honest about the pressure: client in-housing, reduced briefs, billable-hours compression. Not fearmongering — a working map of where the leverage is in a creative agency's operating model.
30 min
Live build on a real workflow
The room picks one real workflow from your agency. We build a working agent on it, in your tools, in the room. Not a mock-up. Not a spec. A running system. Everyone watches it ship.
90 min
Leverage mapping
Each person in the room maps their own workflows to the In-Seat AI framework: Direction, Memory, Production, Proof. By the end, the room knows who is ready for a 1:1 install, who needs a month, and who doesn't need one yet.
45 min
The three names
The workshop ends with explicit install recommendations for the MD: three names, three use-cases, three session plans. These are the people whose work changes most, fastest, with the least disruption to the team.
15 min
The prototype
By lunch, you own a working AI agent on a workflow your team uses every week.
The prototype runs in Claude Code or Claude.ai — whichever your team already has access to. It is not a spec document. It is not a slide showing what it could look like. It is a working system, built on real material from your agency, tested in the room before the session ends.
Past first builds: distributor-pitch-deck generator, new-business brief expander, production-schedule builder, scope-of-work drafting agent, client-report generator. What gets built depends on what the room brings.
The four layers the prototype demonstrates
Direction
You write the brief. The agent ships the work.
Memory
Context that survives between sessions.
Production
The agents that do the repeatable work.
Proof
The receipt that shows it's working.
The 24 hours after
The workshop is not the end. It's the diagnostic.
Within 24 hours of the workshop, every attendee receives a personalised 1:1 recommendation — a one-page note covering exactly what a Role Install with them would look like: which workflow, which agent, which three sessions.
The MD gets a short summary of the three strongest install candidates and the sequencing logic behind the recommendation.
The 14-day credit window opens from the workshop date. Role Installs booked within 14 days apply the full workshop fee as credit.
How the credit works
How the ladder works
The workshop fee credits in full toward Role Installs booked within 14 days.
Book the Founding Workshop. Run the half-day. Within 14 days, book any of the recommended 1:1 installs — and the entire workshop fee applies as credit toward the Role Install price.
The window exists because the workshop is the diagnostic, not the product. The product is the installed agent in the operator’s hands. If the workshop does its job — surfaces the right people, surfaces the right workflows — fourteen days is generous enough to move into the install.
ROLE INSTALL (SINGLE) — from R20,000 · 1 agent, 3 sessions
TRANSFORMATION PACK — quoted on scope · 2–3 agents, 6 sessions
The 14-day window does not flex. It is the commercial promise that keeps the workshop from being a one-day event with no follow-through.
The proof
Case Study #1
13 May 2026 · Cape Town
Mack Brands — Founding Workshop
Mack Brands distributes and markets premium spirits across South Africa. The Founding Workshop ran on-site in Cape Town. Four senior operators in the room. By lunch: one working distributor-pitch-deck agent built on real Mack material, and three install candidates confirmed.
"The workshop didn't show us what AI could do. It made us build something with it — on our actual work. That's a completely different experience."
Em Kendrick Cox · Mack Brands · Cape Town
Common questions
Do we need to prepare anything?
Bring a real workflow. A brief you're currently writing, a proposal in progress, a production schedule, a reporting format your team hates. The build runs on real material — the less hypothetical, the better.
What tools do we need?
Claude.ai or Claude Code. Free or paid — either works for the workshop. If your team hasn't used Claude before, that's fine. You'll be using it by the end of the session.
What if no installs come out of the workshop?
That’s a valid outcome. The workshop surfaces what’s there. If the room does not produce strong install candidates, I will say so — and the 14-day credit window will not apply. The discovery call before the workshop is where we check fit.
Can the workshop run remotely?
On-site is strongly preferred — the live build lands better in a room. Remote is possible for agencies outside Cape Town, but expect a slightly shorter session and more pre-work on your team's part.
How do we book?
Discovery call first — 30 minutes, three questions. If there's a fit, the workshop gets scheduled. If not, I'll say so on the call.
What is the discovery call?
Three questions: what's the team shape, what are the obvious bottlenecks, and who's the person whose job AI could change the most this quarter. 30 minutes. No deck.
The discovery call is
30 minutes.
Three questions.
If there's a fit for the Founding Workshop, we book it on the call. If not, I'll tell you on the call. No deck, no proposal, no follow-up sequence.
Bespoke AI systems for creative production. Installed into your employees. Not in place of them.