You leave with: A readiness score, a signed job description, and either an AI employee running in your systems in Claude under a named manager or, if the score is too low, a remediation plan your managers own.
The hiring process for a worker that is not a person: score whether your department can employ one, write its job description, then onboard it under named supervision. Two days plus thirty days of follow-through, built for the person who will answer for the output.
2 days plus 30-day follow-through. Up to 20 seats. Priced per cohort, never per seat. Travel and venue excluded. The scoping call fixes the dates and the room; the price is the price.
Scoped to your organisation, so it starts with a short brief, not a checkout.
60 percent of the fee follows proof. A tax invoice is issued at each milestone: 40 percent on signature, 40 percent when your named sponsor accepts the outcome, 20 percent at close-out. NDA on request.
The rule behind the score: assist, delegate or autonomous, task by task
12 assets across the program. Every one is a document, a standard or a running system your organisation keeps.
4 of 12 assets built
9 of 12 assets built
12 of 12 assets built
Bring the real work: your processes, your data, your decisions. Nobody builds a demo.
Where a card shows a number, its source is on the card. Clients appear by name only where they agreed to be named; external cases cite their public source. The corporate track opened in September 2026, so this is what we run and what our builder cohorts shipped, not a corporate reference yet. That is why 60 percent of a corporate fee follows proof: 40 percent when your sponsor accepts the outcome, 20 percent at close-out.
22 years inside Google, Cisco, Avaya and Uniphore, 16 of them working with AI, from speech and conversational systems to the agentic systems his team ships today. Founder and CEO of SMOrchestra.ai and founder of Entrepreneurs Oasis, an SMOrchestra company. MBA, University of Leicester; BEng Computer Engineering, University of Jordan. He teaches the room himself.
I spent 14 of those 16 years building AI before AI had a brain. The moment it got one, I rebuilt everything around it.
SMOrchestra is a registered member of the Claude Partner Network (see the proof).
Name the program and your organisation, or book 15 minutes on the academy page. No deck needed.
A short brief: what the program builds for you, the seats, the price, the dates we can hold.
Room, language, on site or remote, and the number, fixed. Then a tax invoice, 40 percent on signature.
The people who answer for output quality: department heads and team managers in operations, HR, finance, service and delivery. Up to twenty per cohort. If your board has not set an oversight model yet, C1 comes first and takes half a day.
Two live days, the 30-day follow-through, the materials and recordings, and, where the readiness score allows, one AI employee in delegate mode left running under a signed oversight sheet. Travel and venue are excluded. Pricing is per cohort, never per seat, and the scoping call fixes the dates before you sign.
It produces the part of it that is usually missing: a readiness score, a signed job description and a decision rule per task type that a manager can actually apply, with the approval gates and the kill switch wired in Claude rather than described in a document.
The two days are consecutive and the 30-day follow-through assumes both. Tell us at scoping: a manager who cannot make both days is better placed in the next cohort than half in this one. Recordings support the room; they do not replace it.
Yes, as two cohorts, not one larger room. Twenty is where every manager's job description still gets reviewed in front of the group; above that, the method turns into a lecture.
A 12-factor readiness score, taken on Day 1. 0 to 17: not ready, and the program tells you what to fix first. 18 to 27: hire in delegate mode (it does the work, a person approves each run). 28 to 36: ready for more autonomy; the program still onboards in delegate mode, and the 30-day follow-through decides whether it moves up.
The leash: approvals, logging, evaluation, a kill switch, and personal data handled under PDPL. Every employee has a named supervisor on a signed oversight sheet. No name, no autonomy.
Into work nobody on the team does today, never into someone's job. We start from four questions: what does your team keep postponing, what do you wish someone did, where do you pay an outside firm, and where is AI simply better? Then four checks before it runs further: is it reversible, can it be checked, is it bounded, and who signs? One no, and it stays in delegate mode.
Keep Copilot. It drafts inside Office. These programs build the operating system around the work: the brain, the skills, the tool keys and the gate. We teach on Claude because that is where we build and run 100+ operating systems ourselves. The decision rule, the job description and the oversight sheet are yours whatever model you run.
You do, directly with Anthropic; they are not in our fee. Each participant needs a seat on your organisation's Claude Team or Enterprise plan. Anthropic lists Team at $25 a seat a month, or $20 billed annually (claude.com/pricing); Enterprise is quoted by Anthropic. The setup checklist sent at signature names the plan and the number of seats.
Not on Claude Team or Enterprise: Anthropic's commercial terms exclude your inputs and outputs from model training by default. Residency and data-protection rules (PDPL, your sector's) are reviewed with your IT and legal at scoping, before any of your data enters a session.
Start with a one-day audit, Revenue & GTM or AI. $1,500 each, credited in full against any build signed within 60 days.