One trigger, one output, one owner. $7,500 per operating system, minimum two, 4 weeks each, no retainer.
A revenue operating system is one link of the revenue chain running itself: one trigger, one output, one named owner, built as a brain, a set of skills, a scoring gate, and a learning loop, wired to the live tools, run in production for a week, then documented and transferred to that owner.
That definition goes in the contract, because per-process pricing invites scope debate and this sentence ends it. The trigger is the event that starts the work. The output is the artifact that leaves the system. The owner is whose name sits on it after we go. A process crossing more than two systems prices as two, decided before you sign.
Most companies systematise the first link and wonder why the rest leaks. That is why contract review and a data pipeline belong on a page called Revenue Operating Systems: revenue compounds after the signature, and those links are the ones nobody instruments.
| Link | What it does | How it leaks when nobody systematises it |
|---|---|---|
| CREATE | Demand, content, campaigns, webinars, training | Publishing depends on whether the founder had a good week. Quality swings with whoever wrote it |
| CLOSE | Proposals, pricing, contracts, approvals | Every deal is custom because nothing is productised. The same paragraph gets rewritten for the tenth time |
| DELIVER | Projects, products, onboarding, client work | Margin leaks after the signature, because nobody scores delivery the way they score pipeline |
| LEARN | Data, reporting, scoring, the lessons loop | Reporting takes three days and is stale on day two. The same mistake gets made in two departments |
Confirmed demoable by the owner on 2026-07-26. Pick the link in your chain that leaks most. That is the first operating system.
Every one of these is demoable on a 15-minute call.
Each role is a seat with a scoped job, toolset, and memory. They talk to each other through a shared 2nd brain, and you talk to them through one surface.
| Component | Owned by your team after transfer |
|---|---|
| Proven framework | Each process encoded as a runnable skill with its inputs, steps, and quality bar written in |
| Best model per task | Cheap models on volume work, strong models on judgement work, routing already decided so cost stays predictable |
| Context | Your institutional knowledge as maintained context files. New joiners read the OS instead of shadowing a senior for six months |
| Results wiring | Connected to the real stack: your document store, ERP or CRM, email, internal channels |
| Learning system | Every correction an operator makes is captured and folded back, so the system improves through use rather than through a rebuild |
| Scoring and validation | An explicit quality gate on every output. Nothing reaches a client or a regulator unscored |
After transfer your operations lead can add a process without calling a vendor.
The most common failure we walk into is a company that bought seats and called it done. Everyone has a licence, everyone prompts differently, and the work comes out inconsistent faster. Nobody can say which output was checked against anything, and whoever got good at it is now the department's bottleneck.
The line between a licence and a capability: the trigger is defined, the steps are written down, the tools are wired, the output is scored before it ships, one person owns the result. A licence produces attempts. A system produces the same output on Friday as on Tuesday, with a different person in the seat.
Why this works in the Gulf when an imported template quietly does not. WhatsApp is a first-class citizen. The AI Tool Admin role treats WhatsApp Business the way it treats email: monitored, orchestrated, logged, auditable. If your customers buy on WhatsApp, your operating system lives on WhatsApp. Dialect-native output: if your buyer in Riyadh closes in Khaleeji, the output arrives in Khaleeji. If Cairo opens in Masri-inflected MSA, that is the register that ships. We do not ship back-translated Arabic that sounds like a telecom IVR. Local-admin-ready: invoices that comply with UAE corporate tax, e-signature flows that hold up in KSA, calendar flows that respect the Gulf weekend, contract templates under ADGM or DIFC framings.
Who signs. The COO, GM, or Head of Operations. In family-owned Gulf groups the owner signs and the GM scopes. Where the pain is bid production, the commercial director signs. Company shape: 20 to 500 people with one function drowning in repeat work: operations, finance back-office, HR, client service, proposals, or reporting.
Who this is not for. Skip this if you are pre-revenue and need product-market fit rather than process; if you want a managed service and would rather not learn the system; if you already have an operations team of three or more running documented processes; if you want the cheapest option, because we are not it; or if you cannot give a named owner four hours a week for four weeks. To build these yourself, the routing is the AI Operating Systems cohort, $1,450 for 6 weeks, where you ship one working OS as the capstone.
| Phase | Week | What happens |
|---|---|---|
| Map | 1 | Process mapped and instrumented. Trigger, output, and owner named in writing. Repeat cost measured, so the OS pays for itself on a number |
| Build | 2 to 3 | The OS built: brain, skills, templates, scoring. Wired to the live tools, up to 2 integrations, one approval gate placed where a human decision matters |
| Operate and transfer | 4 | Run live in production for a week with the owner in the seat. Documented, SOP written, owner trained, handed over |
Transfer sits inside the live week deliberately. The owner learns the system while it runs, with us in the room. A handover that starts after operation ends is a document drop.
| What you buy | Price | Window |
|---|---|---|
| One operating system | $7,500 | 4 weeks |
| The minimum engagement, two operating systems | $15,000 | 4 weeks each |
| Department bundle, 3 to 5 operating systems | $22,000 | scoped |
| Company-wide, multi-department | from $55,000 | scoped |
The minimum of two exists for a delivery reason: the brain, the context files, and the wiring built for the first OS make the second one cheap. A single OS pays for that setup and never uses it twice.
| What you do today | Annual cost | What you own after 12 months |
|---|---|---|
| Hire an operations manager for the function | Not published until the fully loaded figure is verified | Nothing transferable. The context leaves when the person does |
| Add a per-seat tool for another fragment of the workflow | your current renewal | A login and one more integration to maintain |
| Two operating systems, owned | $15,000 once | The brain files, the skills, the scoring gate, the SOP, and an owner who can add the third |
A number invented to look reasonable is worse than an empty cell.
| What is included, per operating system | What is not included, and what triggers a change request |
|---|---|
| Up to 2 tool integrations; 1 approval gate; 1 owner trained; documentation and the SOP; the learning loop, wired and demonstrated; the process mapped and its repeat cost measured; one week of live production running before handover. | Not included: tool subscriptions, which sit on your accounts, a second owner's training, processes crossing more than two systems, which price as two, and custom software development, which is the MicroSaaS build at $22,000. Change requests: adding a trigger or an output to an in-flight process, or a third integration. |
One named owner per operating system, with authority over the output. Four hours a week from that owner across the four weeks. Tool access in week 1. Decisions inside 48 hours at the approval gate. Two hours in week 1 from the person who does the work today, because the process lives in their hands.
We name the trigger, the output, and the owner, and price it before you commit.
Not sure which link leaks? The 1-Day AI Audit is $1,500, delivered in 2 business days, names the five highest-value processes to turn into operating systems ranked by hours returned, and credits 100% against this within 60 days.
| Question | Answer |
|---|---|
| What is an AI operating system? | One function running itself: a trigger, a defined output, a named owner, plus the four parts that make it repeatable, a brain holding the context, skills that do the work, a scoring gate on every output, and a learning loop that folds corrections back in. A licence to a chat tool is not an operating system, because nothing in it is defined, scored, or owned. |
| Why is the minimum two operating systems? | The first one pays for the brain, the context files, and the tool wiring. The second one uses all three and costs the same $7,500 while delivering more. Buying one means paying setup cost for a single use. |
| What exactly counts as one process? | One trigger, one output, one owner. A quote request arriving, a priced proposal leaving, the commercial director owning it, is one. If the work crosses more than two systems, it is two, and we say so before you sign rather than in week 3. |
| Does this work if we already run HubSpot, Pipedrive, or GoHighLevel? | Yes. This is stack-agnostic, and the AI Tool Admin role is designed to operate your existing tools rather than replace them. On GoHighLevel we wire into GHL natively. On HubSpot we wire into HubSpot. On Notion plus a spreadsheet we will still build the brain, and we will usually recommend a real CRM inside the first two weeks. |
| What if our data is a mess? | It will be. Every founder we have worked with has a CRM that is 40% stale, 20% duplicated, and 10% wrong. Cleanup for the process in scope happens in the Map week. We do not install on top of a graveyard. |
| How is this different from hiring an operations manager? | A manager is a person: hard to hire in the Gulf, harder to keep, and impossible to replace without a three-month handover. An operating system is installed on your infrastructure, documented, scored, and transferable. It does not get poached and it does not quit in month 9. It is also not a substitute for a real operations leader once you pass roughly 40 people; it is what lets you wait until you actually need one. |
| What does ownership look like at transfer? | Everything runs under your accounts. Your Claude subscription, your CRM licence, your domain, your integrations, your data. We do not host and we do not hold keys. If SMOrchestra disappeared tomorrow, nothing in your business would stop running. |
| How much does the owner have to learn? | Less than you think, and the training is inside the price. The owner sits in the seat during the live week, not after it, which is why one week of production running beats four weeks of documentation. |
| What happens after the four weeks? | You run it. If you want a further process, that is another OS at $7,500. If you want ongoing optimisation, that is a named expansion module with an end date, $6,000 to $15,000 depending on scope. No retainer is sold as the default. |
| Can we start with a single operating system? | Not at this price. The honest alternatives are the 1-Day AI Audit at $1,500 to find out which processes are worth it, or the AI Operating Systems cohort at $1,450, where your own team ships one. |
Up: SMOrchestra, the three pillars. Across: AI-Native Revenue MicroSaaS. Down: Book 15 minutes.