Definition
What should an AI consultant leave behind?
An AI consultant should leave behind operational independence: a working system in your accounts and the means to change it safely.
By the Foxnut team · Updated
What should an AI consultant leave behind?
An AI consultant should leave behind operational independence: a working system running in your own accounts, plus enough of the build’s reasoning and test material that your team can change the system without breaking it, and can tell when they have. That material changes hands as a defined set of handover artefacts, listed in the artefacts of an AI handover. The test of the work is that the consultant can leave.
Foxnut AI is a senior-only practice in Bengaluru and Paris, and this page describes how we hand over our own work. There’s a link to talk to us at the end. Let’s first go through what should actually change hands.
What is the build, own, operate, transfer model?
Build, own, operate, transfer (BOOT) is the model we typically use to hand a system over. We do not sell AI strategy documents. We build AI workflows into commercial operations until they produce measurable outcomes, and then we hand them over.
In a pure consulting engagement, that means scoping the project with your team, structuring it correctly and implementing it together. Where we also own implementation, we build end to end, including guiding you through buying and setting up every tool the system needs.
The table below shows what each stage obliges, and, in the last column, what you actually hold by the end of it.
| Stage | What the studio does | What the client does | What has changed hands by the end of it |
|---|---|---|---|
| Build | Scopes the system with the client’s team and builds it end to end, inside the client’s own accounts and tools | Names a single point of contact to own the system internally | The infrastructure: the system takes shape in accounts the client already controls |
| Own | Runs the system it built on live business workflows and answers for what it produces | Puts real work through the system and judges the output against it | The proof: weeks of real-world use showing the system holds up, not a demo |
| Operate | Assigns a single point of contact on the studio side and keeps the system running through the support period | Puts its internal owner and a supplementary team through thorough training, with modules for specialised roles | The knowledge: documentation, testing, and a team trained to run the system independently |
| Transfer | Completes the documentation handover, then steps back to on-request troubleshooting and short consulting on new requirements | Runs the system alone | Everything. Transfer happens when the system is proven in real use and the team is trained, or when the project timeline lapses, whichever comes earlier |
One system runs that way today. A D2C cosmetics brand selling to women in India, with multiple SKUs and complex requirements, needed its entire supply-chain flow automated. We built a fully automated, end-to-end supply-chain management tool that does the work of a full procurement and supply-chain team, and the client runs it.
The Transfer stage on this engagement was concrete. The tool runs on the client’s own servers and API keys, so no data leaves their system. The client also received a troubleshooting framework, a maintenance cadence its own IT team was trained to run, and training sessions for the supply-chain team, senior leadership and the adjacent parts of the business the tool now feeds. The fuller, dated account of this engagement is written down in the founder’s own words.
Important note: this library will not answer three kinds of question, each for a stated reason. It gives no prescriptive legal or tax answers, because we are not a law firm, so the regulatory pages describe what the rules ask of operators and what operators do about them, and stop there. It carries no named-client case detail, because client permission covers the name and not the project, so every worked example here is anonymised even where the client list is public. And it publishes no vendor rankings or tool listicles, because we build on these tools and sell work around them, so a ranking from us would not be worth reading.
What should an AI consultant not leave behind?
A dependency. The deliverable of an AI engagement with us is operational independence, so we deliberately leave out the things that create return business for the wrong reason: a system only the vendor can safely change, maintenance framed as a retainer, and documentation that describes the system without helping anyone run it.
The engagement mechanics behind that, including how engagements at Foxnut AI run, and when the studio says no, apply to AI work the same as to everything else in Foxnut AI’s consulting practice on AI enablement.
This page covers one part of our practice in depth: what changes hands, and when. For the wider case, meaning which business processes AI actually changes and the real systems we have shipped, see where AI changes a business.
Who should you talk to about handing over an AI system?
Talk to someone who measures the work by the day you stop needing them. A consultant paid by the month has little reason to leave. We have one tier and no account managers, so the person who proposes your system is the person who builds it, and we launched products of our own before we advised anyone on theirs.
Your first step is a free audit call with a senior practitioner. If your supply chain, or anything shaped like it, needs the same treatment the cosmetics brand above got, tell us about it.
Foxnut AI takes on work like this from Bengaluru and Paris. If you want to see how we do it before you talk to anyone, here is how an AI engagement is scoped and priced.