Engagement model
How AI automation projects are scoped and delivered
Four stages, each with a defined deliverable and a fixed price agreed before it starts. You can stop after any one of them, and everything produced up to that point is yours.
The stages
What happens, in order
Most engagements run through all four. Some stop at stage two because the assessment shows the work is not worth doing, which is a legitimate and useful outcome. Stage two can also be bought on its own as a workflow automation audit, with no obligation to build anything afterwards.
| Stage | What we do | What you get | Typical duration |
|---|---|---|---|
| 1. Consultation | A call about the process that is hurting and the systems involved | An honest read on whether this is worth automating | Under an hour, free |
| 2. Assessment | Interviews, process mapping, technical feasibility and scoring | A written, ranked and costed set of candidates | Two to three weeks |
| 3. Build | Integration, workflow or agent development, testing against real data | A working system in your accounts, plus documentation | Two to six weeks per workflow |
| 4. Handover | Runbook, working session, monitoring and alerting in place | Full ownership and the ability to maintain it yourself | Within the build stage |
The first conversation costs nothing and carries no obligation. If the honest answer is that your problem is a process change rather than software, or that an off-the-shelf product already solves it, we will tell you on that call. That is a cheaper outcome for you and a better use of our time than a project that should not have started.
Pricing
What a project costs, and why we quote per stage
Every stage is quoted as a fixed price against a written scope, and you approve that number before the stage begins. If the scope changes mid-flight, we re-quote rather than absorbing it silently or billing the difference by surprise.
Cost is driven by integration surface far more than by anything to do with AI. A workflow inside a single system with clean rules is a small engagement. The same logical task spread across four systems — one of which has no usable API and needs a workaround — can be several times the size, even though the automation looks identical from the outside. The exception count matters just as much: the cases that occur two per cent of the time frequently account for most of the build.
This is also why we ask which systems are involved before quoting anything. A platform with a documented, well-maintained interface such as the Microsoft Graph API is a known quantity; a legacy system whose only export is a nightly file is not, and pretending otherwise at quoting time simply moves the problem to week three.
Because the number depends on those specifics, we quote after the consultation rather than publishing a price list that would be wrong for most readers. The consultation is free precisely so that getting to a real number costs you nothing. If you want a sense of scale before booking, describing the systems involved and the rough volume in the form is usually enough for us to give you a range by reply.
Who this suits, and who it does not
This model works well for companies with a handful of processes that are demonstrably consuming hours, where the systems involved are mainstream and the decision-maker is reachable. It works badly for enterprise-wide transformation programmes with formal procurement cycles and a dozen stakeholders — that is a genuinely different kind of engagement, and a larger consultancy is a better fit. We would rather say that early than three months in.
You can read more about who is behind the work on the about page, or start with a scoped AI strategy assessment if you are not yet sure which process to tackle first. If you are weighing us against other firms, our guide on how to choose an AI automation consultant sets out the commercial questions worth putting to all of us. Where the work is a one-time port rather than an ongoing build, the stages are the same shape but the end is visible from the start — our Mailchimp to Klaviyo migration service page is a worked example of what a fixed-scope stage that finishes actually contains.
Questions
Practical questions about working together
How is a project priced?
By fixed price per stage, agreed before that stage starts. You approve a defined scope and a number, and that number does not move unless you change the scope in writing. We avoid open-ended hourly billing on build work because it puts the risk of a bad estimate on the client rather than on us.
What makes one build cost more than another?
Three things, in order: how many systems have to be integrated, whether those systems have documented APIs, and how many exceptions the process contains. The AI portion is rarely the expensive part. A single-system workflow with clean rules is a small piece of work; the same task across four systems, one of which has no API, is a different project.
What do you need from us during a build?
Access to the relevant systems, one person who can answer questions about how the process really works, and someone empowered to decide on edge cases. The last one matters most. Projects stall far more often on an unanswered question about what should happen in an unusual case than on anything technical.
What happens after handover?
You own the work outright and can maintain it yourself, since it lives in your accounts on mainstream platforms and arrives documented. Where clients prefer us to keep an eye on it, we agree a small ongoing arrangement separately — but it is genuinely optional, and it is not a condition of the build.
Free consultation
Tell us what is costing you the most time
Describe the process. You get a straight read on whether it is worth automating, roughly what it would take, and what we would tackle first — on a call that costs nothing.
- Under an hour, free, and no obligation
- Fixed price per stage, agreed before any work starts
- You own everything we build, documented and in your accounts
- If you do not need us, we will say so on that call
The form is the fastest way to reach us — we reply from a real person, usually within one business day.