What an AI business app actually costs

"Contact us for a quote" is the right answer when a project's scope is genuinely unknown — and for an AI business app, scope is the whole game. Two businesses asking for what sounds like the same system can be five times apart in price once you know how many systems it touches and how much of the logic is actually custom. A generic band would be honest about neither of them.

What we publish instead is the shape of the ladder: four tiers, what separates them, and one number — the only one that doesn't depend on your workflow.

The four tiers

Tier	Price	What it buys
Workflow Audit	Free, credited against the build if it proceeds	One workflow, mapped — the step most likely to conclude you should not build yet
Prototype Sprint	Contact for pricing	A working slice you can put in front of one real user
Pilot Build	Contact for pricing	A system a real team uses in production, with a run log
Production Build	Contact for pricing	The scoped, hardened version, for volume and reliability beyond what a pilot needs

Ongoing support is also priced at handover, for the same reason — it depends on what was built and how it needs to be kept running.

What separates the tiers

The difference between a prototype, a pilot, and a production build is scope and risk — not ambition. A prototype answers "does this work at all, for one real case." A pilot answers "does a real team actually use this, and what does the run log show over real weeks." A production build is what you commission once the pilot has already answered those questions and the workflow needs to hold up at volume.

Most enquiries should start at the smallest tier that answers the actual question in front of you. Starting at Production Build before a Prototype Sprint has ever been in front of a real user is usually buying certainty you haven't earned yet.

Why the audit is free, and the rest isn't quoted here

The audit is the step most likely to end a project honestly — to conclude a workflow does not need building yet, or needs a smaller thing than was originally asked for. Charging for the step most likely to tell you not to spend money creates an incentive to find a reason to proceed. Free, credited against the build if one happens, removes that incentive, and it's the one number that's true regardless of whose workflow we're looking at.

Everything past the audit genuinely depends on what the audit finds — how many systems the workflow touches, how much of the logic is custom, whether inference needs to be self-hosted. A published band for those tiers would either be wide enough to be useless or narrow enough to be wrong for most of the people reading it. The audit exists precisely so the number you get is real, not a guess dressed up as a range.

What you own at handover

At handover: the code, the data exports, the credentials, and the documentation. Our access is removed. This is not a special arrangement negotiated per client — it is the default on every engagement, because a vendor with permanent access to a system it no longer maintains is a liability we don't want to be for anyone.

Engagement terms otherwise: an NDA before scoping if you want one, a written SOW or proposal per project, an advance up front followed by milestone payments.

Where clients are

Working hours are shifted per project to overlap the client's working day, for both US and UAE clients — there is no fixed support window; it is set to match wherever the work is happening. Payment runs through Wise or Payoneer, not raw SWIFT, and not Stripe.

Self-hosted inference

For anyone whose data cannot leave infrastructure they control, self-hosted inference — the model running on infrastructure the client owns rather than a third-party platform — is a real, offerable option, priced separately from the build itself.

Honest limits

• Arabic language support is model-level only, and has not been tested against Gulf Arabic specifically. It can be scoped as a pilot deliverable with accuracy testing up front and an English fallback — it is not, today, a proven strength, and we would rather say that plainly than let a demo imply otherwise.
• "Contact for pricing" is not a stall tactic. The audit is the fast path to an actual number — most scoping calls end with a one-page quote within a day or two of the audit, not a long sales process.

← Back to AI for owner-led service businesses