Question 1

What do we get at the end?

Accepted Answer

A scored use-case shortlist with the reasoning behind each ranking, a data readiness report, a build-versus-buy-versus-API recommendation, a pilot specification with success criteria, a running-cost model and a governance document covering data flows and human oversight. It is written for a mixed audience, so your finance and operations leads can read it without a technical translator.

Question 2

How long does an AI strategy engagement take?

Accepted Answer

Typically two to four weeks from the first workshop to the final documents, depending on how many departments contribute use cases and how quickly we can inspect the underlying data. The free readiness assessment shortens it, because the first conversation starts from your answers instead of general questions.

Question 3

What does it cost?

Accepted Answer

A fixed fee quoted after a free scoping call, sized by the number of use cases in scope and the number of systems whose data we need to examine. Pilot development, if you decide to proceed, is quoted separately once the specification exists. We do not price the strategy work as a discount on a build we have not agreed to yet.

Question 4

Do you also build the AI, or only advise?

Accepted Answer

We build it, which is precisely why the advice is cautious. Running AI in production teaches you what a demo never will: retrieval quality decides accuracy, evaluation sets decide whether you can change anything safely, and the monthly token bill is a real operating cost. You are free to take the strategy to another team — the recommendation would be identical.

Question 5

What if the answer is that we do not need custom AI?

Accepted Answer

That is a common and completely legitimate outcome, and we say it plainly. Many businesses get more value from a well-configured off-the-shelf assistant subscription, better search over existing documents, or simply fixing a process, than from anything custom. Spending a few weeks to reach that conclusion is far cheaper than discovering it eighteen months into a build.

Question 6

Would an off-the-shelf tool do the job instead?

Accepted Answer

Often, yes — for drafting, summarising, translation and general research, a per-seat subscription plus training beats anything bespoke. Custom work earns its place when the assistant must read your own private data, act inside your systems through tool calls, run inside your product for your customers, or operate at a volume where per-seat licensing stops making sense.

Question 7

Our data is messy. Should we wait?

Accepted Answer

Not necessarily, but you should know what you are dealing with before committing. A narrow use case over one reasonably clean data source is a good first pilot, while a company-wide assistant over everything is not. The readiness audit tells you which parts are usable today and which need cleanup, so cleanup becomes a scoped task rather than an open-ended blocker.

Question 8

How much does it cost to run each month?

Accepted Answer

Model usage is billed per token, so the honest answer is a range tied to your volume rather than a flat figure. We estimate from your real query volume and typical response length, then design in the levers that move the bill: routing simple requests to smaller models, caching repeated context, and capping retries. Human review time belongs in the same budget line.

Question 9

We already have a data or IT team. What do you add?

Accepted Answer

An outside read on which use cases are worth defending to the board, and current production experience of what breaks. Internal teams usually know the data better than we ever will; what they rarely have is uninterrupted time to compare options and the freedom to recommend not building. We work with your team, not around it.

Question 10

How do you handle risk and compliance documentation?

Accepted Answer

We document the practical layer: which data leaves your systems and to which provider, retention and training settings on those accounts, where personal data is masked, refusal and escalation rules, and where a human must approve. That gives your legal or compliance advisor concrete facts to assess. We document practice — we do not certify compliance, and you should be sceptical of anyone who claims to.

Question 11

How does this relate to AI Readiness Assessment and AI ROI Analysis?

Accepted Answer

This page is the entry point and the free tool is the first step. AI Readiness Assessment goes deep on data, systems and organisational maturity — whether you can support AI at all. AI ROI Analysis takes a specific use case and models cost against benefit, including ongoing inference spend. Most clients need one of the two, not both.

How can we help?

AI Strategy
FAQ

Helpful
information

Getting Started

Timeline

Pricing

Support

Still have questions?

AI StrategyFAQ

Helpfulinformation