Who is AI consulting for?
Leadership has asked for "AI", but nobody can say which job to start with. Or a pilot looked promising and then stalled because no one could describe the next step. This is where AI consulting earns its fee: we pick the right problem first, then describe the simplest solution that will solve it.
We recommend a consulting engagement most often when:
- You have a long list of ideas. You need one set of criteria to separate the ideas that will create value from the ones that are merely interesting.
- A pilot has stalled. The model seems to work, but production, integration and ownership questions are still open.
- You do not know the state of your data. Documents, records and systems sit in different departments, and it is unclear what is usable.
- You need to compare vendor proposals. Proposals arrive with different scopes and assumptions, and you want your own needs in writing before you compare them.
- Privacy and KVKK questions are holding up the decision. No one will sign until it is clear which data may be processed where.
What does the engagement deliver?
The output is not a slide deck. It is a set of written documents that stays with the person who has to make the decision. What will be delivered is written into the scope document before work starts.
Use-case inventory and prioritisation
Every candidate is scored on the same criteria: expected benefit, data readiness, risk and integration effort. Jobs that are not worth AI stay on the list with the reason why.
Budget goes to the right problem, not to experiments.
Data and systems assessment
One table shows where each data source lives, how current and consistent it is, who can access it and whether it contains personal data.
The riskiest assumption in the project becomes visible early.
Solution approach and deployment model
For each priority use case we recommend an approach (a cited assistant, workflow automation, an AI agent, computer vision or a classical model) and where the data should be processed.
Technology follows the need.
Pilot scope and success criteria
For the first use case we write down the pilot boundary, the data it will use, the acceptance criteria and how they will be measured.
Everyone knows what the pilot will be judged on.
Risk, compliance and human-approval notes
Personal data, logging, permissions and steps that need human approval are flagged per use case; questions that need legal review are listed separately.
Compliance is discussed at the start, not at the end.
Roadmap and cost drivers
Phases, dependencies and the items that drive cost are written out separately, so you can hand the same document to any team you ask for a proposal.
You compare proposals against one scope.
Let's use a short call to work out which use case is worth testing first.
Book a free intro callHow the engagement works
You decide whether to continue at the end of every step. The schedule is written into the scope document with dates.
- Introductory call
We listen to your goals, candidate processes and constraints. The first call is free and carries no commitment; if you prefer, we sign a non-disclosure agreement first.
- Scope document and proposal
We write down the processes, teams, systems and deliverables in scope. The price is based on this document.
- Interviews and data review
We talk to process owners, IT and data owners, and review sample data in an environment your organisation chooses.
- Quick validation
Where useful, we test the most critical assumption on sample data. The goal is not to build a product but to back the decision with evidence.
- Report and decision meeting
We walk the decision-makers through the findings, the priority order and the pilot recommendation.
- Next step
You can run the pilot with us, implement the plan with another team, or decide that this job does not need AI.
What does fixed scope mean?
What usually inflates a consulting budget is an open-ended scope. That is why we write the boundaries of the engagement down before it starts.
| In scope | Out of scope (separate proposal) |
|---|---|
| Processes and teams named in the scope document | Building and integrating a production system |
| Systems and data sources named in the scope document | Large-scale data cleaning and labelling |
| Limited validation tests on sample data | Hardware procurement and installation |
| Written report, priority list and pilot scope | Legal opinions; flagged separately when needed |
If an important need outside the scope comes up during the work, we record it as a separate item in the report. We do not widen the scope without your approval.
How is it priced?
The engagement is quoted in writing once the scope document is clear, before work starts. For software projects, the first meeting and the discovery/feasibility phase remain free; this consulting is a separate, paid engagement. The main factors are:
- How many processes and teams are in scope.
- How many systems and data sources we need to examine, and how easy they are to access.
- Whether validation tests on sample data are needed.
- Whether on-site work is needed, and where the interviews take place.
- How deep the security, personal-data and documentation requirements go.
Work that follows an engagement is quoted stage by stage, in writing, as the scope becomes clear; after go-live we agree a monthly support arrangement. The report states which one fits and why. Our article on what an AI project actually costs explains the cost drivers in detail.
Advice from the team that builds it
The consulting report is written by the same engineering team that builds its own AI products and has delivered at public-sector scale, so the recommendations have already passed a feasibility filter.
- We build our own product. Our legal AI product Caseon brings document review and contract preparation into one workspace where every answer shows its source.
- We know public-sector delivery. We built the SYD HUB platform for the Turkish Ministry of Environment, Urbanisation and Climate Change; it is live on the ministry's own domain.
- We build operations software. VinçTakip puts a crane fleet's location, operators, jobs and status on one screen.
- We tell you when it is not worth it. If a job is not worth solving with AI, the report says so plainly; if simpler software or a process change is enough, that is what we recommend.
The work that follows an engagement usually becomes one of these services: RAG and enterprise semantic search, AI agent development, visual quality inspection, data engineering or custom software development. For deployment and procurement in government bodies, see AI for public institutions.
If your idea list is ready, we can use the first call to pick the one worth an engagement.
Request a call