# How to choose an AI consultancy in Finland

> Six questions for the first meeting show who will build, who will stay and what comes after the first project.

**Published:** 2026-11-20  
**Author:** [Niina Tuikka](https://www.renessai.com/team/niina-tuikka)  
**Topics:** AI Strategy, Safe & Ethical AI Advisory

**Question answered:** How to choose an AI consultancy in Finland

**Related questions:**
- What should I ask an AI consultancy before hiring them?
- Who builds the solution after an AI strategy project?
- What happens when the first AI project ends?
- What is an AI operating model?
- How should an AI consultancy handle the EU AI Act?
- What kinds of AI consultancies are there in Finland?

---

By now most consultancies in Finland have added AI transformation to their services. Some have the know-how, others do not. This post is not about who is good and who is not. It is about what to look for in a partner who will still be there in the months after the first project ends.

Things move so fast now that it is hard to judge whether a person or a team is right for you. Helping you decide what to do and how to do it, and then actually doing it, takes several kinds of skill at once. Below are six questions that tell you how a consultancy will handle everything around the technology.

## Why AI skills are no longer enough

Two years ago, finding a consultancy that understood large language models was a reasonable way to make a shortlist. Today nearly every consultancy, IT house and global advisory firm working in Finland offers generative AI consulting, and most can build a working prototype in days or weeks.

That is good news for buyers, but it means technical skill, while still essential, rarely separates one serious candidate from another. What is rarer is knowing when a problem calls for something other than a language model: a forecasting model, mathematical optimisation for planning and scheduling, or even no AI at all.

What separates consultancies shows up later. In our experience, AI is at most half of what transformation work is about. The rest is who decides, who owns the result once the consultants leave, and how real people’s work changes from week to week. A consultancy can excel at the first half and still leave you with a well-documented project and the same way of working as before.

## Six questions for the first meeting

### 1. Who builds it after the strategy work?

During the sales process, you meet senior consultants. Once the project starts, it is often passed to other consultants to deliver. Sometimes one consultancy writes the strategy and another builds it. Each time, something is lost, usually the reasoning behind the decisions.

Ask to meet the people who will write the code and design the architecture, and ask whether they will be in the room when the strategy is shaped. If strategy and building sit with the same team, the strategy tends to stay realistic and the building stays focused on the right goal.

### 2. Can the consultancy tell you about a project that did not work?

A great deal of AI spending has produced little in return. Any consultancy that has done real work has seen projects stall, and the useful ones have learned something specific from it. A consultancy that can only show successes has either not done much or is not telling you everything.

Ask for one project that went badly, what the consultancy would do differently, and what it now tells clients at the start because of that experience. The quality of that answer says a lot about how the consultancy will behave when your project has problems.

### 3. Where will the consultants actually work?

Real change happens inside the teams whose daily work has to change. Their leaders rarely have time to lead that change on top of everything else. A consultant who only turns up for steering meetings sees a tidied-up version of your organisation, with fewer moving parts and fewer tangles than the team lives with every day.

Ask how much of the work happens alongside your people, in their tools and their meetings, and how much happens in the consultancy’s own office. If most of it happens beside your people, you are buying help; if most of it happens in the consultancy’s office, you are buying advice.

### 4. What happens when the first project ends?

The end of the first project is where most AI projects stop without results. What comes after is the hard part.

The result of the project can take many forms: a pilot, an AI strategy, an AI operating model, an AI agent or a new way of running a process. Whatever the result, it only makes a difference once your organisation can carry it on without the consultants.

This is why a good consultancy also looks at your [AI operating model](/clientcases/sweco-ai-playbook). Your business strategy decides what to do with AI. Your AI operating model decides how the organisation does it well, again and again. It covers who owns AI work and how it is governed, which capabilities, roles and skills you need, how processes change, how solutions are maintained over their lifetime, and how the change is led. Without it, even a successful first project tends to stay a one-off.

Ask how the consultancy plans the step from the first project to everyday use, who owns the result on your side, and how long the consultancy expects to stay involved. Be wary of an offer that stops at a successful first project, and equally wary of one that charges for headcount by the month with no planned end. Finally, agree before the project starts on how its success will be measured.

### 5. How does the consultancy use AI in its own company?

A consultancy that runs its own work on AI can show you what that looks like: how its knowledge is kept, how its people use agents in their daily work, what went wrong during the building process. An example like this is worth more than a reference slide, because it comes from living with the consequences.

Ask the consultancy to show you, not only tell you.

### 6. How does the consultancy handle governance and the EU AI Act?

Regulation, data protection and risk are easiest to get right when they are [part of the design from the first week](/blog/governance-is-not-the-boring-part-of-ai). Treated as a separate audit at the end, these issues tend to arrive as a list of reasons why the thing you built cannot be used.

Under [the EU AI Act](https://eur-lex.europa.eu/eli/reg/2024/1689/oj), that matters even more. Ask who on the consultancy’s team owns governance, and at what point in a typical project that person is involved.

## Three kinds of consultancy, and when each makes sense

The Finnish market has broadly three kinds of consultancy offering AI work, and each is the right choice for someone.

**Large international advisory and technology firms** bring scale, global reference cases and the ability to staff a programme across several countries at once. They suit organisations running a very large, multi-country change, where coordination across many teams matters more than anything else. The trade-off is usually distance: the people who sell the work and the people who do it are often not the same.

**IT and software consultancies** have added AI to a long record of building systems. They are a sensible choice when you know what you want built and need reliable engineering capacity to build it. They are less often the ones to ask whether the system should be built at all, or what the organisation should look like around the new system.

**Smaller specialist consultancies** tend to combine strategy and building in one small team and work closely with a client’s own people. They suit organisations that want a sparring partner as much as a supplier. The trade-off is capacity: ask how such a consultancy would cover a sudden need for more people, and who you would work with if your main contact left. The strongest of them make up for it by using AI to work efficiently themselves, and by leaving your own people able to carry on with the transformation work themselves.

## How I would answer these for Renessai

So that you can compare, here are Renessai’s short answers to the questions above. Our aim is to make ourselves unnecessary: we help you work out what to do, then teach your own people to lead the work on both the technical and the strategic side, so that you can carry on without us or any other consultancy.

- **Who builds it:** the same team. Renessai covers [strategy, architecture, building and governance](/services), so the people who shape the strategy also build it.
- **Where our people work:** beside the client’s own people, in their tools and meetings. Some of us have taken [interim leadership roles](/services#interim-c-level-services) to see a change through.
- **Our own use of AI:** Renessai runs on a [company brain](/blog/your-agents-are-getting-smarter-your-organisation-is-not), which is a shared knowledge base that AI agents keep up to date and that people check. We keep changing how we work as we learn what AI can do.
- **Governance:** before we started using that company brain, we wrote a data protection impact assessment for it. We work the same way with clients: governance and data protection are part of the design from the first week, so they shape what gets built.

## A last thought

Renessai is one of the consultancies you might be choosing between. I have tried to write the questions I would want a client to ask us, including the ones that are uncomfortable to answer. My aim is that you choose well, even if you choose another consultancy.

I would suggest picturing your organisation a year from now, once the first project is long finished. Then ask each candidate directly: will your consultants still be involved at that point, and what exactly will they be doing?