How to choose an AI consultant: the questions to ask before you decide

A good AI consultant starts with your operations, not with a tool. They assess where you stand today, understand the data in your systems, prioritize the use cases with the biggest impact and can build AI into the systems you already have, with guidelines, ownership and follow-up. This guide covers what an AI consultant does, the questions to ask before you choose and the warning signs that you risk ending up with yet another pilot that never reaches everyday work.

Short answer

• Choose a consultant who starts with the current state. They should want to understand your processes, data and systems before proposing solutions.
• Choose a consultant who understands the data in your systems. AI is never better than the data it builds on, so the consultant needs to understand how the information is structured and connected in your systems.
• Be prepared that the first step may not be AI. With outdated ways of working and systems, a workshop and a more efficient system are often the right start, with automation and AI after that.
• Require that they can build, not just recommend. The value only appears when AI is built into the work, not in a report.
• Ask for real projects. What was built, in which systems and what was the result?
• Make sure you own what gets built, and that there is a plan for ongoing management.

What does an AI consultant do?

An AI consultant helps a company understand where AI creates value and makes it happen. In practice, the engagement spans four parts, and different consultants are strong in different parts:

• Current-state assessment and prioritization. Where does time go, what data exists and which use cases deliver the most impact relative to the risk?
• Guidelines and governance. What may AI be used for, which data may be used and who owns each solution?
• Implementation. AI agents, automations and integrations built into the systems you already use, such as monday.com, Visma Net or Fortnox.
• Adoption and ongoing management. Training, follow-up on the impact and adjustments as the business changes.

Some AI consultants only do strategy, others only technology. For most mid-sized companies, it is an advantage if the same partner can go all the way, from assessment to production, so nothing falls through the cracks.

The first step isn't always AI

To be candid: a lot depends on where you stand today. An organization with outdated ways of working, manual routines and systems that don't talk to each other rarely benefits from starting with AI. It is often more effective to take it in stages, where each stage builds on the previous one.

For an organization that already has modern systems and well-organized data, things can move much faster, and AI can sometimes be the first step. A good AI consultant tells you honestly which path fits, even when the answer is that AI should wait.

Nine questions to ask before you choose an AI consultant

1. How do you start an engagement?

A good answer is about understanding your business: processes, data, systems and organization. If the answer starts with a tool or a platform, it is a sign that the solution was decided before the problem was understood.

2. Can you show projects you have done?

Ask for concrete examples: which problem, what was built, in which systems and what the result was. One example from us: for SKG Storköksgrossisten, a Swedish foodservice wholesaler, we built AI-driven document scanning, where incoming documents are interpreted automatically and the data lands directly in monday.com. Read the case.

3. Can you build what you recommend?

A recommendation that no one builds rarely becomes reality. Ask whether the consultant can build AI agents, automations and integrations themselves, or whether you will need another vendor for that.

4. Do you understand the data in our systems?

This may be the most important question. AI is never better than the data it builds on. An AI agent that creates orders, posts invoices or updates customer records has to understand how the information is structured in the system: which registers and fields exist, how they connect, what is required and where the data is incomplete or duplicated. In Visma Net, that can mean the chart of accounts, subaccounts and the customer and supplier registers. In monday.com, it means how boards, columns and connections between boards are set up.

A consultant who only knows the AI technology but not your systems can easily build something that works in a demo but breaks on real data. Ask how the consultant reviews your data before anything is built, and which systems they know in depth. If you use monday.com, Visma Net or another business system, it is a big advantage if the consultant has worked with it for a long time, not just with AI.

5. Can you explain the difference between data-driven and AI-driven?

A good AI consultant should be able to explain the difference simply, and know when you need which. Data-driven means that decisions and work are guided by data: reports, dashboards and KPIs that people interpret and act on. AI-driven means that AI uses the data to interpret, predict or perform work on its own, such as reading documents, sorting tickets or suggesting the next step.

Data-driven is often the foundation for AI-driven. Without structured, reliable data in your systems, AI works poorly, and many problems are better solved with the right data and a good report than with AI. A consultant who can say "you don't need AI for this" is usually one you can trust.

6. How do you handle data, security and regulations?

Ask where your data is stored, who has access, whether it is used to train AI models and how you comply with GDPR. A good consultant raises this on their own, before anything is built.

7. How will we know if it pays off?

There should be an answer about what is measured before and after, such as time per ticket, number of errors or lead time. Without measurement, you can't know whether the AI investment delivered anything.

8. Who owns what gets built?

You should own the solutions, the data and preferably the code. Also ask how dependent you will be on the consultant: can you manage the solution yourselves or with someone else if you want to?

9. What happens after launch?

The business changes, and AI solutions need to change with it. Ask how ongoing management, support and further development work, and who does it.

Warning signs

• The solution is ready before the question. The consultant talks about a specific platform or product before understanding what you need.
• Only pilots, nothing in production. Ask how many of the consultant's previous pilots are used in day-to-day operations today.
• No one talks about data or ownership. That is often where AI projects get stuck, and an experienced consultant knows it.

How Straviont works as an AI consultant

We are AI consultants based in Stockholm, and we always start with the current state. Sometimes the right first step is a workshop and a more efficient system, not AI, and then we say so. We assess processes, data, systems and organization, and because we work in monday.com and Visma Net every day, we know the data and structure in them in depth. We prioritize the use cases with the biggest impact and then build AI into the systems you already have, such as monday.com and Visma Net. You own what we build, and we follow up and manage it afterward. Read more about how we work as an AI consultant, or about where to start in How to get started with AI in your business.

Frequently asked questions

What does an AI consultant do?

An AI consultant helps a company assess where it stands today, prioritize where AI creates value and implement it in the business. The engagement can include a current-state assessment, guidelines and governance, implementation of AI agents and automations in existing systems, and adoption and ongoing management.

What should you consider when hiring an AI consultant?

Choose a consultant who starts with your business rather than a tool, who understands the underlying data in your systems, who can show concrete projects with results, who can build what they recommend and who lets you own what gets built. Also ask how data is handled and how the impact is measured.

What is the difference between data-driven and AI-driven?

Data-driven means that decisions and work are guided by data, such as reports and KPIs that people interpret. AI-driven means that AI uses the data to interpret, predict or perform work on its own. Data-driven is often the foundation for AI-driven, because AI needs structured, reliable data to work.

What is the difference between an AI consultant and an AI agency?

The line is blurry. An AI consultant usually means advisory and implementation in an existing company, while an agency often builds products or campaigns. What matters is not what they call themselves but whether they can take you from the current state to solutions that get used.

Do we need an AI consultant?

Not always. Many companies can assess the current state and test the first use cases on their own. An AI consultant adds the most value when you have tested but not gotten anything into production, when AI needs to be built into several systems or when no one internally has time to drive the work.

Where can you find an AI consultant?

There are many AI consultants, from large consulting firms to specialized agencies. Use the questions in this guide to compare them. At Straviont, several AI consultants work from our office in Stockholm, on projects in Sweden and across Europe.

Next steps

Write down two or three workflows where you think AI could help, and bring them to the first meeting with the consultants you are comparing. How they ask questions about those workflows says a lot about how they work. If you want to talk to us, you can book a free mapping session.