INNATE Almost every consultancy now says it does AI. Some are strategy firms that have added a slide. Some are development shops that will build whatever you ask for. A few can do both: tell you what is worth building, then build it properly.
Choosing the wrong one is expensive. You lose budget, but you also lose months and internal confidence, which makes the next attempt harder. These are the questions we would ask if we were on your side of the table.
1. Do they start with your business or with a tool?
Listen to the first conversation. If the talk moves quickly to models, platforms or a product they resell, the recommendation was decided before they met you. A good partner asks where time is lost, where revenue leaks and which decisions are slow. The technology should be the last decision, not the first.
2. Can they show work that is live, not just demos?
Prototypes are easy to make look impressive. Ask to see systems that real teams use every day, and ask what happened after launch: who uses it, how often, and what changed. If every example is a mock-up or a slide, treat that as a signal.
3. How do they decide what to build first?
There should be a clear method for ranking opportunities by commercial impact, feasibility and speed. You want a short list of priorities with a projected return for each, not a long list of ideas. Ask what they would not build, and why.
4. Who actually does the work?
Many engagements are sold by senior people and delivered by someone you never met. Ask who leads the project, who builds it, and how much of their time you will get. Strategy and engineering should be connected, so what gets built matches what was agreed.
5. Will it work with the tools you already use?
Most businesses do not need to replace their CRM, ERP or email to benefit from AI. A strong partner builds around your existing stack and explains exactly how data will move between systems. Be cautious of proposals that start with a migration.
6. What happens after launch?
AI systems need monitoring, small fixes and training for the people using them. Ask what documentation you will receive, how your team will be trained, and whether you can run the system yourselves. The goal should be independence, not permanent reliance.
7. How will success be measured?
Agree the measure before anything is built: hours saved, faster response times, conversion rates or cost per order. If a proposal does not name how value will be tracked, it is hard to know whether the project worked.
8. How do they handle your data?
Ask where data is stored, which third-party models are used, who can access what, and how sensitive information is protected. You should get clear, specific answers, not reassurance.
A good AI partner will happily tell you what not to build. That is often the most valuable advice you get.
A simple way to test a partner.
Ask for a short, fixed-fee diagnostic before committing to a large build. It shows how they think, how they work with your team, and whether their recommendations are grounded in your numbers. At Innate AI this is our AI Strategic Assessment: 2 to 4 weeks, a ranked list of opportunities and a roadmap you can act on, whoever builds it.
If you want to talk through an AI decision you are weighing up, book a call with our team.
