You run a company in Minneapolis or somewhere close to it. You have decided AI could help, or someone has told you it could, and now you are looking for a firm to help you do it right. So you search. And you get a wall of companies that all sound identical. The same words, the same promises, the same photo of a glowing brain. It is hard to tell who actually builds things and who just talks about it.
We build software and AI here, so we know what separates the firms worth calling from the ones worth skipping. Here is what to look for, in plain terms.
Look for the ones who audit before they sell
The first and clearest signal is whether a firm wants to understand your systems before it recommends anything. Be careful with anyone who leads with the model. If the first conversation is about which AI they will install and how impressive it is, they are selling a product, not solving your problem.
A good partner starts by looking at what you already have. What software you run, how it connects, where the work actually gets stuck, and which of your problems even need AI in the first place. Often the biggest win is not a new model at all. It is fixing the tangle of tools underneath, so that whatever you add on top has something solid to stand on. A firm that starts with an audit is a firm that plans to build you something that lasts.
Ask where the work will live
AI that lives in its own window gets abandoned. Your team already works in email, in a spreadsheet, in the one tool they open every morning. If a new system asks them to go somewhere else, log in again, and copy answers back and forth, they will quietly stop using it. So ask a simple question: how will this fit into the tools we already use? If the answer is vague, that is your answer.
Look for ownership and a clear finish line
Most AI projects do not fail loudly. They stall. The demo works, everyone is excited, and then months pass and it never reaches the people who were supposed to use it. The usual reason is that no one owned it and no one ever said what done meant.
So ask who will own the work through to production, and ask what success looks like in plain words before anything starts. A firm that can answer both is a firm that has shipped things before. A firm that talks only about possibilities has not.
Local still matters, even now
Plenty of good work happens remotely, and you should not rule out a firm for being far away. But there is real value in a partner you can sit across from. Someone who can come to your office, meet the people who actually do the work, and understand the way business gets done in the Twin Cities. Being able to point at the same screen in the same room still moves projects faster than a string of calls. If you can find a firm that is genuinely local and genuinely capable, that is the short list.
The red flags
A few things should make you slow down. A firm that leads with the model instead of your problem. One that skips the audit and jumps straight to a proposal. Deliverables described in words so vague you cannot tell what you are buying. No named owner. A pitch that cannot be explained without jargon. And the big one: anyone who promises to replace your entire stack on day one. Real work is more careful than that, because your business still has to run while it happens.
How we think about it
For what it is worth, here is our own approach, so you can measure it against anyone else you talk to. We audit everything first, including the parts that are not software. We repair what is worth keeping, replace what is not, and build what you actually need. Then, and only then, on a foundation that holds, we add the AI and the agents that were the point. It is less exciting than leading with a demo. It is the reason the things we build are still running a year later.
If you are in Minneapolis and you want a straight answer about where AI actually fits in your business, that is exactly what an audit is for. No pitch, no glowing brain. Just a clear look at what you have and what to do next.