Why your AI advisor shouldn't be selling you software

Here's a question worth asking anyone who offers to advise you on your AI strategy: how do you make your money? If the honest answer is "by selling you the thing I'm recommending, " then what you're getting isn't independent advice. It's a sales process wearing the costume of strategy.

I want to be careful here, because plenty of excellent software companies give genuinely useful guidance. The issue isn't dishonesty. It's structural. When someone's revenue depends on you buying a particular solution, their advice will bend toward that solution, often without anyone intending it to. Incentives shape conclusions, quietly and reliably.

The conflict hiding in plain sight

The AI market is full of this. Implementation firms that "assess your readiness" and unsurprisingly find you're ready for exactly the platform they implement. Tool vendors whose strategy workshops conclude that their tool is the answer. Consultancies with deep partnerships that nudge every roadmap toward a preferred stack.

You can't get an unbiased answer to "what should we do?" from someone whose income depends on a particular answer. That's not cynicism. It's just how incentives work.

Why neutrality changes the advice

I chose deliberately not to resell software. It costs me a revenue stream that a lot of advisors rely on. But it buys something I consider more valuable: the ability to look a client in the eye and give them the answer that serves their business, with nothing riding on which way it goes.

That neutrality changes the conversation entirely. When I recommend a particular approach, the client knows it's because I think it's right, not because it pays my invoice. When I tell them a popular tool is wrong for them, there's no competing product I'm steering them toward. The advice is just advice.

Sometimes the best advice is 'not yet'

The clearest proof of genuine independence is this: a real advisor will sometimes tell you to do nothing. To wait. To fix your data foundations before you buy anything. To run the cheap experiment before the expensive transformation. To not adopt the shiny thing everyone's talking about because it doesn't fit your situation.

I've told clients to hold off on six-figure projects because the timing was wrong. None of them regretted it. A vendor, structurally, cannot give that advice, their entire model depends on the answer being "yes, now, with us."

How to tell the difference

When you're choosing who to trust with your AI direction, a few questions cut through quickly. Does this person make money based on which solution I choose? Have they ever told a client not to buy something? Can they recommend approaches that involve no purchase at all? Will they put the reasoning on the table so I can challenge it?

None of this means you shouldn't work with vendors, you absolutely will, and many are excellent. It means the person helping you decide which vendors, which use cases and whether to move at all should be someone with no stake in the outcome. Set the strategy with someone neutral. Then buy the tools with your eyes open. Getting those two roles confused is one of the more expensive mistakes I see organisations make.