During the first months of Redstone Labs, our pitch was “AI and automation consulting.” It sounded great. It was what the market was looking for. Every company wanted to “do something with AI” and we were there to help.
The problem was we were selling a hammer. And when you sell a hammer, everything looks like a nail.
The hammer trap
We’d walk into meetings and the conversation always revolved around AI. “What can we do with machine learning?” “Can we use AI to improve our customer service?” “Can you build us a predictive model?”
And we, with all the enthusiasm in the world, found ways to fit AI into every corner. Because that’s what we sold.
Until one project taught us the lesson.
A client asked us for a recommendation system for their e-commerce platform. They had a catalog of 500 products and 2,000 active users. We did the analysis, evaluated the data, designed the model architecture.
Halfway through, we realized something uncomfortable: with 500 products and 2,000 users, a simple rule-based recommendation system (frequently bought together, related categories, popularity) worked just as well as any ML model. And it was 10 times easier to maintain.
We had proposed the sophisticated solution because that’s what we sold. Not because it’s what the client needed.
The shift
That project forced us to ask an uncomfortable question: are we solving problems or selling technology?
They’re different things. Selling technology means convincing the client they need what you offer. Solving problems means understanding what the client needs and using the right tool, whether that’s AI, a Python script, a spreadsheet, or a difficult conversation with their team.
Abraham Maslow described it in 1966: “If the only tool you have is a hammer, you tend to treat everything as if it were a nail.” He wasn’t talking about tech consulting, but he might as well have been.
We decided our differentiator wouldn’t be “we’re AI experts.” It would be “we solve business problems using the right tool.” Sometimes that tool is AI. Sometimes it’s not.
What changed in practice
1. We start with diagnosis, not the solution.
Before: the client told us what they wanted and we built it. Now: the client tells us their problem and we evaluate what they need. Sometimes it matches what they asked for. Sometimes it doesn’t. But we always start by understanding before proposing.
2. We stopped being afraid to say “you don’t need AI.”
At first it was scary. If you tell a client they don’t need what they came to buy, don’t you lose the project? Reality was the opposite. Clients value honesty. They trust someone who says “this can be solved more simply” more than someone who sells them the most expensive solution.
3. We expanded the stack.
AI is still our strength. It’s where we have the most experience and technical depth. But we also do data architecture, process automation, system integrations, and strategic consulting. Because a business problem rarely gets solved with a single tool.
What the market really needs
After years working with companies of different sizes in LATAM, the conclusion is clear: most don’t need “AI.” They need someone to sit with them, understand their operation, identify where they’re losing time and money, and propose a solution that works in their context.
Sometimes that’s a sophisticated ML model. Sometimes it’s connecting two systems that don’t talk to each other. Sometimes it’s automating a manual report that eats 20 hours a month. Sometimes it’s telling the client the problem isn’t technological but process-related.
What the market needs is consultants who solve problems. Not AI salespeople disguised as consultants.
The phrase that defines us
After that shift, we coined an internal phrase that guides everything we do: “We don’t sell AI. We solve problems.”
It’s simple. But it completely changed how we present ourselves, how we evaluate projects, and how we measure success. Success is no longer “we implemented AI.” It’s “the client has a measurable result that justifies the investment.”
And if that result is achieved without a single line of machine learning, perfect. The problem was solved. That’s the only thing that matters.