In a meeting with the board of a financial sector company, someone asks: "We have Copilot, we're testing Claude, the marketing department wants agents, IT is proposing RAG on our documents, and the CEO came back from a conference asking about multimodal models. Where do we start?"
I answer: "By saying 'no' to most of that."
Silence. Then someone laughs. Then we start talking seriously.
Tool Syndrome
I see this regularly in mid-to-large companies. The organization buys licenses for yet another AI tool before it even understands what it actually needs the first one for. Copilot lies dormant — data shows that 3-4% of employees use it regularly. But someone is already proposing the next purchase. Because the competition "has it." Because it was shown at a conference. Because the vendor promised a demo.
This is not an AI strategy. This is tool syndrome — the belief that the next tool will solve a problem whose root cause lies somewhere else entirely.
The problem isn't technical. The problem is that nobody in the organization has the mandate to say: "Let's stop. We don't need this. Not yet."
The Value of "No"
There's a trap in consulting: the client wants to hear "yes." Wants confirmation that their idea is good, that it's worth buying this tool, that this project makes sense. Many consultants say "yes," because "yes" is easier. "Yes" generates revenue — a new project, a new deployment, a new invoice.
But the value of a good consultant doesn't lie in confirmation. It lies in a well-timed "no."
"No — this project doesn't make sense given the current state of your data." "No — this agent won't work because you don't have a process it could automate." "No — another tool won't solve a problem that is process-related, not technological." "Not yet — but in three months, when you've finished cleaning up the data, let's revisit."
These sentences cost the consultant revenue in the short term. But they build something more valuable — trust and real results in the long term.
Why Organizations Can't Say "No" on Their Own
There are several reasons the internal voice of reason doesn't work:
Market pressure. Every conference, every report, every article says: "deploy AI faster." The board feels pressure that they're falling behind. That pressure translates into purchasing decisions devoid of strategy.
Lack of evaluation competence. The IT department can evaluate a tool technically. But it can't assess whether the tool fits the company's strategy, whether the problem it's supposed to solve even exists, or whether the organization is ready for its deployment.
Incentive structure. It's easier for the IT director to buy a new tool (because that's an "initiative") than to tell the board: "Let's not buy anything new. Let's first make use of what we have." The latter sounds like a lack of ambition. The former sounds like innovation.
Vendor push. Tech companies have sales teams whose sole job is to sell the next license. Their demo is always impressive. Their case study is always successful. Their ROI is always positive. Nobody talks about projects that didn't work out.
What Actually Works
From my experience, companies that achieve real results with AI have one thing in common: they started with one thing. Not five tools in parallel. Not an "AI strategy for the entire organization." One process, one team, one measurable goal.
They chose a specific problem — e.g., report automation, customer service support, legal document analysis. They did it well. They measured the effect. Then, with data and experience in hand, they expanded the scope.
That sounds obvious. But in practice it requires something most organizations lack: someone who says "no" to everything that isn't that one thing. Someone who protects the team from distraction. Someone who has the mandate to tell the board: "I understand the pressure. But if we start five things at once, we'll finish none of them."
A Consultant Who Says "No"
My role isn't to sell tools or deployments. It's to help the company find that one place where AI delivers measurable value — and block everything else until that one thing works.
Sometimes that means I say: "You don't need me for more than two meetings. You have everything you need — you're just lacking a decision." Sometimes it means six months of working together. But it always starts with an honest answer to the question: what do you actually need?
If you're looking for someone to help you organize what your company already has before buying something new — let's talk about strategy.
Dear Reader, if your organization is drowning in AI tools but not seeing results — I invite you to get in touch. Leszek Giza.