Is Your AI Adoption FOMO-Led? How to Tell, and What to Do Instead

If you can't finish the sentence "we need AI so that we can ___" with something specific, your AI adoption is FOMO-led, not strategy-led. That's a pattern AI consultants see constantly, and it's the single biggest predictor of whether an AI initiative produces anything or just produces activity.

I talked about this recently on my podcast with David Cohen, founder of Superposition, a consultancy that works exclusively with other data and AI consultancies. When I asked him what clients actually mean when they say "we need AI" his answer was blunt: "I don't think they typically mean anything in particular. They usually say that out of a place of FOMO, trying to capture what they think the market says that they should."

Here's how to tell if that's you, and what to ask instead.

The Tell: You Can Name the Tool, Not the Problem

FOMO-led AI adoption has a specific shape. Someone in leadership says "we need to be doing more with AI". Everyone nods. A budget gets approved. A vendor gets selected. And at no point does anyone answer the question that should have come first: what business problem is this actually solving?

David's read matches something I see in my own work with mid-market clients: a huge amount of frustration in the consulting world right now comes from exactly this gap, table setting correctly for what a client actually needs before building anything, rather than jumping straight to tools.

The tell is simple. If your AI initiative is described in terms of the technology (we're deploying Copilot, we're building a custom GPT, we're rolling out an AI platform) rather than in terms of an outcome (we're cutting proposal turnaround from five days to one, we're reducing research time so the team can take on three more clients), you're likely reacting to market pressure rather than solving a problem you've actually identified.

Why This Happens: Access Isn't the Same as Value

There's a deeper point buried in David's answer that's worth pulling out on its own. He pushed back on the idea that AI has reshaped the broader economy as dramatically as the hype suggests: "the average person does not care or want to use AI tools." What's actually changed is access to intelligence, and access to information has never, on its own, been what companies actually pay for.David drew a direct comparison to what happened when Google made the world's information searchable. Consultancies and service firms didn't disappear. If anything, the value of the right knowledge, applied to a specific, contextualized situation, went up. As he put it: "having access to the right information is the actual value... knowledge that makes sense in the setting they're navigating... factors that an AI system could never understand or pick up, like internal politics, feelings, context of the greater situation."This matters for FOMO-led adoption specifically. If your AI strategy is built around "we now have access to this powerful new tool", you're solving the wrong problem. What's scarce, and what's actually worth building a strategy around, is knowing exactly which problem in your specific business this level of access should be pointed at.

The Question to Ask Instead of "Do We Need AI?"

Skip "should we be using AI more". Ask this instead: what specific task, in what specific function, is currently slow, expensive, or inconsistent enough that fixing it would show up in a number leadership already tracks?That's a very different starting point. It doesn't start with the tool. It starts with a bottleneck that already exists, whether or not AI turns out to be the right fix for it.This is exactly what a structured audit is built to surface, and it's why I put it as the first stage of the AI Flywheel rather than somewhere further down the list. (Read this for the full information on the AI Flywheel: What Is the Difference Between an AI Pilot and a Full AI Transformation?)Before any tool gets chosen, the audit asks which functions have the most repetitive, low-judgment work, where the data risks already sit, and what's actually costing the business time or money right now. Answer that first, and the tool selection becomes a much smaller, much cheaper decision.

Three Questions That Separate Strategy From FOMO

1. Can you name the metric this is supposed to move?

Not "efficiency" in the abstract. A specific number: proposal turnaround time, first-response time on leads, revision cycles on client deliverables. If nobody in the room can name the metric, the initiative doesn't have a target yet, it has a vibe.

2. Would this problem still be worth solving if AI didn't exist?

If the honest answer is that the problem only feels urgent because AI made it feel solvable, that's a signal you're chasing the tool, not the problem. Real bottlenecks were real before generative AI existed. They just might have been harder to fix.

3. Who specifically benefits, and how would they describe the change?

If you can picture the actual person, in their actual role, describing what got easier or faster for them, you're solving a real problem. If the answer is a vague organizational benefit no one person would ever say out loud ("we'll be more competitive"), you're still in FOMO territory.

What This Looks Like Done Well

A marketing team that's been spending a third of its week repurposing long-form content into shorter formats isn't chasing AI because it's trendy. It has a specific, countable bottleneck, and AI happens to be a good fix for it. A sales team that's losing deals because proposals take five days to turn around isn't buying a platform for the sake of buying a platform. It's solving a problem that predates the tool and will still matter regardless of which tool solves it.That's the difference between AI adoption that compounds and AI adoption that generates a press release and then quietly fades. One starts with a real, specific bottleneck. The other starts with a fear of being left behind.

The Short Version

If your AI initiative can't answer "we need this so that we can ___" with something specific and measurable, it's FOMO, not strategy. The fix isn't to slow down on AI. It's to start with the actual bottleneck instead of the tool, and let the tool selection follow from that, not precede it.David Cohen's full conversation on this, including why the services business model itself is being reshaped by easier access to contextualized knowledge, is on episode 276 of AI Literacy for Entrepreneurs.If you want a structured way to find your organization's real bottlenecks before you spend anything on tools, the NorthLight AI Readiness Audit gives you a structured picture in about 10 minutes. Run the audit now.