Are you making AI decisions from fear or from clarity?

"Half of why we are rushing this is that I am terrified of being the leader who got left behind." You hear yourself say it out loud in the budget meeting, almost as a confession, and nobody pushes back. They nod, because they feel it too. The proposal gets approved in eleven minutes. And on the drive home you replay the moment and realise you cannot actually name what the spend is supposed to achieve, only the thing it is supposed to shield you against.

Here is the answer to the question in the title, before the diagnostics. You are making AI decisions from fear when the spend is justified by the fear of standing still, and from clarity when it is justified by a specific outcome you intend to reach. The two feel almost identical in the room. They produce wildly different results in the accounts. The first buys you the appearance of motion. The second buys you compounding capability. Telling them apart is a skill, and most of us were never taught it.

Why do I keep saying yes to AI spend I cannot fully explain?

Because the feeling driving the decision is real and widely shared, and almost nobody names it for what it is. IBM's 2024 study of 3,000 chief executives across 30 countries found that 51% agreed the risk of falling behind drives them to invest in some technologies before they have a clear understanding of the value those technologies bring. A separate 2024 ABBYY survey of 1,200 IT decision-makers (ABBYY is an enterprise software firm that runs an annual automation study) found fear of missing out driving AI adoption in 60% of businesses, with average AI investment of $879,000 in the prior year. The Conference Board (a long-established business research association), working with The Business Council, found 43% of leaders worried about losing competitive advantage to early adopters. This is the climate of the decision.

And it reaches the top of the house. IBM's follow-up 2025 study of 2,000 CEOs across 33 countries found that boards and chief executives keep raising AI investment even as results lag, and named culture, governance and workflow design, rather than the technology, as the constraint on the return. So the pressure is travelling in both directions at once. The board is anxious about the executive, the executive is anxious about the board, and the spend accelerates to relieve a feeling rather than to reach a destination. That is the mechanism. Name it and it loses some of its grip.

What does fear actually buy, and what does it cost?

Fear buys reassurance, and reassurance is the most expensive thing on the menu. When the driver is "so we are not last", the spend flows toward whatever is most visible: a tool everyone is talking about, a pilot you can announce, a vendor with a good demo. It chases the shiny object because the shiny object is what soothes the anxiety. The trouble is that anxiety is a poor strategist. It optimises for looking equipped rather than for being effective, and the gap between those two shows up later in the P&L, the place where it is hardest to reverse.

The numbers on that gap are sobering. IBM's 2025 CEO study found that only 25% of AI initiatives had delivered the return on investment leaders expected over the prior few years, and only 16% had scaled across the enterprise. Deloitte's 2024 State of Generative AI in the Enterprise study, drawing on 2,773 leaders across 14 countries, found more than two-thirds expected 30% or fewer of their AI experiments to reach full scale within the following three to six months. McKinsey's 2025 state-of-AI research found that fundamental workflow redesign correlates most strongly with the earnings impact that does appear. Read together, these say something clear: the spend is rarely the constraint. The coherence behind it carries the result.

There is a quieter cost too, and it is the one I watch for most. When IBM asked why the return stalls, the answer was not the models; it was culture, governance, workflow design and data strategy, the human and organisational layer underneath the tool. Fear cannot bear that emphasis. It wants the part it can see and announce, and it starves the part that actually carries the result. So the most expensive emotion in the AI budget does more than spend money on the wrong tools. It quietly defunds the human work that would have made any tool take hold.

How do I tell which one I am operating from right now?

Start with a single sentence, said out loud. Finish "We are investing in this so that..." If the sentence lands on a destination, a margin you intend to defend, a cycle you intend to halve, a capability you intend to own, you are operating from clarity. If it lands on a competitor, a headline, or a vague dread of being last, you are operating from fear, and that is useful to know rather than shameful to admit. The leaders who win the next decade are the ones who upgrade themselves first, and this is the upgrade: noticing the driver before it spends your money.

There is a reason this is hard in a packed week. A study of 1,112 parole rulings (a 2011 analysis of judges' decisions, published in the journal PNAS) found favourable decisions ran near 65% at the start of a session and drifted toward zero before each break, then recovered after food and rest; later work argues the effect is smaller and more context-dependent, yet the direction is intuitive. Depleted minds default to the safe, familiar, low-effort choice. In AI, the low-effort choice is to say yes to whatever quiets the fear. Coherence, head and heart pointing the same way, is what lets you make the considered call instead of the depleted one.

So before the next proposal reaches you, run it through a short sequence. This is the category of intervention that addresses fear-led spend: get coherent first, then let the money follow the intent.

1. Name the driver. Say the "so that" sentence aloud. Write down where it lands. Fear or destination, you now know which one is signing.
2. Attach an outcome. Tie the spend to one number you intend to move, and the workflow it lives inside, rather than to a tool you intend to buy.
3. Fund the human layer. Check that most of the effort sits on people and process, the part fear always starves, and resource it deliberately.
4. Decide when rested. Make the call with a clear head, away from the depleted end of a long day, where the safe-looking yes always wins.

Do that, and the spend stops being a sedative and starts being a strategy. The fear quietens, and it stops holding the pen. This is the move into what I call Phase Three. Phase One, the Age of Effort, was muscle: work hard, get a little more, linear growth. Phase Two, the Age of Scale, was machine: build once, sell to millions, exponential growth. Phase Three, the Age of Acceleration, is mind: output decoupled from human effort almost entirely, the phase AI unlocks, and a clear mind is the only thing that turns an AI budget into an AI advantage.

About the author

Thomas Green

British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.