The panic response loses. The freeze response loses. Here’s the disciplined third path – How to respond when your competitor just announced a major AI capability.
Maybe it was a press release. Maybe it was a customer forwarding you their deck. Maybe it was your son-in-law asking over dinner whether your company is doing any of “that stuff”. There are myriad flavors the realization may take; they all feel the same… Something shifted, and you’re now inside a window you can feel but can’t quite describe.
Here’s what I want you to do for the next ten minutes: breathe.
Don’t close the seventeen tabs you’ve opened since you heard the news. Don’t book the demo your CIO forwarded. Don’t send the calendar invite to your leadership team. Read this instead.
Because the worst mistake you can make right now is moving too fast. The second worst mistake is not moving at all. And the third, which most CEOs don’t even recognize as a mistake, is treating those two as the only options.
The two traps that cost you the response
The panic-buy looks decisive from the inside. From the outside it looks like concession. The signs: booking vendor demos within 48 hours of the competitor’s announcement; publishing an “AI initiative” before anyone has defined what it’s supposed to accomplish; hiring a Chief AI Officer because the competitor hired one. Every one of these moves cedes the category to whoever went first. You are now playing their game on their timing with their vocabulary.
The freeze looks prudent from the inside. From the outside it looks like absence. The signs: “let’s see how this plays out”; commissioning a committee; deferring the question to next year’s planning cycle. Markets don’t wait for you to be ready. Your competitor is compounding while you’re convening.
Most CEOs do both. They freeze for six weeks; then they panic-buy in the seventh. The result is the worst of both postures, delivered slowly.
Why the first move is the one that counts
You may be wondering whether the “you get one counter-punch” framing is rhetoric or reality. It’s reality, for four reasons that compound on each other.
Internal narrative. Your employees are watching. Your first move tells them whether this company takes AI seriously or performatively. That narrative is surprisingly hard to rewrite once it sets. The engineers you need to retain are forming their opinion of your company’s future this month; the ones you don’t keep will leave with a version of your company’s story that you did not author.
External narrative. The market is scoring this response. Analysts, customers, prospects, and recruits are all making inferences from your move. Being third to respond is forgivable. Being third to respond in a muddled way is not.
Budget commitment. First moves in AI consume budget that compounds. A poorly-directed first move costs you the second move’s funding; the board approved the $4M initiative, it disappointed, and the next ask goes nowhere. Your runway is not infinite and every misfire shortens it.
Data flywheel. The first deployment starts collecting data. If the first deployment is aimed at the wrong problem, you’ve spent eighteen months teaching a model to be good at something you didn’t need. The competitor who aimed correctly has an eighteen-month head start on the thing that actually matters.
Second chances exist. They are more expensive than the first chance, and they arrive with less patience attached.
The Disciplined Third Path: Broaden Your Horizon of Possibilities *Before* You Jump
The mistake in both panic-buy and freeze is the same. They both narrow the response space before broadening it.
Panic-buy narrows to whatever the first vendor in your inbox is selling. Freeze narrows to “do nothing right now”. Both skip the step that determines whether the counter-punch lands: widening the set of possible responses before committing to one.
This is the discipline of our approach to innovation. Divergent thinking first – on its own. Convergent thinking second. In that order, without shortcuts.
Divergence means mapping every possible response, including the ones you immediately dislike. Convergence means narrowing to the top few that actually fit your business and which your leadership has the bandwidth to give real focus. The diamond looks simple on a whiteboard. It is brutally hard to execute under pressure, because every instinct in a CEO whose phone is ringing wants to skip straight to convergence. Making a choice feels like decisive action. Discipline is resisting that pull for the 72 hours it takes to widen properly. It is worth it. Every time.
What divergence actually looks like
The response options most CEOs consider in the first 48 hours: match the competitor’s feature; announce a similar initiative; wait. Three options. All narrow.
The response options that a disciplined divergence produces are closer to these:
- Match the feature directly. Build what they built.
- Match and one-up. Build what they built, with a capability they lack.
- Undercut on price. Commoditize the category they just tried to create.
- Reposition around a weakness their move exposed. Their announcement revealed what they bet on; bet on the opposite.
- Abandon the contested ground. Double down on a different moat where you’re structurally advantaged.
- Partner asymmetrically. Move with a provider they cannot partner with for regulatory, competitive, or contractual reasons.
- Pivot to white-glove service. Make it a strategic advantage for your company that you don’t put AI in front of your customers. Limit it to “behind the curtain” to set your approach apart.
- Acquire your way in. A smaller player already has the capability; buy the capability and the team in one move.
- Attack the adjacent workflow. Put AI into a workflow the competitor hasn’t touched, changing the locus of the fight.
- Rebuild the offer so AI is invisible infrastructure. Not a feature you announce; a layer that makes your existing product better in ways customers feel but don’t name.
- Deploy AI internally, silently. Widen your cost advantage without giving the market anything to benchmark against.
- Attack their flank. Their announcement suggests overextension; find the customer segment they just stopped serving well and go after it.
This is a list of a dozen response categories, not three, and it can be expanded further in many situations. Most will be wrong for your business. But you cannot know which three are right until you’ve seen all twelve on the board.
Divergence is not brainstorming for its own sake. It is exploring every direction to learn enough so that the weak ones eliminate themselves and the strong ones reveal their strength in the convergent phase.
A practical divergence exercise (half a day)
Put your executive team in a room. No laptops. No vendor decks. One whiteboard.
For each of the eleven categories, force a two-sentence answer to: “If we did this, what would it require, and what would it cost us in other directions?”
By lunch, three or four will have eliminated themselves. By the end of the afternoon, you will have two or three serious candidates and a clearer view of your business than you had when you walked in. That clarity is worth the half day even if you learn nothing else.
What convergence actually looks like
Convergence is not consensus. It is the application of three criteria to the candidates that survived divergence:
- Does this play to a strength the competitor cannot quickly copy?
- Does this compound over time, or does it deliver a one-time bump?
- Can we execute this well enough that the market will notice?
Three yes answers here? – you have a candidate response. Any no(s)? – you have a distraction dressed up as a response.
Most CEOs want to add a fourth criterion: “does this match what the competitor did?” That criterion is precisely the one that cost them the counter-punch in the first place. Strike it from the list.
Read their punch before you throw one
A counter-puncher in boxing does not react to the punch that was thrown. He reads what the punch committed the opponent to, and exploits the opening that commitment created. Your competitor’s AI announcement is the commit. What did it commit them to?
Ask these questions before you write a single word of response:
- Which customer segment did they bet on, and which did they implicitly deprioritize?
- Which technology stack did they lock themselves into?
- Which use case did they put on their roadmap that will take 18 months to deliver, whether they admit it or not?
- What did they not announce? (What they did not announce is often more revealing than what they did.)
- What did the announcement reveal about their internal capability, or the lack of it?
- Who on their leadership team now owns this publicly, and what does that tell you about their willingness to pivot?
The counter-punch that lands is the one aimed at the opening the announcement created; not the one aimed at the announcement itself.
The next 72 hours
Not vendor meetings. Not a press release. Not an all-hands. The sequence that preserves the counter-punch:
Hour 0 to 24: Read the punch. Get your team in a room to analyze what the competitor’s move actually commits them to. Refuse to discuss your response yet. Anyone who jumps to response gets redirected.
Hour 24 to 48: Diverge. The eleven-category exercise. Widen before narrowing. The goal of this session is not a decision; it is a map.
Hour 48 to 72: Converge. Apply the three criteria. Narrow to two or three serious response candidates. Assign an executive owner to pressure-test each one over the following week.
At the end of 72 hours you don’t have a response. You have a disciplined shortlist. That is infinitely more valuable than the response your competitor had when they announced, because their response is now fixed and yours is still adapting.
What gets remembered
Ten years from now, nobody in your company will remember how fast you responded to the competitor’s AI announcement. They will remember whether the response worked.
The CEOs who get this right are not the ones who move fastest in the first week. They are the ones who spent the first 72 hours widening their options while their competitors were narrowing theirs. They earned the right to converge by doing the work of divergence first.
Make the counter-punch count. You only get one.
