AI made analysis fast. Judgment is the new bottleneck.
In October, McKinsey Quarterly published The judgment bottleneck: Balancing AI and human speeds in strategy by Alex D'Amico, Bruce Delteil, Laura LaBerge and Matt Banholzer. It's one of the clearest descriptions we've read of something product leaders feel every week, and it's worth reading in full. Below is our summary of it, followed by what we think it means for the people who own product strategy.
What the article says
The authors' argument is simple. AI has made the analytical part of strategy dramatically faster: research, synthesis, scenarios, the first draft of almost anything. The part that turns analysis into a decision hasn't sped up. Someone still has to understand what the analysis implies, challenge its assumptions, weigh the trade-offs, bring the people who will carry it out on board, and own the call. A process moves at the pace of its slowest step, so as analysis accelerates, human judgment becomes the constraint.
Their most striking example is a team that started producing analysis ten to twenty times faster and running several times as many scenarios. Decisions didn't speed up. They slowed down, because the executives reviewing the work couldn't tell which assumptions and data sources sat behind it. When you don't know which part of an analysis to trust, you end up checking all of it.
The article describes three ways this goes wrong:
- Treating strategy as an efficiency problem. To keep up with the volume, review gets skipped or pushed down to people without the experience to do it. AI starts interpreting AI's output, and activity rises while understanding and accountability fall.
- Burying leaders. More analysis arrives than leadership can absorb, and attention drifts from the decisions that matter most.
- Losing the training ground. The analytical grunt work that used to build junior people's judgment is now done by a model, so the next generation of leaders gets fewer chances to practise.
Their remedies are just as practical. Treat senior attention as the scarce resource it is: bring leaders the key assumptions and the "what would you have to believe" questions, not the whole report, and reserve their time for the decisions with the highest stakes and the most uncertainty. Build judgment on purpose by making AI show its working and expecting teams to answer for it. Keep a named, experienced person accountable for every consequential decision. And use the extra speed to revisit decisions more often, not to make more of them.
One chief strategy officer quoted in the piece puts it in a sentence: "the goal of AI here isn't efficiency, it's better outcomes."
The article is written about corporate strategy. The application to product below is ours, not the authors'.
The same bottleneck, one level down
Product teams got AI earlier and faster than most strategy teams did. Interview summaries appear minutes after the call. Support tickets get clustered overnight. A spec, a prototype and a launch plan can all exist by Wednesday.
So the head of product's week has changed. The question is rarely "do we have the analysis?" any more. It's "which of these do I trust, and what exactly am I being asked to decide?" Every AI summary, generated spec and dashboard is a request for judgment, and they arrive faster than anyone can give it.
Most tools answer the first question and make the second one harder. They're built to produce: more summaries, more drafts, more output, sooner. That's the efficiency mindset the article warns about, sold as a feature.
Built for the slower clock
We built Outcomify for the other half of the problem. Its job isn't to generate more analysis. It's to make the judgment step faster, safer and owned by someone. That's what we mean by purpose-built: the design decisions all point at the bottleneck the article describes.
It shows the working, not just the answer. Every opportunity and assumption in Outcomify carries the evidence behind it, down to what customers actually said and where they said it. When a leader reviews a recommendation, they can see what it rests on and check the part that matters, instead of re-checking everything because the source is a black box.
It puts the assumptions on the table. "What would you have to believe?" is the backbone of an Opportunity Solution Tree: the outcome you're after, the customer needs that drive it, the bets you're making, and the assumptions each bet depends on. That structure is the attention triage the authors recommend. A leader reviews the riskiest assumptions on one page, not a forty-page deck, and the team tests them before the expensive work starts.
AI proposes, people decide. Canopy, our built-in AI, does the legwork: it reads your research and drafts opportunities and assumptions, each linked back to where it came from. Nothing changes your strategy until a person approves it. That's deliberate. It breaks the loop where AI checks AI, and it keeps a name on every decision.
It keeps the reasoning, not just the result. Every change is versioned, human or AI, so months later you can replay how a decision was reached. That's accountability, and it's also the apprenticeship the article worries is disappearing: a new product manager can trace why the team believed what it believed, instead of inheriting a conclusion with no history.
It's a living strategy, not an annual one. The tree is where the team works every week, not a diagram redrawn for the quarterly review. When the evidence moves, revisiting a decision is an ordinary week's work rather than a planning cycle, which is exactly the faster cadence the authors suggest using AI's speed for. And because product, design and engineering work in the same tree, the alignment that the article counts as part of judgment happens along the way rather than in a meeting at the end.
What Outcomify won't do
Outcomify doesn't scan patents, model your finances or run market simulations. Use the best AI tools you can find for the fast part of the process. Outcomify is where their conclusions meet your judgment.
And it won't make your strategic decisions for you. That isn't a gap on our roadmap. The article ends by arguing that the winners won't be the organisations that automate the most decisions, but the ones that use AI's speed while protecting the judgment it can't replace. That's the bet we've built Outcomify on.
Read the original
If you lead product or strategy, read the McKinsey article in full. It has more examples than we've covered here, a useful three-tier way to decide which decisions need senior attention, and a lesson from the two-speed IT era.
If it describes your week, try Outcomify free and see what your strategy looks like when every bet shows what it rests on.
Source: Alex D'Amico, Bruce Delteil, Laura LaBerge and Matt Banholzer, "The judgment bottleneck: Balancing AI and human speeds in strategy", McKinsey Quarterly, October 15, 2026. The summary above is our reading of the article and the quote is taken from it. McKinsey & Company is not affiliated with Outcomify and has not reviewed or endorsed this post.