A faster deck is still a deck
I have spent the last few months watching the biggest consulting firms brag about how fast they have gotten. The claims are not marketing fluff. Bain has become an OpenAI Elite Partner and has put money into the OpenAI Deployment Company, the vehicle built to push AI into enterprise work at scale. Bain now says AI and technology enabled work already makes up something like 30 percent of its business, and the firm expects that share to approach half. BCG struck its own partnership, this one with Anthropic, so client teams can reach directly into Claude for synthesis and report writing rather than waiting on an analyst to draft it by hand. BCG expects AI enabled work to reach roughly 40 percent of its revenue by 2026. McKinsey built its own internal tool, called Lilli, and put it in the hands of thousands of consultants.
None of this is theater. A research and synthesis process that used to eat six weeks of an associate's life can now happen in days. The interviews still take time, because people still need to talk to each other, but the model building, the pattern matching across old engagements, the first draft of the deck itself: all of that moved from weeks to days at three of the most conservative firms in the world. If you sell against consulting decks, as I do, that should worry you a little. It worried me.
Then I noticed what the speed does not fix.
A deck is a snapshot. It captures what the market looked like, what the competitor was doing, what a customer said in an interview, on the day someone stopped typing and hit send. It does not matter whether that day arrives after ten weeks of work or three days of work. The moment the deck is finished, it stops learning. Everything that happens after it, a competitor's price cut, a regulator's ruling, a customer who churns the week after the interview, arrives too late to change a single slide. Faster production does not change this. It only means the snapshot gets taken sooner, which leaves more time between the snapshot and the decision for the world to move without anyone updating the picture.
Strategic decisions do not live in a moment. They live in a stretch of time between when someone starts thinking about a question and when someone finally has to answer it, and that stretch is exactly where a deck cannot follow. A CEO does not decide the day the consultants leave the building. She decides three weeks later, after the board meets, after the quarter's numbers come in, after a competitor announces something nobody saw coming. The deck she is holding at that moment was accurate once. It is not lying to her. It simply stopped paying attention the day it was printed, and nothing about how fast it was printed changes that.
So speed narrows a gap without closing it. Consulting used to hand a CEO a ten week old answer to a decision made in week twelve. It might now hand her a two week old answer to the same decision. Better, genuinely. Still an answer that was true once and has not heard anything since.
What a CEO wants, when she is honest about it, is not a faster snapshot. It is an answer that keeps listening. Something still paying attention on the day she has to decide, not just on the day someone handed her a folder. And she wants that answer to come from people who have something on the line if they are wrong, not from a team paid the same fee whether the recommendation holds up or falls apart six months later. An outside team, however sharp, gets scored on the presentation. The people who have to live inside the decision get scored on the outcome. Those are different incentives, and no amount of AI in the research phase changes which one is driving the answer in front of her.
This is the problem I built Assay to solve. Not faster research. A different source, and a different clock. The number a CEO gets from us comes from the company's own people, the ones who will actually work inside whichever path gets chosen, and it moves because those people are putting real money behind their own view, not because someone refreshed a slide. Nobody's name shows up next to a trade on any screen, and no one trades more than a paycheck's net pay either way, so the exposure stays bounded and ordinary, and the number still costs something to move. The number keeps updating for as long as the question stays open: when the facts change, the number changes with them, without anyone commissioning a new round of interviews. And it never makes the call for her. The read is advisory. She still decides. She just decides holding an answer that was still listening five minutes ago instead of one that stopped listening five weeks ago.
I do not think the consulting firms are wrong to invest in speed, and I do not think their diagnosis work is worthless. Reading a market from the outside, spotting a pattern across a hundred companies nobody at any single company has seen: that is a skill worth paying for, and it is not what I am arguing against here. What I am arguing against is the idea that a faster deck solves the actual problem. It does not. A deck was never slow because typing was slow. It is limited, and stays limited, because it is a photograph of a moving decision, and a photograph does not get less fixed no matter how quickly you develop the film.
The CEOs I talk to are not looking for a faster photograph. They are looking for something that is still awake when they finally have to choose.
Sources
Bain and OpenAI, Bain press release: https://www.bain.com/about/media-center/press-releases/2026/bain-company-openai-a-new-venture-to-deploy-ai-at-enterprise-scale/
Futurum Group on Bain's OpenAI Elite Partner status: https://futurumgroup.com/insights/bain-company-elevates-ai-strategy-as-openai-elite-partner/
Techzine on BCG's partnership with Anthropic: https://www.techzine.eu/news/analytics/111351/consulting-group-bcg-gives-clients-direct-access-to-anthropics-ai-tools/
AOL, on AI reshaping McKinsey, BCG, and Bain fees and revenue share: https://www.aol.com/finance/ai-forcing-mckinsey-bcg-bain-063300951.html