At CES 2026, McKinsey didn't demo a gadget. They demoed a process. And it was more disruptive than anything with a screen.
The consulting giant showed a live AI workflow that takes product development cycles that used to run six to nine months and compresses them into roughly two weeks using AI-generated consumer insights, digital testing, and simulated customer personas.
"The key to great product is fast iteration. Try this. Did it work? Okay, it's okay, but these three things are still a problem," Dave Fedewa, a partner at McKinsey, told International Business Times.
Five months later, the data suggests they were underselling it.
100,000 Comments in Two Hours. No Survey Required.
The system McKinsey demonstrated pulls in over 100,000 unprompted consumer comments from TikTok, product reviews, and social media, then clusters them into actionable product attributes that engineering teams can work from immediately.
"We can ingest 100,000 comments in a couple hours on a specific space, way better than a survey," Fedewa said.
From there, visual concepts are generated in about an hour, then tested with large sample sizes and AI "personas" representing specific consumer archetypes: a suburban mom with three kids, a 45-year-old football dad, a budget-conscious Gen Z renter. These aren't demographic labels. They're behavioral simulations that stress-test product messaging, packaging, and design before anything reaches a physical prototype.
The old model was build extensively, then test with 20 people behind a one-way mirror and hope for the best. The new model is tiny build, relentless testing, rapid iteration. Consumer brands are now getting statistically significant feedback from thousands of simulated and real consumers in days.
71% of Organizations Now Use Gen AI. Only 1% Say They're Mature.
McKinsey's own State of AI research, updated through early 2026, confirms that the CES demo wasn't a one-off showcase. Seventy-one percent of organizations now regularly use generative AI in at least one business function, up from 65 percent in early 2024. The most common deployments are in marketing and sales, product development, service operations, and software engineering.
But here's the gap that matters: only 1 percent of company executives describe their gen AI rollouts as "mature." Almost everyone is experimenting. Almost nobody has scaled it enterprise-wide.
In the most advanced organizations, long product requirements documents are disappearing entirely. Instead of writing exhaustive PRDs, product managers move directly to prototypes. AI enables rapid mockups, fast iteration, and real-time testing, often without waiting on a full design or engineering cycle.
McKinsey's revised edition of its best-selling book "Rewired," published in April 2026, argues that the practice with the highest correlation to value was reimagining workflows end to end, not just dropping AI tools into existing workflows.
That distinction is critical. The companies seeing real returns aren't the ones that gave their employees a ChatGPT login. They're the ones that rebuilt entire processes around what AI makes possible.
Generic AI Won't Save You. Proprietary Data Will.
One of the more revealing moments from the CES conversation was Fedewa's bluntness about off-the-shelf tools.
"You could plug all these questions into ChatGPT. The answers that would come out would not be good," he said.
McKinsey's position is that useful AI outputs require proprietary training data built from decades of actual product outcomes. They've been doing consumer product research for 20 years, building a library of cases and results, then tuning AI models on top of that institutional knowledge.
This tracks with McKinsey's May 2026 research warning that AI is "not a productivity revolution" but a "competitive reset," where the winners are not those who adopted the technology fastest but those who understood where value was moving earliest and positioned themselves to capture it.
The Companies That Move First Will Set the Cost Baseline for Everyone Else
McKinsey's latest research frames the current moment as a narrow window. Early movers can scale faster, lock in lower cost positions, and make it harder for competitors to catch up once the benefits begin to compound.
The firms that treat AI product development as a pilot program will find themselves competing against companies that compressed their entire innovation cycle into two-week sprints. The speed gap doesn't close. It widens.
LATAM Airlines, cited in McKinsey's April 2026 "Rewired" update, is probably a year ahead of most companies in terms of adopting and embedding agentic engineering, not just for coding, but for the entire software development life cycle. Singapore's DBS Bank went from taking 18 months to deploy its first AI model to deploying one every two months.
The blueprint McKinsey showed at CES five months ago is no longer theoretical. It's being implemented by companies that decided the old timeline was a competitive liability. The question for everyone else is how long they can afford to keep building the slow way while their competitors stopped months ago.
