The 100X Innovation Funnel

Listen to an AI audio recording of this newsletter

At our next ​Outthinker Summit on May 28​, we will gather with a group of strategy and transformation leaders already asking one of the most urgent questions facing large organizations:

How do we turn AI from a productivity tool into a growth engine?

That question has been on my mind since my ​recent conversation​ with ​Peter Weinberg​, co-founder of ​Evidenza​ and former head of ​LinkedIn’s B2B Institute​. Peter has been working at the frontier of what is called synthetic research: using AI-generated “lab grown customers” to simulate real customer populations and test ideas, messages, products, segments, and strategic choices at a speed and scale traditional market research cannot match.

The implications are bigger than marketing. They go to the heart of strategy.

The Real Job of a Company

In our recent ​Harvard Business Review​ article, my coauthors and I argued that the fundamental job of a company is to match ideas with markets. Most organizations are full of insight, but the insight is scattered across functions, teams, and business units. The challenge is building a system that can surface the right problems, connect them with promising solutions, and move the strongest ideas forward.

In a follow-up piece, “​HBR Nailed the Why — Here’s the How​,” I argued that AI may finally give us the tools to build those marketplaces at speed. Not simply by generating more ideas, but by removing the barriers that keep ideas from being tested, compared, connected, and acted on.

My conversation with Peter brought that possibility into sharper focus.

Because synthetic research gives us a glimpse of what happens when companies can test not just a few ideas, but hundreds or thousands of them.

Why Good Ideas Get Stuck

Inside nearly every large organization, thousands of ideas remain trapped in people’s heads, spreadsheets, meetings, and back channels. Only a tiny fraction get tested. Fewer still get tested with real market input.

The problem is not simply that companies lack ideas. It is that testing ideas with real customers takes time, money, and attention from the very people companies are trying to understand.

A team may have dozens of product, positioning, pricing, or adjacency ideas. But even if the team has the budget to test them, customers do not have unlimited capacity to respond. You cannot ask CFOs, legal executives, healthcare leaders, or busy consumers to fill out 100 surveys every time your organization wants to explore a new direction.

So companies test only a handful of ideas. The rest are filtered through hierarchy, internal debate, sales anecdotes, last year’s assumptions, or executive intuition.

Sometimes that intuition is right. Often, it is incomplete.

That is the hidden cost of missed ideas. As I explored in ​this article​, companies that fail to fix the inefficiencies in their idea marketplaces are not just slowing down innovation. They are losing potentially millions, even billions, in untapped value.

Enter the “Lab Grown Customer”

Peter framed the problem clearly: every strategic choice should in some way be informed by the voice of the customer. But getting that voice is hard.

If your customer is a CFO at a major bank, an in-house legal leader at a Fortune 500 company, or a mining operator in Northwest Australia, they are not sitting around waiting to fill out your survey.

Even in consumer markets, the sheer number of decisions companies need to make means traditional research is often too slow and expensive to guide every choice.

That is where synthetic research changes the equation.

Instead of surveying 1,000 real CFOs, you can create a statistically representative synthetic sample of 1,000 CFOs and ask them the same set of questions. Instead of waiting weeks or months for survey results to come in, you can get their answers tomorrow. Instead of testing the three ideas that survived internal politics, you can test 300, 1,000, or even 10,000.

Peter was careful not to oversell it. Synthetic responses are not “perfect” (he shows they are “90%” identical to human assessments) but, then again, human research is not accurate either. People get tired, misunderstand questions, and often answer aspirationally rather than truthfully. An AI customer doesn’t get tired, is less likely to misunderstand a question, and is more likely to answer truthfully — so it could be that synthetic customers may be even more accurate.

But the practical question is not whether synthetic research is flawless. The practical question is whether it points you toward the same decision a human sample would. In Peter’s experience, it does … much faster.

From Testing Three Ideas to Thousands

Imagine a company that currently tests three product concepts a year. Not because it only has three ideas, but because that is all the research budget and timeline allow.

Now imagine that same company testing 1,000 ideas.

Maybe not all the way to launch. But enough to see patterns. Enough to 10X the idea funnel. Enough to identify which customer needs recur. Enough to discover ideas that would otherwise have died in a spreadsheet, a brainstorming session, or a hallway conversation.

This is what AI does when applied to the idea marketplace. It removes barriers. It lowers the cost of experimentation. It increases the flow of signal.

It allows organizations to move from “Which few ideas can we afford to test?” to “which ideas are actually worth pursuing?”

That shift could radically expand the front end of the innovation funnel.

In “​AI-Powered Blueprint for Scalable Innovation​,” I described how AI can help internal idea marketplaces discover, match, value, prioritize, and mobilize ideas faster.

Synthetic research adds another powerful layer: the ability to pressure-test ideas against a simulated market before committing real resources.

What if the limiting factor were no longer research cost, but strategic imagination?

The Real Promise of AI

This is the larger promise of AI: not doing old work faster, but creating organizations that continuously sense and respond to the market.

But more signal does not automatically mean better strategy.

In a recent ​Outthinkers conversation​ with ​Neil Hoyne​, chief strategist at Google, we explored why many organizations say they want to be data-driven, but still use data to confirm what they already believe, delay hard choices, or create the appearance of rigor without real clarity. Peter made a similar point about customer research: some companies say they want the voice of the customer, then ignore it when it contradicts internal beliefs.

That is the leadership challenge AI will make impossible to avoid.

AI can generate more signals. Synthetic research can make customer insight faster and more scalable. But leaders still have to decide whether they are willing to listen, especially when the market tells them something they did not expect.

Why This Matters on May 28

That is why our members want to have this conversation at our May 28th Outthinker Summit in New York, and will certainly want to again in future summits.

If you are already registered, I encourage you to show up ready to engage. If you are not a member, we will record, summarize, and share the session. This is not a passive topic. It is not another abstract AI conversation. It is a practical question about how strategy work itself is changing.

How should ideas move through your organization? Who should have access to customer intelligence? How do we prevent AI from simply accelerating bad assumptions? And what strategic choices become possible when we can test 100 times more ideas than before?

Pressure-Testing What Comes Next

The New York Outthinker Summit is not about hearing generic predictions about AI. It is about comparing notes with leaders who are facing the same pressures: how to grow, how to adapt, how to make better decisions faster, and how to build organizations capable of turning uncertainty into advantage.

AI will not replace strategy.

But it may replace the slow, narrow, politically constrained systems through which many companies currently make strategic choices.

The companies that win next will not simply be the ones that use AI to save time. They will be the ones that redeploy that time into testing more ideas, learning faster from the market, and making bolder choices with better evidence.

That is the opportunity in front of us.

And it is exactly why May 28 matters.

Learn how to test better, learn faster and create precise strategy by becoming a member of ​Outthinker​ today.