Author: Gregg Quy, Senior Director, Insights & Strategy

There is a famous story about Michelangelo. Asked how he created such transcendent sculpture, he replied simply: “I saw the angel in the marble and carved until I set him free.” The masterpiece was always there. The artist’s gift was knowing what to remove.

This is, increasingly, what great market research consultancy looks like.

For decades, our value proposition rested on access. We knew how to find insight. We held the tools, the methodologies, the networks of respondents. Clients came to us because they could not easily attain what we could.

That world has ended. Today, any client with a browser and a well-formed question can summon synthetic market intelligence in seconds. LLMs democratise retrieval. AI flattens the information hierarchy. The raw material of insight, data, opinion, pattern, signal, is no longer scarce.

What is scarce is clarity.

The modern client does not drown for lack of information. They drown in its abundance. Competing data points. Contradictory syntheses. Plausible-sounding noise dressed as evidence. The challenge has shifted from finding knowledge to trusting it, from retrieval to curation, from search to sense-making.

Nowhere is this tension more acute than in pharmaceutical market research. Our clients operate in environments where the stakes of a misread signal are enormous. A misjudged physician sentiment, an overestimated payer threshold, or a poorly contextualized patient journey can shape the trajectory of a brand, a launch, or a therapy that reaches millions of people. In this world, the seductive fluency of AI-generated synthesis is not just unhelpful…it can be actively dangerous. Confident-sounding outputs built on shallow inference have no place in decisions of this magnitude.

But here is where the Michelangelo frame reaches its limit. Marble is honest. It does not perform. It does not generate the appearance of structure where none exists. Synthetic data does. AI outputs are fluent, confident, and often structurally convincing. They mimic the shape of insight without necessarily carrying its substance. This means our task is not purely sculptural. It is partly adversarial. We must interrogate before we carve, test before we trust, and learn to distinguish material that will hold from material that only looks like it will.

This is where the consultant becomes a knowledge broker: not a gatekeeper of information, but a skilled navigator of it. Someone who understands not just how to prompt an AI, but how to interrogate its outputs, reading the grain of synthetic data the way a sculptor reads marble, pressing where it yields truth and pulling back where it deceives. The result of this work is what we might call brokered knowledge: insight made more robust, more accountable, and more usable – knowledge that has been deliberately de- and reassembled for the specific moment, the specific decision, the specific client it must serve.

The best research consultants of the next decade will be artists of reduction. They will distinguish signal from noise, sharpen a client’s question before they touch a tool, and understand that insight is not what AI produces. It is what remains when everything unreliable has been carved away.

Our clients don’t need more marble. They need someone who knows where to place the chisel.