The Prompt

October 29, 2026

7 min read

What AI Still Can’t Touch: Where Researcher Value Concentrates Now

What AI Still Can’t Touch: Where Researcher Value Concentrates Now

Explore where human judgment matters most in AI-assisted market research, from consumer experiences to strategic decisions and accountability.

Editor’s Note: Following our recent IIEX.AI event, I’m excited to share Ashley's synthesis of another central question facing our industry: as AI accelerates research tasks, where does human judgment matter most? (You can read the previous piece she wrote here.)

A special thank you to Carol Chong (Thompson & Chong), Jory Des Jardins (Candor Partners), Dan Jenkins (Product Hub by MMR), Bronwen Aldridge (Alliance Pharmaceuticals), and Abran Maldonado (Create Labs) for contributing the practical insights, real-world experiences, and accountability frameworks featured in this piece.


An AI-generated recommendation often looks boardroom-ready long before anyone checks whether it can actually survive contact with the business.

The deck is polished, the logic sound, and the next steps neatly organized. Then someone with institutional experience spots the missing constraint, the faulty assumption, or the strategy tried two years ago without success.

That moment captures a central thread running through IIEX.AI: as AI accelerates research output, the human judgment surrounding that output becomes exponentially more consequential.

Speakers approached the issue through product testing, consumer immersion, strategy, and accountability. Together, their arguments offer a more useful response to replacement anxiety than a general reassurance that humans will always matter. They identify specific contributions researchers need to protect, practice, and make visible.

The question is where human involvement changes the quality of the evidence or the decision—and how researchers organize their work around those moments.

Who Decides Whether We’re Asking the Right Question?

A research brief often arrives with a method already attached: run a survey, test a concept, compare the options. The work begins by understanding what the business needs to decide and whether the proposed research will help.

AI can assist with that interrogation. It can suggest alternative questions, organize existing knowledge, and expose gaps in a brief. Choosing which gap matters requires a view of the organization’s priorities, constraints, and appetite for risk.

In “Closing the Gap Between Strategy & Research,” Carol Chong, Co-Founder of Thompson & Chong, put that responsibility directly:

“And the one big thing to remember about the role of AI is that it will help you interrogate the question behind the question. But ultimately, you are the one that’s going to decide what matters most.”

Consider a team asking which product concept consumers prefer. The more consequential question might be which concept addresses an unmet need the company can credibly serve. Preference still matters, but it belongs within a larger decision about opportunity and feasibility.

Researchers add value by recognizing that distinction early enough to shape the work. Faster execution makes that intervention especially important: a poorly framed question can now generate a persuasive answer with very little friction.

Who Recognizes What the Answer Is Missing?

Judgment also enters after the output arrives. A recommendation can be internally coherent while overlooking something essential about the market or the organization expected to act on it.

In “What AI Can’t Do (And Why That’s Your Job Now),” Jory Des Jardins, Founder of Candor Partners, described using AI to accelerate strategy development, then recognizing proposals that her experience told her would not work. Previous campaigns and execution experience supplied context that the output lacked.

For researchers, this is a familiar responsibility. An unexpected result might signal a meaningful change, a poorly understood subgroup, or a problem with how the question was asked. Deciding which explanation deserves investigation takes methodological knowledge and attention to context.

The practical contribution is the next question: What would have to be true for this conclusion to hold? What evidence contradicts it? What information did the analysis never have?

This kind of scrutiny needs space in the workflow. If every time saving becomes a shorter deadline, researchers may have less opportunity to exercise the judgment that makes the output useful.

What Requires an Actual Human Experience?

Product research makes the limits of inference especially tangible. Describing a texture, predicting a taste preference, and observing someone use a product are different sources of knowledge.

In “Robots Don’t Get Goosebumps,” Dan Jenkins, Global Head of Product Hub at Product Hub by MMR, identified the distinction:

“AI can imagine a product but only people can experience one. There’s an irreducible human core to product research and development, and that’s where this kind of generic AI in the research thesis stops working.”

The physical experience extends beyond a formulation or specification. A handle can feel awkward despite looking ergonomic. A texture can disappoint despite an appealing description. These are illustrative examples of questions that require people to encounter the product itself.

Researchers must decide which parts of that experience need direct observation and how to capture them without stripping away the context that gives them meaning.

Jenkins and Alex Dobromir’s session also described a supporting role for specialist AI: helping interpret collected data and connect learning across studies. That approach depends on a continuing source of human evidence. Its usefulness rests partly on the quality of the experiences researchers choose to study.

What Do Researchers Learn by Being There?

Consumer immersion reveals another kind of evidence: the context people may never think to report. Participants can explain their behavior, but familiar routines and surroundings may be so ordinary to them that they go unmentioned.

In “Fueling Alliance Pharmaceutical’s Next Wave of Innovation,” Bronwen Aldridge, Global VP Consumer Insight & Analytics at Alliance Pharmaceuticals, described what visiting consumers’ homes contributed:

“We then went out into real consumers’ homes… And that of course has the double benefit of providing what only true consumer immersion does, the rich context and understanding and ultimately the insights that the consumers just can’t tell you. I learned a lot about China just from being in someone’s apartment and there’s no way that that lady could have told me the insights that I gleaned.”

The observation matters because a person’s environment contains information they may take for granted. Researchers can notice a workaround, an unexpected storage arrangement, or a mismatch between a stated priority and a daily routine. Those observations become starting points for further questioning, rather than conclusions that should go untested.

Who Takes Responsibility for the Recommendation?

Eventually, the research must support a choice. Someone has to explain what the evidence supports, where uncertainty remains, and what the team recommends doing next.

Researchers remain responsible for explaining and defending recommendations informed by AI. As Jory Des Jardins said in “What AI Can’t Do (And Why That’s Your Job Now)”:

“AI can produce an answer. It can’t own what happens if it’s wrong. Only you can, and if your name is going to be on that research, if your name is going to be on those insights, that is what you need to ensure you are in charge of.”

For insights professionals, ownership means being able to defend the path from evidence to recommendation. It includes separating findings from interpretation, explaining material limitations, and knowing when more research would change the decision.

It also requires a willingness to take a position. An exhaustive list of possibilities can leave stakeholders with the same uncertainty they brought to the project. A useful recommendation weighs those possibilities against the business context and makes the reasoning clear.

Make the Human Contribution Visible

The case for researcher value becomes stronger when it is specific: reframing a brief to address the right problem, capturing a physical experience through product testing, or investigating an observation that changes how the team interprets a finding.

In “The Human Advantage: Why AI Makes You More Valuable—Not Less,” Abran Maldonado, Founder of Create Labs, described the contribution that extends beyond a deliverable:

“You are more than your output. You are what you notice, the questions you ask, the relationships you build, the experiences you bring, and the responsibility you take. All of that falls on you.”

For researchers, developing that value means strengthening business understanding alongside methodological expertise. It means practicing observation, challenging interpretations, and learning to make recommendations that stakeholders can assess and act on. Those skills need deliberate use, including when AI can produce a plausible first pass.

Teams can support that development by using some of the time saved on production for consumer contact, deeper scrutiny, and discussion with decision-makers. Researchers also need to explain how those activities improved the work.

As AI makes research easier to produce, researchers need to make their judgment more visible: what they questioned, what they noticed, and how it changed the decision.

Comments

Comments are moderated to ensure respect towards the author and to prevent spam or self-promotion. Your comment may be edited, rejected, or approved based on these criteria. By commenting, you accept these terms and take responsibility for your contributions.

Ashley Shedlock

Ashley Shedlock

Content Producer at Greenbook

101 articles

author bio

Disclaimer

The views, opinions, data, and methodologies expressed above are those of the contributor(s) and do not necessarily reflect or represent the official policies, positions, or beliefs of Greenbook.

More from Ashley Shedlock

Beyond Bots: Rethinking Fraud Prevention in Research
Data Quality, Privacy, and Ethics

Beyond Bots: Rethinking Fraud Prevention in Research

Learn how layered fraud detection and transparent risk assessments help research teams evaluate samp...

How to Choose a Market Research Firm: A Buyer’s Checklist
Research Methodologies

How to Choose a Market Research Firm: A Buyer’s Checklist

Learn what to ask a market research firm, which red flags to watch for, and how to compare proposals before choosing a partner.

The Governance Gap: Is Trust in AI Outpacing the Evidence?
Artificial Intelligence and Machine Learning

The Governance Gap: Is Trust in AI Outpacing the Evidence?

Explore why AI confidence can outpace validation and how research teams can strengthen oversight, accountability, and trust in AI outputs.

Synthetic Research: When Are Findings Decision-Ready?
The Prompt

Synthetic Research: When Are Findings Decision-Ready?

Learn how researchers validate synthetic findings, test their limits, and determine when results can support business decisions.

What Happens When a Persona Stops Being a PowerPoint Slide?
The Prompt

What Happens When a Persona Stops Being a PowerPoint Slide?

Can AI persona bots deliver trustworthy insights? Explore why data quality, behavioral evidence, and human validation matter more than realism.

Abhinav Kothari

Abhinav Kothari

Chief Information & Technology Officer at Escalent

Trust as a Global Currency: Why the ICC-Esomar Partnership Matters More Than Ever
Artificial Intelligence and Machine Learning

Trust as a Global Currency: Why the ICC-Esomar Partnership Matters More Than Ever

How can market research maintain trust in the age of AI? Andrew Wilson explores the role of ethical standards and the ICC-ESOMAR partnership.

Andrew Wilson

Andrew Wilson

Deputy Secretary General at International Chamber of Commerce (ICC)

Lessons from the Past: Why AI Won't Break Market Research
The Prompt

Lessons from the Past: Why AI Won't Break Market Research

What can past disruptions teach us about AI in market research? Explore why industry change can create opportunities for growth and innovation.

Stephan Sigaud

Stephan Sigaud

Chief Marketing Officer at Phase 5

Not All Synthetic Data Is Created Equal
Artificial Intelligence and Machine Learning

Not All Synthetic Data Is Created Equal

What's the difference between ML and LLM-generated synthetic data? Julian Highley explains how each approach impacts research reliability.

Julian Highley

Julian Highley

EVP, Global Data Science and Product at MarketCast

Sign Up for
Updates

Get content that matters, written by top insights industry experts, delivered right to your inbox.

67k+ subscribers