The Risk of Letting AI Get Between You and Your Customer

Following my previous essay, I received a number of thoughtful comments and other verbal feedback -- and a few that really impressed me.

One, in particular, stood out. Nancy Levy shared a story from an interview with a patient living with Pompe disease. From a clinical standpoint, we understand all of the clinical facets of the condition: progressive muscle weakness, physical limitations.

But what stayed with them wasn’t the symptom profile.

It was the moment when that patient described being yelled at in a Home Depot parking lot; he was called ‘lazy’ for not helping his wife load the car. The embarrassment. The frustration. The loss of dignity. These are the moments that rarely show up in datasets or get elevated in patient journey frameworks -- but they are essential to understanding the full, lived experience of disease.

And it’s a useful reminder as we think about where AI fits into all of this.

‍

Moving Beyond the Wrong Question

Since publishing that first essay, I’ve been asked a version of the same question several times:

If AI isn’t going to replace primary research, then what is it going to do?

It’s the right question: but I think we’ve been framing it the wrong way. This isn’t about replacement.

It’s about redistribution.

AI is fundamentally shifting where value is created in the research process. It compresses the lower-value, repeatable tasks and elevates the areas that require judgment and interpretation. But if we’re not deliberate, we risk over-investing in what’s easiest to automate rather than what actually matters. My colleague Himavanth Chandra, MBA has been ahead of this curve for years, helping to clarify where AI meaningfully enhances our work -- and where human insight remains essential.

‍

Where AI Is Already Making a Difference

There’s no question that AI is already reshaping how we work. It can synthesize enormous volumes of historical data in seconds. It can connect dots across past studies, CRM data, and social listening in ways that would have taken weeks. It can generate hypotheses, customer segmentation solutions, and strategic possibilities at a scale that simply wasn’t feasible before.

In that sense, AI is becoming a powerful starting point. It helps us quickly address the seminal question: What do we already know?

And that’s incredibly valuable.

Because it allows us to move faster past the basics, and spend more time on the questions that actually require human judgment.

‍

But Insight Is Not Just Information

Another comment on my last post from Susan Schwartz McDonald captured this perfectly:

AI may serve up synthetic calories -- but it should whet the appetite for real data.

That’s spot on.

AI is very good at summarizing information. But insight, genuine strategic insight, is something very different.

It comes from understanding history and context. From hearing hesitation in someone’s voice. From understanding all of the micro-moments in a medical encounter. From a skilled moderator asking a follow-up question that wasn’t on the discussion guide due to her gut feel.

AI can tell you that patients are dropping off a therapy. It cannot sit with a patient long enough to understand what that experience actually feels like.

And in pharma, that distinction is relevant.

Because decisions are rarely driven by data alone. They’re shaped by emotion, instinct, trade-offs, and lived experience, often in ways that aren’t immediately visible, even to the patient themselves.

‍

Primary Research as a Reality Check

One way I’ve started to think about this is as a form of grounding. AI can generate possibilities. Primary research tells you which of those possibilities are real.

And in that sense, it serves as something essential:

A reality check.

In an industry where the cost of getting it wrong is high -- not just financially, but in terms of patient outcomes -- that grounding matters.

Because the more we rely on modeled outputs, the easier it becomes to drift, even slightly, away from what’s actually happening in the real world.

And that’s where the risk begins.

Any time we introduce distance between ourselves and the customer, particularly in a human, patient-centric industry like pharma, we increase the likelihood of missing something that matters.

‍

Rethinking the In-House Conversation

This also changes how we think about bringing research in-house. On paper, it sounds efficient, especially with AI handling more of the upfront work. But there’s a different question worth asking:

What is the highest and best use of an insights professional’s time? Or: how can we allow our experienced insights professionals to operate "at the top of their license?"

• Is it running analyses and managing tools?
• Or is it sitting with stakeholders, shaping decisions, and translating insight into action?

The more we pull internal teams into execution, the less time they have for influence.

And that’s where external partners continue to play a meaningful role, not just for objectivity, but for allowing internal teams to stay focused on the decisions that actually move the business forward.

‍

From “What” to “So What?”

If AI is changing anything fundamentally, it’s this: The value of simply reporting “what is happening” is declining. That information is becoming faster, cheaper, and more accessible.

Where value is increasing is in answering the next question:

So what does this mean...and what should we do about it?

AI can help us get to the “what” faster than ever.

But getting to the “so what,” and doing it with confidence, still necessarily requires human judgment, context, and often, direct engagement with the people behind the data.

‍

Growing the Pie

There’s a tendency to view AI through a cost lens: a way to do the same work faster and more efficiently. A more productive way to think about it is as an opportunity.

If AI allows us to compress weeks of data synthesis into hours, then the question isn’t:

How do we spend less?

It’s:

How do we go deeper?

How do we ask better questions? How do we explore areas we previously didn’t have time for, given the constraints of our data collection methods, etc.? How do we get closer to the human experience behind the data?

That's ultimately where the real value is created.

‍

What We Can’t Afford to Lose

There’s one final point that feels increasingly important. AI is democratizing access to information at an incredible pace. Soon, everyone, including competitors, clients, even patients, will have access to similar tools, similar summaries, and similar “intelligence.”

When that happens, information alone is no longer a differentiator. What matters increasingly is how well we understand people.

The stories we hear. The questions we ask. The context we bring.

That’s not something we should be trying to automate away. If anything, AI should push us to value it more.

‍

The Implications for How We Work

So no, AI is not going to replace primary marketing research. However, it is going to change how we use it. It will push us to move faster, focus more deliberately, and spend less time gathering information -- and more time understanding it.

And if we use it well, it won’t distance us from the customer. It will make us more intentional about staying close.