What great problem framing looks like in life sciences

We’ve all been there. Someone walks into a meeting with a vague ambition: “We want to be more innovative” or “We need to do something with AI.” Everyone nods, the room fills with ideas, and for a moment the energy is real.

But fast forward a few weeks and that energy has fizzled out.

Not because the ambition was wrong, but because the problem wasn’t clearly framed. There was no shared understanding of the challenge, no clarity on who it was for, or what value it was meant to deliver. And when it comes to AI in particular, this kind of vagueness is especially risky. Stakeholders may imagine AI doing everything, but unless we are solving something real, all we end up with is a shiny tool no one actually uses.

The difference between a theme and a challenge

A lot of the briefs we see are really themes in disguise. Statements like:

• “We want to help medical teams get closer to customers.”
• “We need a clearer picture of the competitive landscape.”
• “We should be doing something with automation.”

These are good signals, but they are not yet problems.

A well-framed challenge takes that theme and pushes it further. It narrows the focus and sharpens the intent. It asks:

• Who is struggling?
• What are they trying to achieve?
• What’s stopping them?

Let’s say the original theme is: “We want to help medical teams get closer to customers.”

Framed properly, it might become:

“Medical directors tell us they’re overwhelmed by scattered insights across countries and therapy areas. They’re missing trends, duplicating work, and can’t see what good looks like. How might we bring intelligence together in one place so they can build a clear global story?”

Now we are not just talking about a theme. We are solving a real problem for a real user.

How we help teams get specific

Most of our work starts with questions that are too big or too vague. And that is okay. The skill is knowing how to shrink the problem down without stripping it of meaning.

Here is how we usually do that:

🔍 Zooming in on the user

We start by anchoring the problem to a real person with a real job to do. We think about their environment, their decisions, and the pressure they are under.

When someone says, “We need better market insight,” we ask: “For who? For what decision? And when do they need to make it?”

That context changes everything.

⚠️ Unpacking failure

It is easy to focus on what a great outcome might look like. But we learn more by asking what is going wrong today.

Is there too little data, or too much? Is the issue timing? Is the data fine, but no one trusts it? When we understand the type of failure, we can shape a solution that actually fixes it.

🎯 Defining value early

Before we design anything, we ask: How will we know this is working for the user?

Not in terms of dashboards or KPIs, not yet, but in terms of impact:

• Will they save time?
• Make faster or more confident decisions?
• Spot something they could not see before?

If we cannot answer those questions, we are still guessing.

Why this matters more with AI

AI has made it easier than ever to build impressive-looking tools. You can summarize, synthesize, simulate, and generate at scale. But that same ease of building can send teams in the wrong direction quickly.

We have seen stakeholders fall in love with the idea that AI will do everything for them. But product teams need to stay grounded. AI is not a magic fix. Automation should be pointed directly at real friction in real workflows, not just at what can be automated, but what should be.

If you are not deeply familiar with what users are struggling with, it is far too easy to create something flashy that nobody needs. Or worse, something that adds complexity instead of removing it.

Good problem framing cuts through that noise. It helps teams prioritize, stay focused, and build systems that actually make a difference.

In the end, simplicity always wins

The best product work we have done has not come from the most advanced technology or the biggest ideas. It has come from teams willing to slow down, ask sharper questions, and get crystal clear on what the real challenge is.

So next time someone brings you a vague brief or wants to “do something with AI,” don’t rush to the solution.

Help them frame the problem properly. It is the most useful, and most powerful, thing you can do.