In the second part of my conversation with Dr. Ari Zelmanow, we explored some of the most pressing challenges facing researchers today: the need for speed, the democratization of research, the role of AI, and how to position yourself as a strategic advisor to the business.
Why Speed Matters in Research
One of the most common tensions I hear from researchers is the struggle with speed. The business moves too fast, stakeholders are impatient, and there’s never enough time for proper research. But Ari offers a different perspective that completely reframes this challenge.
“People think that speed and rigor are at odds,” Ari explains. “That’s not necessarily the case. What speed means is time to learning, not time to end learning.”
The traditional research model existed because the risk of being wrong was worse than the risk of being slow. Software used to be shipped in boxes and placed on shelves—once it was out there, you couldn’t update it. You had to get it right the first time.
But that world doesn’t exist anymore. Products can be updated constantly, yet research hasn’t adapted to this new reality. Instead of reimagining the process, many teams simply compressed the traditional linear academic approach. The same process that took eight weeks now takes three weeks—but it’s still plagued by the same fundamental problems.
The Real Cost of Slow Research
Ari makes a compelling point about the economics of research speed: engineering teams are a fixed cost. Whether those engineers are coding or sitting around eating bagels, the business is paying for them. When research can’t keep pace, product teams face a choice: wait for research or move forward anyway and fix things later.
We all know what usually happens—they move forward. And then research ends up working the product roadmap, always lagging behind. This creates two predictable outcomes: either you validate what’s already been decided (and product managers celebrate being right), or your findings contradict the plan (and suddenly your sample size is too small or your methods are questioned).
The Backhoe vs. The Shovel
Ari offers a brilliant analogy to illustrate the difference between old and new approaches to research:
“Imagine I buried a gold coin in my backyard and asked you to help me find it. You could take a backhoe and dig one big hole, then deliver all that dirt to me so I can sift through it to find the coin. Or you could take an abductive approach—with every shovel scoop, we learn from it. It’s not here? Cool, let’s change locations. Not here either? Let’s adjust again.”
This is what research needs to become: an ongoing learning process rather than something with a finite endpoint. Instead of one massive study that delivers results weeks later, we should be learning continuously and sharing insights as we go.
The Democratization of Research: Friend or Foe?
The democratization of research is happening whether we like it or not. But I recently came across a study showing that researchers’ fears about democratization have shifted. The top concerns are no longer about whether non-researchers should be doing research at all, but rather:
Cherry-picking results
Missing important connections in the data
Bias in presenting findings
In other words, researchers worry that stakeholders will manipulate research to support whatever story is most convenient for them.
Attacking the Wrong End of the Problem
Ari points out that we’re often attacking the front end of this problem—trying to teach people better methods and more rigorous approaches. But there’s a more fundamental shift that needs to happen in how we think about evidence and argumentation.
He breaks down three types of arguments:
Arguments of blame: “Why did you only do five interviews?”
Arguments of value: “Quantitative methods are way stronger than qualitative methods.”
Arguments of choice: “Given what we know today, where should we go in the future?”
Only the last one is truly useful. When researchers engage in blame and value arguments, we create conflict rather than collaboration. Instead, we should be focusing on evidence and decision-making.
“If you came to me with a point of view and I came to you with a point of view, we should be able to say, ‘Oh, that’s tacit knowledge—I have a low level of certainty on that evidence. I have a higher level of certainty on this. What do we need to do to move forward?’”
Making Room for Imperfect Evidence
This approach allows someone who’s had two sales conversations to still bring that evidence to the table, because it’s meaningful to them. Rather than being completely dismissive, you acknowledge it as evidence with unexamined assumptions—lower certainty, yes, but still valuable when corroborated with other sources.
This creates a dialectic rather than a debate. You share information together to reach a shared viewpoint focused on the future, not arguing about what methods should have been used in the past.
As Ari puts it: “We deal with the quality of evidence on the backend, because not all evidence is created equal. You could have a bad interview that still yields good insight, just like you could have a great interview that yields bad insight.”
AI and the Future of Research
Of course, we can’t talk about the future of research in 2025 without discussing AI. I’ve noticed the research community is pretty split—some people are scared, others see huge opportunities, and many are still trying to figure out how to use these tools in their daily work.
Ari’s perspective is refreshingly pragmatic. He frames it around the three parameters of the research program of the future: faster, cheaper, and good enough rigor.
Where AI Can Help Today
AI can create efficiencies in analysis and help make research more accessible. One example Ari mentions really resonated with me: “Academics love to write at a 12th grade level. Maybe AI could help you write at a fifth grade level, which will help your audiences better understand what you’re talking about.”
AI is also enabling cheaper solutions—searchable repositories that work like Perplexity but for your internal research, reducing operating expenses while making insights more accessible.
Looking Forward, Not Just at Today
“We could evaluate AI in terms of what it can do today for research,” Ari notes, “but I don’t think that’s really where we should be looking. It’s what is it going to do tomorrow?”
The interview tools available now aren’t as good as human interviewers—not even close. But are they good enough for some things? Sure. And they’re getting better rapidly.
The question isn’t “should we use AI or not?” It’s “how can we use it to improve in collecting, connecting, and communicating in a way that’s faster, cheaper, and with good enough rigor?”
Becoming a “consigliere” to the Business
One of my favorite concepts from Ari’s work is this idea of the researcher as consigliere—the trusted advisor who provides counsel to the business. But how do you actually become that person when you’re stuck running usability tests and deciding whether buttons should be red or green?
Ari’s answer is both simple and profound: Just do it.
Act As If You Already Belong
He references William James, the father of American psychology, who said that if you want to be seen as something, act as if you already had that quality.
“Stop asking for permission and instead ask for feedback. When you’re asking for a seat at the table, you don’t get a seat at the table. But when you pull up the chair and sit down, it’s just assumed that the seat was yours.”
Act as if you belong in those conversations. Start interjecting yourself. Show the business how your work connects to growth, value, adaptability, risk, and speed. Demonstrate how your research helps capture or keep more customers and ultimately drives revenue.
Everything Can Be Strategic
This connects to something I feel strongly about—the false dichotomy between “tactical” and “strategic” research. Ari completely dismantles this distinction:
“Using the word strategic to say this is important totally minimizes the work that everybody else does. Strategy is defined as having an intense focus on the things that matter most. All research by nature is strategic.”
A usability test might be strategic from a UX perspective, but it could also be strategic to the business in other ways. And all research is tactical too—you have to collect data, connect the dots, and communicate findings. Those are tactics.
By creating these artificial hierarchies, we’re cutting our own legs out from under us.
Three Tips for the Researcher of the Future
I asked Ari for concrete advice for researchers who want to transform their practice but don’t know where to start. His three tips perfectly encapsulate everything we discussed:
1. Collect Evidence Like a Detective
Take an abductive approach to evidence gathering. Yes, think about methods, but remember that five methods account for the majority of the work. Focus on continuous learning rather than perfect methodology.
2. Build Cases, Not Just Reports
Think like a detective building a case. How do you construct “likely event” stories? How do you take data and turn it into stories, stories into strategies, and strategies into outcomes? Build narratives that help the business move forward.
3. Become a Better Communicator
This is the number one thing you can do. The better you can communicate insights, explain problems, articulate desired outcomes, and present solutions with supporting evidence, the better you’ll be.
As Ari explains: “The better you can tell stories that show ‘I’ve identified this problem, we want to get to this outcome, we’ve got this solution as the bridge, here are the tactics and evidence to support that, and here are the risks of action and inaction’—the better you’re going to be and you’ll be seen as the expert that you really are.”
Final Thoughts
What strikes me most about this conversation is how Ari reframes challenges as opportunities. Speed isn’t the enemy of rigor—it’s about continuous learning. Democratization isn’t a threat—it’s a chance to elevate the conversation about evidence. AI isn’t replacing researchers—it’s a tool we can leverage to do our jobs better.
And perhaps most importantly, we don’t need to wait for permission to become strategic advisors to our organizations. We just need to start acting like we already are.
