Agencies

Successful AI experimentation depends on being the scientist, not the lab rat

Jul 8, 2025

Paramark News Desk

Credit: Outlever

Key Points

  • AI is often implemented in customer service without a clear plan, resulting in ineffective workflows and diminished control.

  • DISQO's David Karp advocates for intentional AI use, prioritizing human problem-solving over mere efficiency.

  • AI initiatives at DISQO are approved only if they provide tangible value, such as revenue or efficiency gains.

  • Karp's cautious approach to AI includes starting with low-risk projects and maintaining human oversight to protect clients.


When you’re in a world of experiments, you get to choose: Am I the scientist or the lab rat? We’re going to keep being the scientists.

David Karp

Chief Customer Officer
,
DISQO

Too many teams are rushing AI into CS without a clear hypothesis. It’s chaos, not experimentation, and the result isn’t just bad workflows. It’s a loss of control, with AI calling the shots instead of the people meant to guide it.

David Karp is the Chief Customer Officer at DISQO, a leading ad effectiveness measurement platform. For him, cutting through the noise requires a "ruthlessly intentional" mindset.

Lab coats only: “When you’re in a world of experiments, you get to choose: Am I the scientist or the lab rat? We’re going to keep being the scientists,” says Karp. Being the scientist in an era of constant tech upheaval means taking control, setting the terms, and using AI with intention. The alternative is to become a passive subject, reacting to change instead of directing it. In his view, the choice is clear: lead the experiment, or be shaped by it.

AIs and ears: Karp’s first principle flips the script: AI's primary role isn't boosting efficiency, but solving a deeply human problem most businesses fail at. "If I'm a customer, I want to know that everything we've ever told you, you've heard," he explains. "But what's really hard is to listen to all of your customers. You just run out of human bandwidth."

His first foray into AI wasn't about cutting costs, but about creating an organizational memory backed by a structured framework. His team uses AI to mine every conversation for four critical categories—customer risk, product feedback, advocacy opportunities, and growth signals—turning raw data into an intelligence engine that funnels insights across the business.

Our clients are solving real problems, and we have to protect that. If we rush into AI the wrong way, we’ll disrupt their work—and that’s not okay.

David Karp

Chief Customer Officer
,
DISQO

Earning its slot: When it’s time to build new AI workflows, Karp is strict. “I could apply it in a zillion places, but where will it actually drive real impact?" he asks. His team ranks projects based on one rule: no dreamy ideas, no theoretical value. To earn a green light, every AI initiative must deliver one of three things: direct revenue, indirect revenue, or meaningful efficiency. “Our goal is tangible value, not just the ability to say, ‘Hey, we use AI.’ Who cares, really?”

Caution meets conviction: Karp brings the same discipline to how AI projects are rolled out. His team starts with low-risk automations—no client touchpoints, no room for error—and only moves to customer-facing workflows once the value is proven. Every step includes a human-in-the-loop safeguard to pause, review, and approve before anything reaches a client.

It’s a balance of experimentation and care. “We’re going to try things, and we’re going to get some things wrong,” Karp says. “That’s why the human’s there—to catch it before it impacts the client.” Because on the other end, he reminds, are people under real pressure. “Our clients are solving real problems, and we have to protect that. If we rush into AI the wrong way, we’ll disrupt their work—and that’s not okay.”

Agencies

Successful AI experimentation depends on being the scientist, not the lab rat

Jul 8, 2025

Paramark News Desk

Credit: Outlever

Key Points

  • AI is often implemented in customer service without a clear plan, resulting in ineffective workflows and diminished control.

  • DISQO's David Karp advocates for intentional AI use, prioritizing human problem-solving over mere efficiency.

  • AI initiatives at DISQO are approved only if they provide tangible value, such as revenue or efficiency gains.

  • Karp's cautious approach to AI includes starting with low-risk projects and maintaining human oversight to protect clients.


When you’re in a world of experiments, you get to choose: Am I the scientist or the lab rat? We’re going to keep being the scientists.

David Karp

Chief Customer Officer
,
DISQO

Too many teams are rushing AI into CS without a clear hypothesis. It’s chaos, not experimentation, and the result isn’t just bad workflows. It’s a loss of control, with AI calling the shots instead of the people meant to guide it.

David Karp is the Chief Customer Officer at DISQO, a leading ad effectiveness measurement platform. For him, cutting through the noise requires a "ruthlessly intentional" mindset.

Lab coats only: “When you’re in a world of experiments, you get to choose: Am I the scientist or the lab rat? We’re going to keep being the scientists,” says Karp. Being the scientist in an era of constant tech upheaval means taking control, setting the terms, and using AI with intention. The alternative is to become a passive subject, reacting to change instead of directing it. In his view, the choice is clear: lead the experiment, or be shaped by it.

AIs and ears: Karp’s first principle flips the script: AI's primary role isn't boosting efficiency, but solving a deeply human problem most businesses fail at. "If I'm a customer, I want to know that everything we've ever told you, you've heard," he explains. "But what's really hard is to listen to all of your customers. You just run out of human bandwidth."

His first foray into AI wasn't about cutting costs, but about creating an organizational memory backed by a structured framework. His team uses AI to mine every conversation for four critical categories—customer risk, product feedback, advocacy opportunities, and growth signals—turning raw data into an intelligence engine that funnels insights across the business.

Our clients are solving real problems, and we have to protect that. If we rush into AI the wrong way, we’ll disrupt their work—and that’s not okay.

David Karp

Chief Customer Officer
,
DISQO

Earning its slot: When it’s time to build new AI workflows, Karp is strict. “I could apply it in a zillion places, but where will it actually drive real impact?" he asks. His team ranks projects based on one rule: no dreamy ideas, no theoretical value. To earn a green light, every AI initiative must deliver one of three things: direct revenue, indirect revenue, or meaningful efficiency. “Our goal is tangible value, not just the ability to say, ‘Hey, we use AI.’ Who cares, really?”

Caution meets conviction: Karp brings the same discipline to how AI projects are rolled out. His team starts with low-risk automations—no client touchpoints, no room for error—and only moves to customer-facing workflows once the value is proven. Every step includes a human-in-the-loop safeguard to pause, review, and approve before anything reaches a client.

It’s a balance of experimentation and care. “We’re going to try things, and we’re going to get some things wrong,” Karp says. “That’s why the human’s there—to catch it before it impacts the client.” Because on the other end, he reminds, are people under real pressure. “Our clients are solving real problems, and we have to protect that. If we rush into AI the wrong way, we’ll disrupt their work—and that’s not okay.”