Dev Nag is the Founder and CEO of QueryPal, where he’s building AI systems that transform customer support economics, enabling teams to deliver faster resolutions while dramatically reducing operational costs. Previously, he founded Wavefront, a monitoring and analytics company that provided real-time insights into cloud infrastructure and applications, leading it as Founder and CTO until its acquisition by VMware in 2017. At VMware, Dev served in the Office of the CTO, developing and launching VMware’s flagship AIOps product.
Before Wavefront, Dev was the founding engineer at GLMX, the leading electronic securities trading platform for money markets, and held senior engineering roles at Google, PayPal, and eBay. He holds more than a dozen patents in AI/ML and security, published academic papers in computational biology and medical informatics at Stanford, and is an active investor and mentor to startups in AI, technology, and biotech.
Dev sits down with our Technology and Software Partner, Robert Dunn, to share game-changing perspectives on AI leadership.
How do you define great leadership and how has that evolved throughout your career?
Great leadership is about creating the conditions where people can do their best work while feeling genuinely valued for who they are. Early in my career at Google and PayPal, I thought leadership meant having the smartest technical solutions. Building Wavefront taught me it’s actually about understanding what motivates each person on your team and connecting their individual aspirations to something larger.
The biggest evolution has been recognizing that the best leaders are translators. At QueryPal, I spend most of my time translating between different worlds: helping engineers understand customer pain points, helping customers see technical possibilities, helping investors grasp market dynamics. Leadership becomes less about being the smartest person in the room and more about being the best connector of ideas and people.
What is the best example of ‘Transformational Leadership’?
I think about Andy Grove at Intel during the transition from memory chips to microprocessors. He didn’t just pivot the company’s strategy but fundamentally changed how Intel thought about itself. Instead of being in the memory business, they became a company that enabled the personal computer revolution.
What made this transformational was Grove’s ability to help people see that their core skills weren’t tied to a specific product but to a way of thinking about precision, innovation, and manufacturing excellence. The engineers who made memory chips discovered they could apply that same rigor to processors. That’s true transformation: when people realize their capabilities are bigger than they imagined.
What is the most important thing you have learnt in your career?
The most valuable insights come from the gaps between what people say and what they actually do. When we were building Wavefront, customers would tell us they wanted more dashboards and metrics. But observing their actual behavior, we realized they were drowning in data and desperately needed the system to tell them what to pay attention to.
This led to our focus on intelligent alerting and anomaly detection, features customers didn’t explicitly ask for but were exactly what they needed. At QueryPal, we see the same pattern. Support teams often say they want better knowledge bases, but what they really need is AI that understands the intent behind customer questions and directly accesses recent tickets to provide an up-to-the-minute, accurate response.
Was your path to leadership strategic, opportunistic, or a bit of both? What key moments shaped your journey?
Completely opportunistic, disguised as strategy in hindsight. I never planned to start companies when I was a math major in undergrad. I just kept encountering problems that seemed solvable with better technology.
A pivotal moment was being a founding engineer at GLMX. Being part of building something from zero taught me that most “impossible” problems are actually just coordination problems. When you’re processing billions in daily trading volume, you can’t afford to have humans manually reconciling transactions. You have to build systems that think ahead of human needs.
That experience shaped my approach to every subsequent company: start with the human workflow, identify where people are spending time on tasks that machines could handle better, then build AI that amplifies human judgment rather than replacing it.
What type of people do you like to work with, and what makes them good leaders?
I gravitate toward people who are naturally curious about why things work the way they do, combined with a deep empathy for user frustration. The best engineers I’ve worked with don’t just solve technical problems; they feel genuinely bothered when someone has to do repetitive work that a computer could handle.
Good leaders share this quality of productive dissatisfaction. They notice friction that others accept as normal. At Wavefront, our best product decisions came from engineers who were personally annoyed by having to manually investigate system outages. At QueryPal, our breakthrough insights come from people who feel genuinely frustrated when a customer has to wait hours for a simple answer.
What fascinates you about your job?
I get to work at the intersection of human psychology and algorithms (not so coincidentally, those are my two degrees — psychology and mathematics!) Every day, I’m thinking about how to make AI systems that understand not just what people are asking, but what they’re trying to accomplish.
The fascinating part is that the hardest problems aren’t technical; they’re anthropological. How do you build an AI that understands when a customer is frustrated versus confused? How do you create systems that know when to escalate to a human versus when to provide more information? These questions require understanding human nature as much as understanding algorithms.
What strategic role do you see AI playing in your business over the next 3–5 years, and how are you preparing your teams for it?
AI will fundamentally reshape how businesses think about customer relationships. Right now, most companies treat customer support as a cost center — something to minimize. AI enables a complete inversion: customer interactions become intelligence-gathering opportunities that drive product development and revenue growth, and customer support owns the most valuable data in those customer conversations.
At QueryPal, we’re preparing teams to think like context engineers rather than traditional software engineers. This means understanding how to bridge the gap between human intent and machine capability. We’re teaching people to recognize the unspoken assumptions that humans make but that AI systems miss: the cultural context, emotional subtext, and implicit priorities that never make it into training data.
In your experience, where does AI add the most value, and where are its limitations most clear?
AI excels at pattern recognition across massive datasets and maintaining consistent quality at scale. Where it struggles is with context that humans take for granted: the unstated preferences, cultural nuances, and situational constraints that shape how people actually want problems solved.
The biggest value comes from designing AI systems that recognize their own limitations and know when to escalate to humans who can provide that contextual understanding.
How are you addressing potential AI-related risks and challenges in your organisation?
The biggest risk isn’t AI making mistakes; it’s AI being confidently wrong in ways that erode customer trust. We’ve built what I call “confidence calibration” into our systems. Our AI doesn’t just provide answers; it provides confidence scores that determine how much human oversight each response needs.
We also maintain human-in-the-loop workflows for edge cases, but more importantly, we use those edge cases as training opportunities. Every time our AI encounters something it can’t handle well, we analyze what contextual information a human would need to solve that problem, then figure out how to make that context more accessible to the AI system.
In what ways has AI already transformed your industry, and what changes do you anticipate will be most disruptive in the next 18-24 months?
AI has already commoditized basic customer service interactions. The companies still routing simple questions to human agents are burning money and frustrating customers who expect instant responses.
The next wave will be AI systems that understand customer intent across entire journeys, including in the product itself, not just individual interactions. Instead of answering isolated questions, AI will start recognizing (right in the product, itself!) when a customer’s questions indicate they’re considering an upgrade, experiencing a technical issue, or planning to churn. This shifts customer support from reactive problem-solving to proactive relationship management. Within a few years, customers will expect every product and service to be self-supporting, only having to check a help center or email support for edge cases.
How do you foster a culture of continuous learning within your teams?
We treat every customer interaction as a learning opportunity. Our platform doesn’t just review failed AI responses, but also the successful ones. Understanding why the AI got something right helps us replicate that success in new contexts.
We also run “context discovery sessions” where we examine customer interactions that required human intervention and identify what contextual clues the human used that the AI missed. This makes the whole team better at recognizing the gap between human and machine understanding, which is critical for building more effective AI systems.