10 Questions for AI leaders: Scott Judson

Trust is the word Scott Judson keeps coming back to. As Director of Product for Sales Hub at HubSpot, he's spent more than 11 years watching sales technology evolve -- and he's convinced that in the AI era, earning genuine trust from both customers and colleagues isn't a nice-to-have. It's everything.

Scott Judson: I’m most excited about AI’s ability to automate traditional, mundane tasks and rote work, unlocking more human creativity. When AI handles the repetitive, operational tasks, it gives people more space to think, create, and really shine on the creative side. That, to me, is the most interesting and exciting element of AI.

What concerns me is how society will change as a result of AI — especially how we’ll distinguish what’s genuinely human and authentic versus what’s not. Figuring out how we preserve trust and authenticity in a world where AI can generate almost anything is going to be a really important societal challenge going forward.

My AI philosophy is simple: AI should be used to solve real customer problems — it’s not about the technology itself, but about solving problems more effectively and efficiently than we could before.

Over time, my thinking has evolved. Early on, there was a lot of focus across the industry on “bringing AI into the product” — almost as a box to check. But there’s been a healthy and promising shift toward asking a better question: how do we solve the customer’s problem in the most effective way possible, and where can AI act as an enabling technology to make that happen? The best solutions don’t make AI the main event — they use it quietly and powerfully to create better outcomes.

At this point, I don’t fully automate entire workflows with AI. Instead, I use it to solve what I think of as the “blank slate” problem — getting from zero to a first draft.

My typical workflow is to dictate my ideas and thoughts, especially when I’m writing emails or longer-form content. AI helps turn that into a structured draft. From there, I review, edit, and often rewrite sections — sometimes even starting over if it hasn’t captured the nuance correctly. For me, AI accelerates the starting point, but I still want to stay closely involved in shaping the final output.

Communication will always be one of the most important skills, especially in business — and I think it’s even more important in the AI era.

When it’s incredibly easy to generate large volumes of content with AI, clarity becomes a differentiator. Your authentic voice and communication style need to shine through. It’s also critical to be clear and precise, because AI-generated content can sometimes be verbose or sloppy. The ability to communicate simply, directly, and authentically is becoming more valuable, not less.

No, I don’t think we’re in a bubble.

The value of AI is already very clear, and as the technology continues to improve, that value will increasingly accrue to software, businesses, and the teams that use it well. We’re not in a speculative phase detached from reality — we’re in the early stages of a foundational technology shift.

To me, this is a moment of opportunity. It’s a chance for people to get close to the technology, experiment with it, and become comfortable using it. AI is real, it’s delivering real value, and it’s here to stay.

It’s the most important thing.

We’ve seen this firsthand while building products like the Prospecting Agent. For customers to adopt AI in their sales workflow, they have to trust what it’s generating. That shows up clearly in the data — roughly half of customers using the Prospecting Agent review the outputs before approving them for send. That tells you trust is still being earned.

Earning that trust is a big responsibility. It requires consistent quality, transparency, and a clear commitment to producing outreach that truly represents the customer well. Over time, as the outputs improve and customers see reliable results, trust grows.

On a personal level, that’s also why I don’t fully automate workflows yet. I trust AI’s potential and capability, but I still want a human in the loop. Trust in AI isn’t binary — it’s built gradually through repeated, high-quality outcomes. And in the AI era, trust is everything.

One decision that had an outsized impact was choosing to ship an early version of the Prospecting Agent before it felt perfect.

We believed there was a strong opportunity to apply AI to the prospecting workflow, but in practice, we didn’t fully know where the real value would materialize. The only way to truly understand that was to get the product into customers’ hands and learn quickly.

By shipping earlier, we learned how critical trust is. We saw in the data how often customers reviewed outputs before sending them, and it reshaped how we thought about quality, control, and the role of AI in the workflow. I’m not sure we would have learned those lessons as clearly if we had waited until everything felt polished.

In product development, I’m a big believer in iteration speed and learning velocity over perfection. Getting real-world feedback early is how you ultimately build something great. That bias toward action and learning made all the difference.

One of the best examples is company and prospect research.

Traditionally, sales reps spend a significant amount of time researching every target account — understanding the company, the decision-makers, the context, and potential opportunities. Being able to deploy an agent to research a company and contact, and generate a clear briefing or dossier on where the opportunity might be, represents massive time savings.

We’re also excited about expanding this even further — helping sales teams prioritize which companies are actually in market and sourcing the right contacts within those accounts. That kind of signal-driven prioritization replaces hours of manual effort.

On top of that, the Prospecting Agent already adds real value by drafting outreach. Taking research and turning it into thoughtful, high-quality outreach is time-consuming work. Automating that first draft while keeping the rep in control is saving customers a significant amount of time today.

For me, it’s Tesla’s Full Self-Driving capability, which I’m a subscriber to and use regularly.

It’s an incredible product because driving is a problem with an almost infinite number of variables. Every situation can be slightly different, and many are novel or unpredictable. Yet the system handles driving remarkably well. I often tell people that I trust Tesla’s Full Self-Driving more than I trust myself. The car effectively has multiple “pairs of eyes” and constant awareness of everything around it, while I only have one pair!

Seeing AI master something as complex and variable as driving fundamentally shifted my belief in what’s possible. If AI can navigate an open-world, real-time environment like that, it reinforces the idea that it can meaningfully support complex business contexts too — even when goals, constraints, and customer situations vary from organization to organization. It strengthened my conviction that AI isn’t just incremental improvement; it’s an enabling technology capable of handling real-world complexity.

AI will make work feel more like true knowledge work — less repetitive, more energizing, and more value-creating. It has the potential to help people spend more time on the parts of their jobs that actually require human judgment, creativity, and insight.

Scott Judson is a Director of Product for Sales Hub at HubSpot, where he leads product strategy for prospecting and sales engagement tools. He focuses on shaping product direction, balancing customer needs, and building solutions that help sales teams be more effective and successful in their organizations.