10 Questions for AI leaders: Neha Monga

Neha Monga has spent her career turning emerging technology into practical products that actually get used. As AI shifts from assistant to agent, she has a clear-eyed view of what that means for the people doing the work, and what it will take to earn their trust.

Neha Monga: What excites me the most is that AI is shifting from being a passive assistant to being an active operator. Earlier, it helped with research, drafts, and answers.

Now, we are entering territory where AI can actually accomplish tasks and take meaningful work off our plate. The shift from "copilot" to "agent" is changing what it means to work.

What worries me is how quickly we may start delegating human judgment and accepting AI outputs as ground truth. If we accept outputs without interrogation, we risk becoming faster but less thoughtful. Today's AI is powerful, but it's not yet reliable enough to carry high-stakes judgment on its own.

AI is democratizing access to expertise, and that should make us demand more from AI, not less. Early on, we thought AI was an efficiency play, but now its role is to give anyone access to capabilities that require specialization (e.g. coding, content creation). That changes what's possible, not just what's faster.

Automation works best for mechanical effort, repetitive tasks that require little ongoing judgment or creativity. When work needs high judgment, original thinking, taste and creativity, humans need to be in the loop. Automating customer feedback summarization, yes. Automating product strategy and roadmap decisions - definitely not.

Taste and original thinking. AI is excellent at pattern matching and predicting plausible answers. But it struggles with messy context, judgment and discernment. That's what separates high-quality output from AI slop.

There is definitely a hype cycle, and especially a disconnect between investors' expectations and what actual users and practitioners experience. Some things that we thought would be transformational, like digital workers, are still far from reality.

There is a real, powerful underlying capability that is improving fast. Software development, marketing and sales functions have experienced a step-change that none of us could have imagined.

Trust is not a feature - it is the product.

If users have to double or triple check every output, it's not adding efficiency; it’s adding cognitive overhead. Real adoption happens only when AI is predictable, transparent and consistent, where users feel comfortable delegating meaningful work.

Insisting that AI-generated output always show reasoning, sourcing or other signals supporting the answers.

It sounds small, but this is what helped even the most skeptical customers to verify the work. And once they did, they started to trust it.

Winning is not about the tool, it's about which team actually applies AI. Teams that pull ahead are training reps to use and verify the outputs, re-wiring their workflows, and deliberately delegating tasks to AI so they can focus on higher-value creation. The teams that don't just tack it on.

The rise of highly personalized agents. Unlike traditional software, agents do not have to be one size fits all. They can be fully tuned to my preferences, workflows and even quirks.

That shifts AI from being "software we use" to being our digital extension, with big implications for productivity, identity, and how we design products.

More ambitious. Because when execution becomes easier, we create the opportunity to tackle more complex problems and audacious goals.

Neha Monga is a product executive at HubSpot who builds AI-powered technology platforms that translate emerging tech into practical, real-world products that help customers deliver work.