Most AI conversations start at the front office. Sarika Garg starts in the basement. The HubSpot Revenue Hub VP and former Cacheflow CEO has a blunt diagnosis: your AI is stalling because your data is a mess, and no amount of clever prompting fixes dirty infrastructure. Clean that up first, then talk.
Mindstream: If you had to explain your AI philosophy in a single sentence, what would it be?
Sarika Garg: For me, the promise of AI was never faster software. It's that humans finally stop behaving like machines and get to do the things machines can't: imagine, judge, and craft with taste.
Everyone talks about AI helping marketers and developers. Which back-office workflows are being overlooked?
The front-office mess is usually just the back-office mess showing up downstream. Duplicates, dirty data, disconnected systems, processes no one ever mapped end to end. This is exactly what AI is great at: untangling data messes and using the underlying architecture to produce clean, contextual, connected data at scale. Front-office AI stalls until that's fixed. The best companies already know this and are going after the source, not the symptom.
What job in revenue operations do you expect will change the most over the next five years because of AI?
The deal desk and reporting analyst. Today the job is mostly mechanical. Building and approving quotes, joining data into reports, re-slicing them per stakeholder, cleansing data, forecasting in Excel, and assembling board reports on ARR, NRR, LTV and GRR. AI absorbs nearly all of that. The role shifts from producing the numbers to architecting the data model and revenue system underlying them, and using judgment to ensure they're effective.
Which AI use case has blown your mind the most?
As a leader, having revenue context at my fingertips has been game-changing. At any moment, I know which deals are in play, which quotes are out, what contracts got signed and on what terms, what's renewing, what's at risk, where the expansion opportunities are, and which payments are delayed. This used to be a quarterly exercise. The moment we saw an upmarket pull in the data, we changed the product's direction to focus on upmarket features. That's sense-and-respond in the truest sense.
As a founder, what signal tells you a company is building real AI value versus just adding AI features?
Look at where the AI sits. Helping an individual work faster is a feature. Real value is when AI is woven into the business process itself and can run jobs autonomously, not just assist one person. Features speed people up. Real AI changes how the work happens.
What advice would you give someone early in their career who wants to build a company at the intersection of business and AI?
Start with the pain, not the technology. Find a genuinely painful business problem, then ask how AI could solve it 100x or even 1000x better, not 10% better. And to get that kind of result, the AI almost always has to do a full job, not just assist with a task.
Which human skill do you think is becoming more valuable in the AI era?
Systems thinking. As AI absorbs the tasks, humans move from driving the work to architecting it, and that takes seeing how the whole system connects rather than just executing one piece. We become architects of the business, not task executers.
What's the biggest misconception people have about how AI will reshape the world we live in?
That we'll all become jobless. It's silly. Every job will change, yes, but that's not the same as disappearing. I think we're heading into a golden era with more businesses being built and more interesting work for humans, because when the busywork goes away, what's left is the judgment, creativity, and building that people actually want to do.
What's a task you personally spend less time on today because of AI?
Gathering and synthesizing signals about how the business is actually doing. I have a series of agents that listen across everything: community chatter, CSAT, support tickets, prospect calls in Gong, and the CRM, then summarize it for me every single day. That collecting and summarizing used to take ages and could only happen occasionally. Now it's a daily read, and I spend my time acting on the signal instead of assembling it.
Finish this sentence: In five years, AI will make people feel ___ about their work. Why?
Joy. In five years, the mindless tasks will be gone, and what's left is the strategic, interesting, creative work, the part people actually find meaningful.
When you stop behaving like a machine and spend your day thinking, building, and using judgment, work starts to feel less like a grind and more like the thing you'd choose to do anyway.
Sarika Garg is GM and VP of Product for Revenue Hub at HubSpot. She joined HubSpot in November 2024 after it acquired Cacheflow, the Quote-to-Revenue software startup she co-founded and led as CEO.
She also serves as a board director at OneSpan. Sarika credits her global experiences - growing up in Africa, studying and working in India and Germany - as key to her success. She actively invests in and mentors startups.
You can get our 10 prompts for strategic deep research below!