Teamday
← Back

Ron Gabrisko: How Databricks Rebuilt Sales Around AI

SaaStr AI
agentsbusinessenterpriseai-sdrinterview

How Databricks CRO Ron Gabrisko Runs an AI-First Sales Team

Ron Gabrisko joined Databricks as chief revenue officer when it had under $1 million in ARR. On the SaaStr AI podcast, host Sam Blond (co-founder of AI-native sales platform Monaco and former CRO of Brex) introduces Databricks as a $200 billion company with many billions of ARR. Two years after their last conversation, Blond asks what AI has actually changed in a sales organization that size. Gabrisko's answer is concrete: hiring tests, team ratios, reply rates and a pricing mistake he fixed.

On hiring sellers who build: Databricks runs Databricks on Databricks, and its account executives use Genie, the company's own AI product, to build custom demos for accounts like Pepsi or Rivian. The interview process tests this live: "here's Genie, here's access to it, build us a demo, and then walk us through the demo". Candidates also have to deliver an industry-specific, CEO-level AI pitch for an account they choose, and reps keep certifying on pitches and Genie demos throughout the year.

On a flat SDR team: Databricks keeps adding AEs and solutions architects, but SDR headcount has stayed mostly flat because agents now handle account research, contact selection, personalized outreach, exec prep docs and proposals. Gabrisko puts the reply-rate change in numbers: the old "automated spammer" approach got 1-2% response rates, and personalized outreach now gets 6-10%. He still won't cut the SDR team to zero, because it is how junior sellers learn to sell before becoming AEs.

On time with customers: Automation buys selling time. Gabrisko pushes for at least 60% of time in front of customers, with a simple rule: prep and admin on Monday and Friday, customer meetings Tuesday through Thursday. The buyer has changed too. AI is now a CEO-level initiative, so Databricks sells to CFOs, CMOs and business leaders, not just IT.

On closing with working software: "I think the day, the old day of the show up with a PowerPoint is gone." Databricks closes by building the use case first, such as stock-out prediction for a retailer, on industry-realistic dummy data, then deploying it with a forward deployed engineer motion. Blond describes the same shift at Monaco, where every account executive works like a forward deployed engineer and offers to deploy what they built.

On usage pricing: Databricks has always charged on consumption, and Gabrisko says that is why agents are a tailwind: millions of agents will query data, not just people. His pricing lesson came from a mistake. Customers ramped slowly because of per-user fees, one telling him every user cost a thousand dollars. He made users free, and "consumption went out the roof once I did user base for free."

5 Lessons from Databricks on AI-First Selling

  • Test for building, not just talking - Have sales candidates build a working demo with your product during the interview, then certify reps on it all year.
  • Grow closers, not prospectors - At Databricks, agents absorbed outbound research and personalization, so AE and SA headcount grows while SDR headcount stays flat.
  • Personalization pays in reply rates - Moving from mass outbound at 1-2% to agent-personalized outreach at 6-10% is the measurable gain.
  • Show the outcome, not the slides - Build the customer's use case on realistic data and offer to deploy it; business buyers trust what they can see.
  • Price on the unit that scales with value - Pick one base unit, add multipliers for premium features, avoid seat fees and big upfront commits, and let customers start pay-as-you-go.

What Databricks' Playbook Means for AI-Era Sales Teams

Gabrisko's view is not that AI replaces sellers but that it moves them up the stack: agents do the research and outreach, and humans spend the freed time with executives, building and deploying working solutions. For companies selling into the agent era, his pricing point may matter most: if agents will be your heaviest users, seat-based pricing caps your growth. See also AI SDRs, forward deployed engineers and agent pricing.