Databricks Holds Its SDR Team Flat as AI Outbound Response Hits 6-10%
Databricks CRO Ron Gabrisko says AEs and solutions architects kept growing while the SDR team stayed roughly flat as a training bench. He puts outbound response at 6-10%, up from 1-2% on mass sends.
TL;DR
- Databricks CRO Ron Gabrisko says SDR headcount stayed roughly flat while AEs and solutions architects grew, and outbound response rose to 6-10%.
- Hiring went to closers and technical sellers, while the SDR team stays on as the training ground for future AEs.
- Sales leaders planning 2027 should set the SDR-to-AE ratio deliberately and test any AI reply rate against list warmth first.
Databricks CRO Ron Gabrisko was the guest on episode 884 of the SaaStr podcast, and SaaStr's show notes say the company's SDR headcount stayed flat while its closing org scaled. In SaaStr's October 28 write-up of the interview, Gabrisko says the SDR team has stayed roughly flat while account executives and solutions architects kept growing. Outbound now runs on buying signals and personalization for each contact, and response rates have risen to 6-10% from 1-2% on the old mass loads of about 5,000 contacts. Gabrisko calls the old approach "the automated spammer." Every one of those figures is his, stated on a podcast, and he gave no SDR or AE counts and no timeframe. Databricks said on August 13 that it had passed a $7 billion revenue run-rate with more than 80% year-over-year growth, on a $190 billion valuation.
Gabrisko's account is mostly about where the headcount went. The prospecting team held still while hiring went to the people who close and the people who make the product work inside a deal. He keeps the SDR team because that is where future AEs learn the job.
Databricks treats its SDR team as the bench for future account executives
Gabrisko said AI may shrink the SDR team over time. For now he runs it as an academy whose graduates move into AE seats. AE candidates build a live demo during the interview itself, using Genie, one of the products the August release says Databricks is scaling. He puts the target for customer-facing time at 60%. The same release counts more than 1,000 customers each running at over $1 million. Accounts that size likely need sellers who can build in front of a buyer, and that's where the new headcount went.
The response-rate jump is the shakier claim. SaaStr ran AI SDRs itself and concluded that response rates were "a function of lead quality, not who was doing the outreach," with 5.5% on colder lists and 11-12% on warmer ones. Gabrisko also described signal-based targeting, so part of Databricks' lift likely comes from who receives the email. A second SaaStr report says its outbound agents had a 6.7% overall response rate and a 4% positive response rate over six months, against an industry average it puts near 3%. That puts Databricks' 6-10% in range for well-run AI outbound rather than out on its own, though Gabrisko didn't say whether he meant all replies or positive ones. Neither SaaStr dataset carries a date on the page, and SaaStr isn't independent of AI-SDR vendors.
Host Sam Blond is CEO and co-founder of Monaco, an AI sales platform, and the write-up cites Monaco's claim that one SDR books 400 outbound meetings a month. By that vendor's standard, even a flat SDR team carries too many people. Gabrisko's answer is the academy. A team sized only for meeting volume would shrink toward Monaco's number, while a team that also trains the next AE class has a second job that pays for its seats.
What this means for sales leaders setting 2027 headcount
Plan on the ratio before you plan on cuts. If AI outbound is lifting your response rates, decide how many SDRs each future AE class needs to come from, then size the team to that. If your 2027 plan takes SDRs to zero, write down where your next AEs will come from and what they'll cost to hire.
Before you scale an AI sequence, split all replies from positive replies and cold lists from warm ones, then compare against your own old baseline. Treat Databricks' 6-10% as one company's stated result on signal-targeted lists. When a vendor shows you a reply rate, ask whose list produced it and what counted as a reply.
Put new headcount where deals need building. If your buyers decide after watching the product work on their own problem, the AE and solutions-architect seats carry the revenue. At your next hiring round, copy the Databricks screen and have AE candidates build something live in the interview.
Our bet is that by the end of 2027, more SDR job postings at large B2B companies describe the role as a path to AE, with targets set on pipeline that converts rather than meetings booked. Teams that hire SDRs only to book meetings will find AI cheaper per meeting.
The prospecting team held still while hiring went to the people who close and the people who make the product work inside a deal.
Before you scale an AI sequence, split all replies from positive replies and cold lists from warm ones, then compare against your own old baseline.
