Jason Lemkin

Running around twenty AI agents now forces one team-mate to spend 10–15 hours weekly reviewing outputs, so leaders must budget real human time.

Jason urges most companies to purchase ready-made AI GTM tools rather than build in-house, as the classic software buy-vs-build logic still applies.

Jason prefers former engineers who no longer code much but obsess over the product to fill FDE roles and drive launch success.

Jason says we still need C-suite and VPs but they must roll up sleeves and handle 2026-2027 GTM skills, not rely on old playbooks.

A well-trained agent outperforms mid-pack SDRs because most humans send poorly researched, generic emails.

Keep manual, deeply researched outreach for a short list of high-value targets where deal size makes the effort worthwhile.

Jason urges picking one agentic tool, ingesting data, answering training questions daily for 30 days, and becoming hyper-employable.

Facing two-week support delays, SaaStr launched a Delphi-based agent that now handles event queries and even sells sponsorships autonomously.

A general-purpose digital Jason agent independently sold a $70k sponsorship, validating agent-led sales.

With entry-level SDR roles disappearing, reps who truly know the product and can manage AI agents will become increasingly valuable.

Owner.com targets US$3–5 million revenue per rep—10× historic norms—by equipping its 100-person sales team with AI tools.

Companies with ~300 customers and 30k contacts still gain from agents; you need some traffic, not giant databases, to see value.

Lemkin cites Marc Benioff envying Palantir’s model where customers go live before paying.

Founders should not quit companies with happy revenue customers; instead build new AI features for them.

Instead of hiring unproven GTM engineers, promote someone who already enjoys data and AI experiments to manage sales agents.

SaaStr swapped 8-9 salespeople for 20 AI agents plus 1.2 humans, keeping revenue steady while cutting cost.