So in early 2024, Lenny Rachitsky invited us on to Lenny’s podcast for what became an iconic deep dive on sales and GTM. How and who to hire, the classic GTM and sales mistakes we all make, and more. It was great.
It was also … an entire era ago. The end of the pre-AI era, really. Everything in that classic GTM convo from Q1 2024 holds true today. But it also was about a world where all of sales, marketing, onboarding, support and success were done by humans.
So Lenny invited us back to talk about AI and GTM in 2026. His first show of 2026, in fact. HNY!!
And so much has changed, and more importantly, is just beginning to change.
At SaaStr, we’ve replaced nearly our entire go-to-market team with AI agents. The results? The same revenue with 1.2 humans instead of 10. Here’s what every founder and sales leader needs to know.
Top 10 Takeaways
- The math is real: SaaStr went from 8-9 human salespeople to 1.2 humans + 20 AI agents, with equivalent business performance.
- Classic email-based SDRs are going extinct: Cadence-based SDRs sending emails and qualifying inbound leads will be 90% displaced within 12 months. The role simply won’t exist.
- The $250K SDR is coming: Top performers won’t send emails—they’ll manage 10 AI agents each. The job title stays, the job description transforms entirely.
- AI closed a $70K deal autonomously and a $100k deal on New Year’s Eve!: A Qualified AI agent sold a sponsorship at 11 PM on a Saturday. No human touched it until procurement. And on New Year’s Eve, another $100k deal came in through an AI Agent.
- Training takes 30 days, not 30 minutes: You need to upload data, answer questions, correct mistakes, and iterate. Budget 50-60 hours to get it right.
- Don’t build—buy: Unless you have a skilled engineer who genuinely wants to build custom AI tools, use existing products. They’re evolving too fast to DIY.
- Human oversight isn’t optional: You need a “Chief AI Officer” type—someone nerdy, quantitative, who loves data and can sit in front of dashboards for hours.
- The plays work, the playbooks are broken: Go-to-market fundamentals haven’t changed, but the execution playbooks from 2019 are obsolete.
- Support is the easiest starting point: 50-80% of support is already AI-handled at leading companies. Start there.
- 70% of AE jobs are safe for now: But that number drops to 40-50% as AI capabilities expand. The clock is ticking.
The SaaStr AI Experiment: What Actually Happened
Jason didn’t plan to rebuild his entire sales org around AI. It started when two sales team members quit simultaneously in May 2025. Rather than hire replacements, he doubled down on AI agents that were already showing promise.
The setup today: 20 AI agents across different workflows, managed by one full-time human AE and 20% of a “Chief AI Officer’s” time. Different agents handle different sales cycles—low-ticket sales run entirely through AI, while high-end sponsorships get human involvement at the negotiation stage.
The surprising part? Lead response and qualification actually improved. AI agents respond instantly at 11 PM on Saturday nights. They never get tired. They never cherry-pick the best leads. They follow up on every single opportunity.
What’s Dying, What’s Surviving, What’s Thriving
Going Extinct (12 months):
- SDRs who send templated emails and qualify inbound leads
- BDRs running classic cadence-based campaigns
- Any sales role where the primary function is information routing
Surviving (for now):
- AEs handling complex negotiations and procurement
- Enterprise sales requiring genuine relationship building
- High-dollar deals where personalization ROI justifies human time
Thriving:
- Sales professionals who manage and train AI agents
- “Nerdy” demand gen marketers who love data and segmentation
- Forward-deployed engineers who ensure customer success with AI products
- Anyone willing to spend 50-60 hours learning to work alongside AI
The Practical Playbook
Where to start: Support. Most companies can’t provide 24/7 support, but AI can. It’s the lowest-risk, highest-impact entry point.
How to train AI agents well:
- Take your best email copy—not average, your best
- Upload it as the template baseline
- Provide data sources (Salesforce, CRM) for personalization
- Run A/B tests constantly—AI is excellent at creating variants
- Review outputs obsessively for the first 30 days
- Correct mistakes explicitly; AI learns from corrections
The “Incognito Mode Test”: Sign up for your own product with a fresh Gmail address. Experience your support, sales, and onboarding as a prospect would. The gaps you find are your AI opportunities.
Vendor selection tip: Pick the vendors who offer to help you the most. Implementation support separates the winners from the also-rans.
The Uncomfortable Truth About Jobs
Jason is refreshingly direct: AI is displacing midpack and mediocre sales professionals. Not the best, not (yet) leadership—but the middle of the pack is getting squeezed.
The media narrative of “mass layoffs” is mostly wrong, though. What’s actually happening is more subtle: companies are backfilling roles with AI instead of hiring humans. The job that would have gone to a new SDR now goes to an AI agent. The headcount just never gets added.
For salespeople, the advice is clear: Don’t wait. Pick an AI tool today. Train it yourself. Become the person at your company who knows how to make AI agents productive. That skill set is the new job security.
Top Mistakes Founders Make with AI Agents in Sales
- Expecting plug-and-play magic: AI agents require 30 days of training and 50-60 hours of setup. Companies who expect instant results get instant disappointment.
- Building instead of buying: Custom AI tools become obsolete fast. Unless you have genuine engineering capacity and desire, use commercial products.
- No human oversight: You need someone watching the AI outputs daily. “Set it and forget it” leads to embarrassing mistakes at scale.
- Training on mediocre content: If you train AI on your average email copy, you get average AI emails. Train on your best performer’s best work.
- Skipping segmentation: Multiple AI agents need careful base segmentation or they’ll conflict with each other and spam the same prospects.
- Underestimating change management fatigue: Implementing 20 AI agents is exhausting. Companies burn out their teams by moving too fast.
- Waiting for perfection: The technology improves monthly. Waiting for “the right tool” means falling behind competitors who started six months ago.
- Ignoring the human handoff: AI qualifies and nurtures, but humans still close complex deals. Build the handoff process before you need it.
- Assuming “people person” skills are enough: Being likeable no longer differentiates. AI can build relationships too. You need to add genuine strategic value.
- Not testing from the customer’s perspective: Use the incognito mode test. Experience your own sales process as a prospect would. Fix what’s broken.
Magical But Uncomfortable Times
These are what Jason calls “magical times”—uncomfortable, uncertain, but genuinely transformative. The playbooks that got us here are obsolete, but the fundamental plays still work. Find prospects. Qualify them. Help them buy.
The tools for executing those plays have changed forever. The winners will be the founders and sales professionals who adapt fastest.
As Jason puts it: “The best people who master AI will be more employable than ever.” The question is whether you’ll be one of them.
