Jason Lemkin transformed SaaStr's sales team from 10 humans to 1.2 humans and 20 AI agents while maintaining revenue. He argues that the 'mediocre middle' of sales is obsolete and that the future belongs to technical leaders who can orchestrate and train AI agents.
Overview
In this conversation, Jason Lemkin, founder of SaaStr, details his radical restructuring of the company's Go-To-Market (GTM) organization. Driven by the frustration of high turnover and mediocrity in junior sales roles, Lemkin replaced nearly his entire sales development team with AI agents. He reveals that the business now runs with '1.2 humans' (one AE and a fractional manager) and 20 specialized agents, achieving similar revenue results with significantly higher efficiency. The discussion moves beyond theory into the practicalities of implementation, emphasizing that 'out-of-the-box' AI fails without rigorous training and 'Forward Deployed Engineers.' Lemkin warns that traditional SDR roles are effectively dead, but predicts a lucrative future for 'AI Orchestrators'—professionals capable of managing agent fleets. The interview serves as a blueprint for the imminent transition from human-heavy sales floors to agent-led revenue operations.
Key Points
The Agent-First Pivot: SaaStr transitioned from a team of ~10 sales professionals (SDRs/AEs) to a structure of 1.2 humans and 20 AI agents. This shift was precipitated by staff attrition and the realization that AI could handle volume work without the turnover or mediocrity of junior hires. The agents now handle inbound qualification, outbound campaigns, and support. Why it matters: Demonstrates that AI displacement in sales is not theoretical; it is already viable for complex B2B businesses to maintain revenue with 90% fewer humans. Evidence: Now you have 1.2 humans, 20 agents... The business is doing very similarly to what it was when you had 10 humans.
Extinction of the 'Mediocre Middle': Lemkin argues that AI is replacing tasks people dislike and displacing 'midpack' performers. While he would hire 'great' humans instantly, he refuses to hire average SDRs who fail to learn the product. The bar for entry-level sales has been raised to a level that excludes non-experts. Why it matters: Signals a massive workforce shift where entry-level sales roles (SDR/BDR), traditionally the training ground for sales careers, are evaporating. Evidence: AI is replacing the jobs people don't want to do today, and it is displacing the midpack and the mediocre.
The Necessity of 'Forward Deployed Engineers': Success with AI agents requires more than just buying software; it requires vendor support in the form of Forward Deployed Engineers (FDEs). Lemkin advises against building internal GTM tools unless you are an engineering company (like Vercel), instead advocating for selecting vendors who will actively help deploy and train the agents. Why it matters: Redefines vendor selection criteria: the software's capability matters less than the vendor's willingness to provide technical implementation support. Evidence: You got to do another column which is your forward deployed engineer... talk to them and say who is going to help me.
Orchestration as the New High-Value Skill: The new critical role is the 'AI Orchestrator' or 'Chief AI Officer'—a person who spends hours weekly reviewing agent logs, correcting hallucinations, and refining prompts. This role requires a data-driven, 'nerdy' mindset rather than a traditional sales management approach. Why it matters: Identifies the immediate career pivot for sales professionals: those who learn to configure and manage agents will be 'hyper employable.' Evidence: Amelia... spends 20% of her time managing the agents, orchestrating the agents... It's not a good job for lazy people.
Training on the 'Best' Human: To make AI agents effective, they must be trained on the emails and scripts of the organization's single best salesperson. Most companies fail because they rely on generic models or untrained agents. A well-trained agent mimics the top performer, vastly outperforming the average human rep. Why it matters: Provides a specific tactical playbook for implementation: quality data ingestion from top performers is the primary lever for AI success. Evidence: Take your best person on your sales team... take their email copy and use that as a template for your AI.
Lead-Rich Environments via AI: AI allows companies to follow up with 100% of leads, including those previously ignored by humans due to low perceived value. SaaStr used Salesforce's Agentforce to reactivate 'dead' leads humans wouldn't touch, achieving a 70% response rate. Why it matters: Highlights hidden revenue: AI captures value from the 'long tail' of leads that human economics previously made unviable to service. Evidence: We just took Asian Force just on those... It had 70% response rate. Those are people that were dying to interact with us.
Sections
Memorable Quotes
Key verbatim statements from Jason Lemkin regarding the shift to AI sales.
We have 10 desks that used to be go to market people. They're all just labeled with our agents. Reply for replet, quali for qualified, arty for artisan.
We're done with hiring humans in sales. We're done.
If you can go do this, you're hyper employable.
Agents work all night and they work weekends and they work on Christmas.
The classic SDR junior kid that is hired out of college to send emails... we don't need them.
Meta-Level Insights
Synthesis of implied trends and patterns in the AI GTM landscape.
The 'Chief AI Officer' in GTM is a tactical, not strategic, role. Unlike a traditional C-suite executive, this person is essentially a high-level systems administrator and editor, spending significant time auditing logs and tweaking prompts to prevent brand damage.
Implementation Fatigue is the next headwind. Companies are rapidly adopting agents now, but the sheer effort required to integrate, train, and orchestrate these tools will lead to a 'closed window' where enterprises refuse to add new agents due to exhaustion.
The 'People Person' defense is failing. The traditional view that sales requires a human touch is being eroded by the fact that AI (like ChatGPT or fully context-aware agents) often provides a more responsive, patient, and knowledgeable 'relationship' than a commission-hungry, slow-responding human.
Future Forecasts
Jason Lemkin's predictions for the near future of sales and AI.
The classic email-based SDR and inbound lead qualifier roles will be 90% displaced/extinct within 12 months.
We will see the rise of the $250,000/year SDR, who earns executive pay by managing a fleet of 10+ agents rather than doing the outreach themselves.
Revenue per rep targets will shift from $300k-$500k to $3m-$5m as AI leverage becomes standard.
Mentioned Tools & Resources
Software and concepts discussed during the interview.
Artisan - AI Agent for outbound sales.
Qualified - AI Agent for inbound lead qualification.
Salesforce Agentforce - AI agents built on the Salesforce platform (used for reactivating dead leads).
Replit - Used by Jason to build internal apps and calculators without coding expertise.
Reeve (formerly Revar) - AI image generation tool used for B2B marketing assets.
Momentum / Attention - AI RevOps tools for CRM automation and call tracking.