More than 50% of our little sales team at SaaStr is now AI.
Not because we’re chasing headlines or trying to be the poster child for the AI revolution. We are trying to learn, so maybe we are pushing harder here than most. But the reality is much more pragmatic—and honestly, more human—than that.
The Learning Imperative (But That’s Not the Real Story)
Part of this is about staying current. We’re in the Age of AI, and if you’re running a B2B company today and you’re not actively experimenting with AI in your operations, you’re already behind. We believe we’re ahead of 95% of companies with four AI agents in full production handling three distinct types of sales, plus support, content review, and community management.
But that’s why we started to experiment on AI in sales. Not why we went deep.
The Uncomfortable Truth: We Got Tired of the Arguments
The real reason we went deeper into AI sales isn’t about innovation or competitive advantage. It’s about operational sanity.
We got exhausted by the recurring human dynamics that were eating away at our productivity and, frankly, our patience:
- The Follow-Up Failures: “Did anyone reach out to that enterprise lead from the demo yesterday?” Silence. Then scrambling. Then discovering the lead went cold because basic follow-up discipline doesn’t exist.
- The Two-Week Expert Syndrome: New hires who’ve been with us for 14 days questioning strategies we’ve refined over a decade of building SaaStr. The confidence-to-competence ratio was consistently inverted.
- Generic Outreach at Scale: Watching team members send the same templated message to hundreds of prospects, then genuinely wondering why response rates were sub-1%. Despite training. Despite examples. Despite repeated conversations about personalization.
- The Phantom Meetings: “No one was interested” became code for “I didn’t actually make the calls” or “I sent one email and gave up.” When we dug deeper, we’d find zero meaningful prospect engagement had occurred.
- CRM Black Holes: Getting sales reps to consistently log activities and summarize key meetings felt like negotiating international treaties. Every. Single. Week.
Here’s what changed with AI: When our AI SDR references specific SaaStr events people attended, congratulates them on new roles it found on LinkedIn, and suggests relevant programs based on their company profile, it’s not because it “felt like being thorough that day.” It’s because that’s how we programmed it to operate.
The bar we set: would I have written this email myself? And consistently, the answer became yes.
What AI Actually Delivers (The Unglamorous Reality) — And Why
Here’s what our AI sales agents do that solved these problems:
- They follow up. Every time. Within the specified timeframe. With relevant context from the previous interaction.
- They don’t argue with proven methodologies. They execute the playbook we’ve spent years refining.
- They personalize outreach because they’re programmed to research prospects before reaching out. Not because they feel like it that day.
- They log every interaction immediately with detailed notes. Because that’s literally how they’re designed to function.
- They work 24/7 without vacation requests, sick days, or attitude adjustments.
The training reality: We spent the same time training our AI as we would a human SDR—two weeks of intensive work. The difference? We could train at 6 AM or 11 PM. For our first two months with AI chat on our website, we spent 30 minutes every morning auditing conversations, catching errors, and manually writing better responses for the AI to learn from. Now we do it once a week.
The data requirement: Our AI needed training on everything—20+ million words of SaaStr content, CRM data, marketing automation platforms, website content, past event attendees, sponsorship history. All of it. The uncomfortable discovery? Our data was messier than we thought, with opportunities never logged in Salesforce and missing context from AEs who never used the system properly.
Is AI perfect? Absolutely not. A great human salesperson will outperform our AI agents on complex deals, relationship building, and nuanced negotiations every single time.
Is a great human better than AI? Without question.
But here’s the thing: We weren’t consistently getting great humans. We were getting humans with great potential who required constant management overhead to execute basics.
The Numbers Don’t Lie (But They’re Not What You Expect)
Here are the hard metrics from our AI SDR experiment that validate this shift:
In just two weeks: 4,495 hyper-personalized emails sent, with response rates that made us the #1 performer on our AI SDR platform. Not #1 in our industry—#1 across all users on the platform.
But here’s the reality check: this wasn’t magic. It required 90 minutes every morning, an hour every night, plus real-time responses throughout the day. For two weeks straight.
The performance breakdown by campaign type:
- Website visitor reactivation: Hit-or-miss results
- Lapsed sponsorship accounts: Strong performance
- Previous event attendees who hadn’t returned: “Money” (as we put it)
- Cold outbound: Performed worst, as expected
The math isn’t always straightforward on individual deal closure. Our AI agents might close deals at 60-70% the rate of our best human reps. But when you factor in:
- Zero management time for basic execution
- No motivation or coaching needed on fundamentals
- No salary negotiations or career development conversations
- Consistent execution regardless of personal circumstances
- Instant scaling when we need more capacity
- 24/7 operation capability
The total cost of ownership equation shifts significantly. Especially when you add the management time, training costs, turnover, and emotional energy we were spending on basic accountability issues.
What This Means for Sales Teams
We’re not advocating for replacing all humans with AI. :). But we are saying that if your sales team consistently struggles with basic execution discipline, AI might solve your immediate problem while you figure out your hiring and management challenges.
The uncomfortable truth many sales leaders aren’t discussing: Sometimes the “people problem” is easier to solve with technology than with better people management.
Our AI agents don’t complain, don’t have bad days, and don’t need constant reminders about basic professional behaviors. They just execute.
That consistency has value, even when the peak performance isn’t as high as what a great human could deliver.
The Future We’re Building Toward
Our utlimate goal:
- Almost all outbound handled by AI. Managed by humans every single day. – Done
- 50%+ of all tickets to SaaStr events sold by AI, rest by classic self-serve (no humans) – Done
- All inbound handled by AI in real-time. – Getting close
- 20% of sponsorships sold by AI, rest closed by humans – Looking for tool
- Automate customer success check-ins with AI – To come
We’re about half way there.
The Hard Lessons from our AI Implementations:
Expect it to be more work, not less. Doing AI right requires what we call “S-tier human orchestration.” You can’t hand this off to junior ops folks or agencies. It requires someone who knows your business deeply and can provide high-quality training and oversight. Every single day.
Plan for daily auditing. We spent 30-45 minutes every day reading samples of what our AI sent. When we found something wrong, we didn’t just fix it—we taught the AI why it was wrong and what to do instead. After 60 days of daily improvements, we had something we were proud of.
Human-in-the-loop isn’t optional. When prospects responded to our AI, I had to respond instantly at the same quality level as the AI. The expectation is that you’re always “on” because the AI created that impression of responsiveness.
Segment everything. We ran separate campaigns for lapsed sponsorship accounts, previous event attendees who hadn’t returned, recent website visitors, and cold outbound. The reactivation campaigns were money. Cold outbound performed worst, as expected.
For us, AI sales agents solved an immediate operational problem. They might not be the optimal long-term solution, but they’re the right solution for right now.
And sometimes, that’s enough to keep moving forward while you figure out the bigger picture.
The reality of scaling B2B sales in 2025 is that consistent execution often matters more than peak performance. AI gives us that consistency while we work on building systems that help humans deliver their peak performance more reliably.
