Podcast Summary: SaaStr 828 — "The AI Revolution in B2B"
Episode Date: November 5, 2025
Host: Kraig Swensrud (Founder & CEO, Qualified)
Guests: Jason Lemkin (Founder & CEO, SaaStr), Amelia Lerutte (Chief AI Officer, SaaStr)
Episode Overview
This episode dives deep into the practical AI transformation happening in B2B SaaS. Jason, Amelia, and Kraig discuss their firsthand experiences deploying AI agents in customer-facing workflows, explore how organizational behaviors must adapt, where the real opportunities and pitfalls are, and share actionable advice for go-to-market teams. The episode is candid, pragmatic, and rich with both strategic perspective and operational detail.
Key Discussion Points & Insights
1. "You Don't Need to Learn AI. You Need to Do AI."
(00:01, 39:07)
-
Jason opens and closes the show by emphasizing action over theory:
“No, you need to do AI, don’t learn it. Buying a subscription to Claude does not count as learning AI. The way you’re going to do it is pick an agent, an agentic product. Pick the simplest possible one use case... If you deploy an agent yourself, train it, QA it, test it, you will be ahead of 90% of the world.” (Jason, 00:01 & 39:48)
-
“Hands on. Yourself, not your team." (Jason, 40:08)
2. Fast Evolution of Agentic Tools & the Imperative to Start
(03:43–07:00)
-
The landscape is changing rapidly; most leading agentic tools didn’t work before early 2025.
-
Instead of waiting for a “winner,” pick a reputable vendor and invest deeply in training and customizing the tool.
“Training is more important. There are very few agentic products… where you want it to do something like interact with customers, speak with authority, close deals... it wouldn’t have happened if we put zero minutes of training into it.” (Jason, 03:43–05:03)
-
The previous model of year-long, expert-assisted deployments is dead—AI time to value is measured in days or weeks.
3. Budget Is Flowing Into AI, Not Legacy SaaS
(07:00–09:02, 27:31–29:59)
-
Almost all new software budget is for AI projects, not legacy tools.
-
Buyer urgency now resembles the early-2000s tech paradigm shifts.
“No one is putting more budget into their old SaaS software. It’s all going into AI. And so everyone’s in market.” (Jason, 07:00)
-
"Classic SaaS" is "geriatric"; B2B software is exploding, but only if you ride the AI wave.
“Traditional budgets are frozen... All the energy is in AI for solutions. If that’s you, it feels dead... But meeting with your customers works. Outbound works—if anyone wants to buy your product.” (Jason, 27:31)
4. SaaStr’s Real-World AI Agent Deployment Journey
(09:20–18:59)
- At the start of 2025, SaaStr had zero AI agents; now it runs 20, with 12 “prime” agents across workflows.
- First step: Clone Jason Lemkin’s knowledge with ‘Delphi’ (Jason AI), answering founder questions 24/7.
- Horizontal → Vertical:
– “We went from horizontal to, you know, 12 verticalized agents in a sense.” (Jason, 13:07) – Start broad to build confidence, then advance to specialized, process-driven agents (SDR, BDR, event support, etc.). - Real wins: Immediate automation of previously poor or missing processes (e.g., outbound emails, qualifying leads, instant event support).
5. Many Workflows Get Better, Faster, and More Automated
(14:48–21:06)
- Example: Contact forms were slow, manual, and often neglected by humans—now instantly handled by an AI agent.
- Amelia details how her morning shifted from single-threaded human management to efficiently “checking in” with each agent, increasing productivity and eliminating dropped leads.
“That hour you spend with the agent is so much more productive. I just go through all of our agents and I just check on them for that hour.” (Amelia, 18:59)
6. Cognitive Load Shifts, Not Disappears
(21:06–22:18)
- “It is a higher cognitive load. You have to think more. You don’t spend time arguing with people... But you use all your brain cells for an hour [with agents]. It’s so many brain cells.” (Jason, 21:14)
- More release velocity, but “doing more” increases mental effort for both line and technical staff.
7. Start with Obvious "Layup" Roles
(22:18–26:46)
- Best first AI agent targets: Support (customer/internal), outbound SDR, neglected workflows—areas with high pain or no reliable human owners.
- “Look at where there’s literally no one doing the work… The two areas that took off most quickly were support and programming.” (Jason, 22:37)
- Avoid “hero” AI purchases pitched to impress leadership; incremental automation where there’s nothing to lose is the highest return.
8. AI Increases Transparency and Performance Standards
(19:44–21:06)
- Deploying agentic tools exposes bad data, poor performance, and “ghost employees” quickly.
- Real-time, automated reporting and interaction surfaces previously invisible problems, making organizations sharper.
9. Deployment Tactics: Stair-Step Approach
(14:58–16:27, 16:52–18:59)
- Crawl-walk-run: Win “small,” learn, iterate, then scale up.
- “If you can stairstep this stuff, you will build confidence by your second, third agent or deployment.” (Jason, 16:10)
- The entire agent journey at SaaStr happened in months, not years.
10. If You’re Not Moving Fast, You’re Already Behind
(31:03–33:02)
- "Halloween" 2025 is cited as the fail-fast deadline for startups; “If you’re not in market with a disruptive AI agent... you need a brand new team.” (Jason, 32:14)
- VC sentiment: “Every investor I know, where the startup hasn’t made the jump yet, they’ve given up—that now they’ve lost hope.” (Jason, 31:16)
- “There’s no excuse in 2025, no excuse… You have to reboot your entire [organization] if you’re not in market [with AI].” (Jason, 32:14)
11. Advice for Go-To-Market Leaders
(39:00–43:46)
- Get personally hands-on with agents, QA, and training—not “delegated learning.”
- “If you don’t do it hands-on keyboard, you will not have a job.” (Jason, 41:50)
- Demand for true AI-literate marketers and operators is higher than perceived—no one wants “the 2019 Marketo playbook” anymore.
- For founders: Don’t stall AI progress due to internal resistance—some team attrition is expected and acceptable.
“You can’t forsake the future of your company... to not hurt people’s feelings. You have to just go for it. Rip the band aid.” (Amelia, 41:56)
Notable Quotes & Memorable Moments
-
“Buying a subscription to Claude does not count as learning AI. You need to do AI.”
– Jason Lemkin (00:01, 39:48) -
On Support as a Layup:
– "A lot of us have terrible support... Go try your own support yourself as a founder... I bet nine times out of ten you’re going to cry how bad your support is."
(Jason, 23:48) -
On Culture Shock:
– "The agents don’t cry. Okay? That is exhausting. But it’s exhausting in a different way… You have to use all your brain cells for an hour, do you? With your agents, it’s so many brain cells."
(Jason, 21:14) -
On Market Urgency:
– "If you’re not in market with a disruptive AI agent as a startup on Halloween, you need a brand new team... give them a nice Thanksgiving bonus and a turkey and tell them they gotta go, because you’ve had 10 months."
(Jason, 32:14) -
On Sales Resistance:
– "I have some agents, but I’m worried about... rolling this out to our sales team... some of them might quit. And I said, that’s okay, let them. Like, it’s okay. It’s almost Halloween... You can’t forsake the future of your company... to not hurt people’s feelings. You have to just go for it. Like rip the band aid."
(Amelia, 41:56) -
On Personal Accountability:
– "If you deploy an agent yourself, train it, QA it, test it, you’ll be ahead of 90% of the world... Yourself, not your team."
(Jason, 40:09)
Suggested Starting Points & Timestamps
- (00:01, 39:48) – “You need to do AI, don’t learn it.”
- (03:43–05:42) – Training, not just vendor choice, defines success.
- (09:20–13:44) – SaaStr’s journey from zero to 20 AI agents, starting with “Jason AI.”
- (17:44–19:18) – Productivity shift: why hands-on agent management outperforms human SDR/BDR management.
- (21:06–22:18) – Cognitive load changes, not shrinks, with AI.
- (22:37–25:04) – Layup roles and practical starting points for agent deployment.
- (31:03–33:02) – How fast you must move: Halloween is the deadline.
- (39:00–43:46) – Final advice to go-to-market leaders; rip the band-aid off.
Key Takeaways
- Stop theorizing—deploy, train, and iterate with agents directly.
- Start with obvious, broken, or neglected workflows for fastest ROI.
- AI is raising organizational standards and visibility, for better or worse.
- Move fast, don’t wait for the “perfect” product or consensus; it’s already late.
- Expect and accept cultural friction—focus on results over feelings.
Closing Note
This episode delivers a jolt of practical energy for SaaS founders and go-to-market leaders. If you haven't moved from AI curiosity to deployment, you’re already behind. If you’re facing organizational resistance—remember, as Jason and Amelia repeat: “Just go for it. Rip the band aid.”
