Sep 17, 2025 · 36m · saastr
Building AI-First Companies: Insights from Zapier’s CEO Wade Foster
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Zapier co-founder and CEO Wade Foster joins SaaStr's Swapping Notes podcast to share practical insights on scaling enterprise AI orchestration, managing autonomous agents, and cultivating company-wide AI fluency.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
Jason as informed peer 4.9 Guest teaching 4.2 Guest disagreement 0.6 Jason pushing back 1.4
speaking balance: gold is Jason, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 6:09 Wade calls out agent market hype
Wade forcefully dismisses viral claims on social media about ubiquitous autonomous agents as mostly hype compared to real-world hybrid deployments.
Hardest push from Jason ▶ 30:39 Guillaume redirects to hiring implementationGuillaume interrupts to steer the discussion back to the core question of how AI fluency is actually tested during candidate screening.
Biggest teaching moment ▶ 24:10 Wade breaks down MCP vs API coexistenceWade explains the technical and economic realities of MCP versus REST APIs, correcting the notion that MCP will simply replace traditional integrations.
Jason holds their own ▶ 12:04 Guillaume reframes agent workflows around human QAGuillaume provides deep practitioner insight by conceptualizing agentic output as an instructional prompt for human operators with built-in feedback loops.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Jason as informed peer | Guest teaching | Guest disagreement | Jason pushing back | Why |
|---|---|---|---|---|---|---|
| The Evolution from Simple Triggers to AI Orchestration | 4 | 6 | 2 | 1 | Guillaume asks about the threshold separating deterministic workflows from agents. Wade pushes back on market hype, explaining that practical AI deployments combine deterministic steps with agentic prompts rather than relying on pure autonomous agents. | |
| Case Study: Zapier’s Autonomous Renewals Agent | 5 | 4 | 0 | 1 | Wade details Zapier's internal autonomous renewals agent aggregating Gong and Zendesk data for reps. Amelia affirms the concept by introducing the term prescriptive selling, leading to shared agreement that traditional discovery calls are obsolete. | |
| Human-in-the-Loop Oversight and Error Management | 7 | 3 | 1 | 3 | Guillaume demonstrates strong analytical authority by reframing agentic outputs as prompts for humans and highlighting the necessity of human variability in QA. Wade fully aligns, detailing the risks of unmonitored AI in last-mile interactions. | |
| Fostering Internal AI Culture and Automation Mindset | 6 | 3 | 0 | 1 | Amelia and Guillaume share SaaStr's operational realities with LLM hallucinations and acceptable error rates. Wade details Zapier's cultural ethos of building robots rather than acting like robots. | |
| User Adoption Trends and Lowering the AI Learning Curve | 4 | 5 | 0 | 1 | Guillaume inquires about customer adoption patterns. Wade breaks down the behavioral difference between experienced deterministic workflow users and beginners who struggle with open text boxes. | |
| The Impact of Vibe Coding and Ecosystem Growth | 5 | 6 | 1 | 2 | Guillaume probes whether MCP will replace traditional APIs. Wade educates on the architectural distinction, clarifying that deterministic APIs remain necessary alongside token-consuming MCP agent protocols. | |
| Top Enterprise AI Automation Use Cases | 3 | 5 | 0 | 0 | Amelia asks for primary enterprise use cases. Wade provides a structured taxonomy spanning automated case study generation, customer communications, and cross-team voice-of-customer aggregation. | |
| Defining AI Fluency and Operationalizing Hiring Standards | 6 | 4 | 1 | 4 | Guillaume presses Wade on concrete hiring standards for AI fluency after Wade initially focuses on internal hackathons. Wade gives actionable advice, criticizing superficial CEO memos in favor of live skills evaluations. | |
| Rapid Fire: Top AI Productivity Tools | 4 | 2 | 0 | 0 | A rapid-fire discussion where both guest and hosts bond over unanimous adoption of Granola and voice-to-text LLM workflows. |
Statements from this episode (11)
Foster: Social media posts about autonomous AI agents are mostly hype
“If you go read X, if you go read LinkedIn, you would think that everybody is deploying agents Everywhere. Here, look at this crazy agent I've done that has solved all these magical problems for me. Mostly that is hype.”
Foster: Scaled automation mixes deterministic and agentic AI workflows
“So what are the folks that are actually deploying like automation at scale doing? They're finding ways to mix and match deterministic workflows with agentic workflows and do those together.”
Foster: Sales and support reps have no excuse for lacking account context
“There's no excuse for a sales rep or a CSM or a support rep to not have context about your account, to know what you're doing.”
Foster: Highly agentic workflows are highly imperfect and require human intervention
“I think a reason a lot of folks are taking this approach of inserting the human at the last mile is the more agentic these workflows, the more imperfect it is.”
Foster: AI Will Eventually Act Autonomously as Sales Reps and CSMs
“I think it takes a lot of work to get that last mile good enough where you would say, Hey, we're actually going to feel confident actually having the AI act as the sales rep or act as the CSM in these equations where it's going to get there. It's going to get …”
Foster: Zapier's sales team touches under 1% of total customers
“Our sales folks touch probably less than one percent of our total customers. And it's not meaningfully infringing on their job. It's just helping them like touch more accounts that otherwise we never could have touched.”
Foster: Vibe coding tools expand the market for software stack products
“Nowadays, that person can go use a Lovable or a Replit or whatever and get started. And so it's increasing the market for all of the tools in the stack because the entry point is just so much simpler.”
Foster: MCP will not replace APIs; both will remain valuable and coexist
“I think the answer is we're going to end up having both. I think APIs are very, are deterministic and cheap. Works the same way every single time. And so there's a whole set of use cases that Do you want it to work that way? Where MCP is a lot more flexible. T…”
Foster: Zapier uses internal AI agents to aggregate cross-team customer feedback
“And so we have a whole automation system agent system that like takes all that stuff. It then generates a bunch of themes based on that and then generates reports based on those themes. Then the second thing we have built up is a chat bot around all that stuff…”
Foster: Zapier drives internal AI fluency with hackathons and weekly show-and-tell
“We run these hackathons usually every four to six months just to keep fresh with it. And that's really important because AI moves crazy fast. And so every time we do that, we see the internal adoption just go higher and higher as people start to figure out new…”
Foster: 'AI-first' executive mandates cause chaos without function-by-function fluency baselines
“You get the CEO memo, like we're an AI first company, right? Which you need to do that step is an important step. But I talked to a lot of people who that's the only step they do. And then it's just like chaos inside the organization being like, I don't know w…”