Skip to content

No Code Is Code: Zapier CEO Wade Foster on Headless Tools, Zapier MCP & Automation Bench

The Cognitive Revolution

Hosted by Nathan Labenz

1h 8m
Transcript

Original podcast

Where to listen

Summary

Written by a model from the transcript; not yet reviewed.

Wade Foster, co-founder and CEO of Zapier, returns to discuss how the company has evolved alongside rapidly advancing AI models. The conversation explores the shift toward headless tools and MCP servers, where users work inside a daily driver like Cursor while pulling in third-party data and context. Wade explains that approximately 80% of agentic tasks are actually better and more reliably handled with traditional deterministic code, reserving AI strictly for tasks that require genuine reasoning. They also examine why white-collar work has changed slower than expected, how Zapier uses multi-agent loops to troubleshoot customer issues, and the internal operational strategies Zapier employs to foster AI fluency across its 800-person team.

The discussion provides a grounded look at enterprise AI adoption, highlighting specific benchmarks like Zapier's automation bench where models currently top out around 40% success. Wade shares practical insights on overcoming the recommendation gap by having automated weekly scripts review activity across communication tools to suggest custom automations. This conversation is worth listening to for anyone trying to bridge the gap between AI hype and practical business productivity, offering concrete examples of how to combine deterministic software with AI reasoning for real-world ROI.

  • Reserve AI agents for tasks requiring genuine reasoning while using deterministic code for the roughly 80% of workflow tasks that require reliability and lower cost. 0:15
  • Improve troubleshooting accuracy by spinning up multiple independent agents to evaluate issues and comparing their outputs for consensus. 25:34
  • Encourage internal team adoption by competing to share communications in public channels so both humans and AI agents can access full context. 51:20
  • Maintain rigorous editorial standards and personal verification when using AI for writing to avoid generating low-quality volume and hallucinated details. 59:13

Chapters

  1. 0:00
    Podcast Introduction and Host Welcome

    The host greets listeners, introduces Wade Foster, and sets the episode’s focus on Zapier’s AI evolution.

  2. 0:41
    Challenges of AI Adoption in White‑Collar Work

    Discussion of why AI hasn’t dramatically altered white‑collar tasks, citing model limitations and the need for deterministic code in many scenarios.

  3. 1:39
    Automation Bench and Model Performance

    Exploration of Zapier’s Automation Bench, its 40% success rate, and how Zapier’s tools improve model outcomes compared to raw LLMs.

  4. 2:18
    Recursive Self‑Improvement and Multi‑AI Strategies

    Wade explains Zapier’s internal loops, running multiple AIs on the same problem, and the benefits of diverse model usage.

  5. 25:00
    Human‑AI Co‑Authorship and Workflow Trust

    The conversation turns to how humans and AIs collaborate, the importance of explaining AI outputs, and building trust in automated work.

  6. 35:00
    Future of AI Pricing and Market Competition

    The final segment covers pricing models, token economics, competition with large model providers, and Zapier’s strategy to stay ahead in a crowded AI landscape.

  7. 46:24
    AI Transformation Leadership and Company Culture

    Wade discusses shifting AI responsibility to the Chief People Officer, the company’s AI fluency journey, and scaling AI from individuals to enterprise workflows.

  8. 48:02
    Reskilling and Reorganizing Teams

    Explores how Zapier’s leadership rethinks job roles and reskilling to align with new AI‑driven workflows.

  9. 48:48
    Public‑by‑Default Communication Culture

    Discusses the shift to transparent, public Slack channels and the incentives that encouraged executives to share more openly.

  10. 51:37
    Balancing Transparency with Sensitive Topics

    Outlines which conversations remain private (HR, security) while most work stays public to maximize context for humans and AI agents.

  11. 52:22
    AI‑Driven Security and Credential Stewardship

    Examines Zapier’s approach to managing vast credentials, leveraging AI for vulnerability detection, and the evolving security posture.

  12. 55:24
    Token Budgeting and Governance

    Details the challenges of high token spend, self‑policing tools, and the future of budget allocation based on AI fluency.

Highlights

  • 7:40

    Headless Tools and Daily Driversinsight

    The guest explains that knowledge work is increasingly consolidating around a single daily driver tool, making headless integrations like MCP servers the winning strategy for platforms like Zapier.

  • 15:14

    Deterministic Code Versus AI Reasoninginsight

    The guest argues that about 80% of what people use agents for should actually be handled by old-fashioned deterministic code for better reliability and lower cost.

  • 25:34

    Multi-Agent Troubleshooting Strategystrategy

    To solve support and workflow failures, the company spins up five independent agents to evaluate the situation, proceeding with the solution when four of them agree.

  • 42:27

    The Death of Seat-Based Pricinginsight

    The guest predicts that seat-based pricing is dying in the age of AI, forcing companies into usage-based or outcome-based models that meter the actual work being performed.

  • 48:25

    Putting the People Officer in Chargestory

    The guest explains why their Chief People Officer was put in charge of AI transformation, focusing on managing job descriptions, retraining, and structural shifts rather than technical adoption.

  • 55:18

    Managing High Token Spendersdata

    Some individual engineers at the company are spending upwards of thirty thousand dollars a month on tokens, prompting the creation of self-policing visibility dashboards.

Quotes

  • “When I last spoke to Wade in September of 2024, some 400,000 customers had already used Zapier to delegate more than 100 million tasks to AI.”
    Nathan Labenz · 0:09
  • “Astra setting a new high of just 40% success on Zapier's automation bench, which consists of roughly 600 knowledge work tasks across marketing, sales, HR, and operations.”
    Nathan Labenz · 1:27
  • “The vast majority of what people are using an agent for, 80% in fact, they probably should be using actually old-fashioned deterministic code.”
    Unnamed speaker · 15:14
  • “We have tons of data on what it takes to do. We hook into everything. And so how do we take that data to actually make our products better?”
    Unnamed speaker · 26:58
  • “The market for automation was like orders of magnitude bigger than I ever thought it was when we started the company 15 years ago. It's probably a thousand times bigger than what we set out.”
    Unnamed speaker · 33:34
  • “And so the challenges start to look like a lot more like people issues where it's like, okay, we kind of have to rewrite certain job descriptions.”
    Unnamed speaker · 48:01
  • “We have to rethink how we do compensation.”
    Unnamed speaker · 48:08
  • “One interesting thing, Zapier's had this value of default to transparency for, gosh, forever, it feels like.”
    Unnamed speaker · 50:04
  • “And we looked at that and thought we could use a little bit of a reminder.”
    Unnamed speaker · 50:52
  • Mercury 3:33–5:18
  • Athena 20:00–20:56
  • Outsystems 21:02–22:57
  • Anthropic 34:01–35:32
  • Zapier 1:03:05–1:03:28
  • Cognitive Revolution 1:06:01–1:06:15

About this episode

Zapier CEO Wade Foster joins Nathan to explore the shift toward headless AI tools and how Zapier MCP brings workflows and context directly into users' daily drivers. Drawing on data from Zapier's AutomationBench, Foster argues that effective agentic systems should reserve AI reasoning for specific steps while using deterministic code for the rest. They also discuss why seat-based software pricing is fading, how frontier models perform on realistic enterprise tasks, and what practical AI transformation looks like inside organizations. For full show notes, links, and references, read the episode page:https://www.cognitiverevolution.ai/no-code-is-code-zapier-ceo-wade-foster-on-headless-tools-zapier-mcp-automation-bench/ Sponsors: Mercury Command: Mercury Command brings powerful conversational AI directly into your banking, providing real-time natural language access to your finances without exposing data to third-party tools. Learn more and apply online in minutes at https://mercury.com Athena: Athena matches you with a dedicated, top 1% executive assistant to handle your inbox, calendar, and daily workflows so you can save an average of 15 hours a week. Get matched with your EA today at https://athena.com/cognitive OutSystems: OutSystems is the leading agentic systems platform that enables enterprises to engineer, orchestrate, and govern AI applications on a single unified platform. Learn more and see how it works at https://outsystems.com/tcr Claude: Claude is the AI collaborator for problem solvers, helping with writing, coding, financial models, strategy, and more. Get started with Claude and explore Claude Pro at https://claude.ai/tcr CHAPTERS: (00:00) About the Episode (03:33) Sponsor: Mercury Command (05:21) Daily drivers and MCP (10:20) Benchmarking automation models (18:01) Replacing no-code with code (Part 1) (19:31) Sponsors: Athena | OutSystems (23:02) Replacing no-code with code (Part 2) (30:34) Navigating market competition (Part 1) (34:01) Sponsor: Claude (35:36) Navigating market competition (Part 2) (39:30) Passive workflow discovery (46:09) Internal AI transformation (55:46) Managing tokens and slop (01:03:37) Episode Outro (01:05:59) Outro PRODUCED BY: https://aipodcast.ing

Transcript

0:00
1:08:01

12,265 words · 3 speakers

Speaker names marked are inferred from speaking time and the show’s known hosts — they are not voice-verified.

Topics

  • Claude
  • Cursor
  • tokens
  • Slack
  • Astra
  • Claude Code
  • Astro
  • Composio
  • Zero XYZ
  • AI agents

Page updated · transcript and summary are machine-generated