a16z Podcast
Episode: Software is Eating Labor
Date: October 3, 2025

Episode Overview

This episode, featuring a16z General Partner Alex Rampell, explores the transformative shift underway as software moves beyond simply digitizing processes to actively replacing labor. With the global SaaS market at $300B and the US labor market at $13T, the episode delves into why the real opportunity for software is no longer just tools for workers—but automating the work itself. Rampell uses history, practical examples, and live AI “agent” audio clips to illustrate how business models, pricing, and the very scope of automation are evolving.

Key Discussion Points & Insights

From Filing Cabinets to AI Agents

• Thesis: Historically, most software has just turned filing cabinets into digital databases. Now, software is set to perform the labor previously done by humans, not just support it.
• Historical Evolution: Examples include airline reservation (Sabre), sales (Siebel, Salesforce), manufacturing (SAP), legal (LexisNexis), accounting (Intuit), healthcare (Epic), and payroll (ADP).
  • These systems digitized records but required humans to interpret and act—"The digital records are read by humans." (11:50, Rampell)
• Key Takeaway: The real market opportunity is automating what people do, not just storing information.

Market Size & Opportunity

• Labor vs. Software Spend:
  • US labor market: $13T vs SaaS market: $300B (02:40)
  • Even narrow labor categories (e.g. nursing: $650B) exceed total global software spend.
  • "What software is now going after, the prize it's going after, is the labor market." (01:17, Rampell)

The “Tall, Grande, Venti” SaaS Model Dilemma

• Old Model: SaaS licenses sold by seat; software as a tool for people.
• New Reality: AI could replace most or all seat-based work (e.g., call center agents), making per-seat pricing obsolete.
  • "If every one of your agents is 9,000 times more productive, how many seats do you need?" (13:42, Rampell)
  • Example: Zendesk's ARR could drop to zero or triple, depending on whether it shifts to outcome-based pricing.

Outcome-Based and Operational Software

• AI as Operator:

  • Software will transact, negotiate, collect, and serve—acting instead of just recording.
  • Sales: “Salesforce should just sell for me. I don’t want to pay for a thousand seats. I want to pay for customers.” (16:21, Rampell)
  • Accounting: Software collects payments on your behalf.
  • Healthcare: AI follows up with patients post-surgery, in multiple languages.

• Pricing Models:

  • Shift from license/seat to value/outcome-based billing (e.g., charging per support ticket resolved or dollars collected).

Notable Audio Demonstrations (AI at Work)

• Freight Booking Negotiation Example (Happy Robot) (17:53):

  • A recorded call between an AI and a human, haggling prices for a trucking load.
  • "Who's the robot and who's the human? Human. This is like the new Turing test." (18:40, Rampell)

• AI Collections Agent Example (Salient) (19:06):

  • An AI–not a person–politely and persistently attempts to collect on an overdue loan, in multiple languages.
  • “AI doesn't get bothered… demoralizing jobs, very, very good for AI.” (20:37, Rampell)

Deeper Insights

Why AI Isn’t Just About Cost Reduction

• Intermittent Demand:

  • Humans are expensive to hire/train/fire for seasonal swings; AI can instantly scale or pause.
  • E.g., Holiday retail surges, airlines during weather disruptions.

• Demoralizing Tasks:

  • Collections, outbound calls, late-night support—AI’s tolerance for rejection is infinite.

• Regulatory Certainty:

  • AI can be programmed to strictly follow compliance protocols, reducing risk.

• Language Accessibility:

  • “If I only spoke Farsi…now they do. An AI nurse, an AI collections agent, an AI negotiator, all of these things can be done in dozens of languages instantly.” (21:58, Rampell)

AI Opening New Markets

• Expanded Market Size:

  • Software can enter markets previously considered too small or labor-intensive to automate, such as compliance officer roles and collections.
  • Example: “There is no software company for collections...Now you can roll into Citi and say, ‘Hey, I will do compliance for you end to end.’” (23:06)

• Enabling Non-AI Startups:

  • By reducing customer acquisition and service costs, AI enables business models (e.g., "Airbnb for bicycles") that were impractical before.
  • “The cost per AI sales rep per year is a few hundred bucks, not a hundred thousand dollars. No coconut water necessary.” (24:30, Rampell)

Examples of AI "Eating Labor" in Real Life

• Craigslist Job Posting:

  • AI “applies” for a front desk job at an optometrist's office, promising to do all tasks except physical ones (like locking the door), at less than half the cost of a human. (16:16)

• Healthcare Calling:

  • After surgery, an AI nurse calls Rampell to check on recovery—demonstrating scalable, multilingual patient engagement. (15:37)

Memorable Quotes

• On the Shift:

  • “We're not giving you software, we're going to do a job for you.” (06:16, Rampell)
  • “The people cost you so much higher than the software cost. Zendesk is at the precipice: revenue could go to zero or 3x.” (14:15, Rampell)

• On Labor Automation History:

  • “You could always take capital, make a machine, and then have more efficient labor on the other side. But what's happening now is the whole thing is effectively done end to end.” (08:07, Rampell)

• On Multi-Language Access:

  • “AI nurse...can totally call a mid-40s patient like me and say: How are you doing? Is there anything we can help with? Do you have a fever?” (15:23, Rampell)

• On Future of Software:

  • “Our job...is to find the best companies that will make software look small.” (25:18, Rampell)

Important Timestamps & Topics

• 00:01 – 07:00: Introduction, market sizing, the limits of traditional SaaS
• 07:00 – 13:40: History of software: filing cabinets, digitization, old SaaS business models
• 13:40 – 17:50: The “seat” pricing dilemma, shift to outcomes, using nurses as a market size example
• 17:53 – 18:39: Live negotiation call: AI agent books a freight load (Happy Robot)
• 19:06 – 19:25: Live collections call: AI agent (Salient) asks for overdue payment
• 20:00 – 24:40: Deep dive: intermittent demand, demoralizing jobs, regulatory compliance, language
• 24:40 – 25:18: How AI enables new kinds of businesses, the global opportunity

Tone & Language

Alex Rampell uses a conversational, vivid, and sometimes wry tone—peppering in personal anecdotes, pop culture references (Glengarry Glen Ross, Starbucks "tall/grande/venti" analogy), and direct engagement with the audience ("Who's the robot and who's the human?"). The discussion is engaging and often humorous, while providing sharp analysis rooted in business and technical realities.

Final Thoughts

This episode provides a compelling argument that the next wave of software innovation isn’t about building better tools, but about turning software into the worker itself. As AI becomes ever more adept at performing "white collar" tasks, the potential market for automation expands from a few hundred billion to trillions of dollars annually—unlocking new business models and challenging entrenched pricing and value paradigms. For founders, investors, and anyone interested in the intersection of software, labor, and AI, this is a must-listen conversation.