MurmurCast
InsightfulDiscussion

Turning AI Tokens Into ROI (That You Can Measure) | CEO @ Paid, Manny Medina

Topline1h 5m

Manny Medina, CEO of Paid, discusses how AI agents are disrupting software business models by shifting from seat-based to outcome-based pricing. He argues that all software will be agentic within five years, and companies must rethink pricing strategies around the work agents perform rather than traditional SaaS metrics.

Summary

Manny Medina, former CEO of Outreach and current CEO of Paid, explains how his platform helps companies monetize AI agents by metering and pricing their work. The core problem Paid solves is that traditional seat-based SaaS pricing doesn't work when agents perform work autonomously without needing user seats. Instead of selling seats, companies need to price based on the actual work and value delivered by agents.

Mediana predicts that within five years, all software will be agentic, though companies won't simply disappear—they'll be forced to build agents on top of their existing platforms or lose market share to competitors. This is already happening through MCP (Model Context Protocol) layers that make SaaS platforms accessible to any agent, creating a competitive pressure for companies to capture value from agentic work.

The transcript covers key challenges in pricing agents: tokens are not a unit of value because they're internal to LLM architecture and vary by lab; transparency and explainability are essential first steps, where customers must understand exactly what work the agent performed and what they received; and pricing should be flexible based on customer-specific definitions of outcomes rather than one-size-fits-all models.

Mediana discusses the end of seat-based pricing by using a phone bill analogy—customers want itemized visibility into what they're paying for and why. He contrasts this with the SaaS era where companies sold 100 seats but only 20 were heavily used, leading to contraction at renewal. With agents, if two users drive value through an agent, the company doesn't lose revenue from the unused 90 seats because they're charging for work output, not seats.

On margins, Medina argues that the era of 70-80% gross margins in SaaS was an anomaly created by cheap money and will end. Future software businesses will likely run on 40-60% gross margins, though agent-based businesses in narrow verticals could be highly profitable like Viva Analytics. He notes that companies like Sierra are proving this is possible by implementing outcome-based pricing catalogs with customer-specific definitions.

Mediana shares lessons from Outreach: he regrets not hiring a product marketer earlier (which he's now doing at Paid), and he shouldn't have taken the company's final fundraising round. He's deliberately building Paid differently—keeping the team small, prioritizing revenue growth over headcount growth, maintaining high ACV, and building a system of record rather than workflows, making Paid responsible for the authoritative billing record. He emphasizes that winning should be the priority, and good culture is the exhaust of winning rather than the other way around.

The conversation touches on the tension between open-source models and frontier models like OpenAI and Anthropic. Medina argues that foundation model companies operate like hardware firms, continuously releasing expensive models while bringing costs down, making it difficult for distilled/open-source models to catch up. However, he acknowledges this dynamic depends on maintaining low interest rates and affordable GPU infrastructure.

Mediana shares examples of Paid customers solving complex problems: Oliver Invent uses agent chemists to reformulate drug dyes in hours instead of the $20 million, multi-month exercises that would normally be required; Squint uses AI to generate manufacturing training videos to address labor shortages. These examples illustrate why he's excited about the space—the variety of real-world agentic problems being solved daily.

About this episode

<p>Manny Medina, Co-Founder and CEO of Paid, the monetization platform for AI agents, and the founder who took Outreach from nothing, joins Sam Jacobs, AJ Bruno, and Asad Zaman in person on how to price AI agents once seats stop making sense. Topics include why every SaaS company needs its own agents before someone else's run on its MCP layer, why a token is the wrong unit of value, and the itemized phone bill every agent builder owes its buyers. Plus, why fixed-price deals win logos but stall expansion, the case for 40% to 60% gross margins in the agent era, and what Satya Nadella told Manny in a green room.<br /> <br /> Key Takeaways:<br /> <br /> - Price the work your agent does, not the seats it fills. In SaaS you sell 100 seats, procurement sees 20 in heavy use, and the renewal shrinks to 50. With agents, "if you've sold 100 seats and only 2 were hyper-users, you don't care because you're charging for all the uses and all the value those 2 seats were getting." For now that pricing lives between effort and output, because true outcome pricing needs attribution, and as Manny put it, "Attribution is really hard."<br /> <br /> - Make the bill readable before you argue about ROI. Manny compares it to the old itemized phone bill, where you went straight to the biggest line and asked, "Did I really call Pakistan from New York?" Tokens can't play that role: "A token should not be a unit of value for charging because a token is really kind of like the inside secret." Show the buyer the research, deck, or analysis the agent produced, because "explainability is like the first rung in the ladder."<br /> <br /> - Fixed pricing is easy money that boxes you in at renewal. Enterprises are buying one of each right now, so "new logo is not hard. Expansion is hard." Teams that land logos on fixed price hit the first renewal with nowhere to go, which is why Paid opens with a monetization consultant and takes a percentage of revenue: "You grow, I grow." - Plan for thinner margins and narrower markets. Manny learned Amazon's margins cold (books 20%, electronics 7%, jewelry 3%) and calls SaaS's 70 to 80% an anomaly of cheap money: "I feel like a normal, you know, software business will be running on 40% margins, 60% tops."<br /> <br /></p> <div> Connect with the Hosts & Guests:</div> <div>  </div> <div> Host: Sam Jacobs, CEO at Pavilion - https://www.linkedin.com/in/samfjacobs/</div> <div> Host: AJ Bruno, CEO at QuotaPath - https://www.linkedin.com/in/ajbruno3/</div> <div> Host: Asad Zaman, CEO at STA - https://www.linkedin.com/in/azaman1/</div> <div> Guest: Manny Medina, Co-Founder & CEO at Paid - https://www.linkedin.com/in/medinism/</div> <div>  </div> <div> Topline is more than a podcast:</div> <div>  </div> <div> Subscribe to Topline Newsletter: https://toplinemedia.substack.com/</div> <div> Check us out on YouTube for the #1 video podcast for founders, operators, and investors in B2B tech: https://www.youtube.com/@TOPLINE-Media</div> <div> Join the free Topline Slack channel to connect with 600+ revenue leaders to keep the conversation going beyond the podcast: https://www.joinpavilion.com/topline-slack</div> <div>  </div> <div> Chapters:</div> <div> 00:00 Introducing Manny Medina</div> <div> 01:37 What Paid Is Building</div> <div> 04:38 Will All Software Go Agentic?</div> <div> 06:35 The 100-Seat Renewal Problem</div> <div> 10:02 An Itemized Phone Bill for Agents</div> <div> 13:33 Why Outcome Pricing Takes Time</div> <div> 19:04 Open Source vs. Frontier Models</div> <div> 26:41 Pricing Agents Is an NRR Problem</div> <div> 33:11 Why Manny Won't Build Workflows</div> <div> 35:39 Fundraising Lessons From Outreach</div> <div> 41:10 Life Is Path Dependent</div> <div> 45:46 Culture Is the Exhaust of Winning</div> <div> 47:50 Is AI Spend Still Experimental?</div> <div> 58:22 The End of 80% Software Margins</div>

Key Insights

  • Medina claims that within five years, all software will be agentic, forcing non-agentic companies to either build agents or disappear, as MCPs enable any agent to plug into platforms and capture value.
  • Medina argues that tokens should not be a unit of value for charging customers because tokens are an internal LLM implementation detail that varies by lab and doesn't reflect meaningful work output.
  • Paid's core insight is that customers need three things: transparency (visibility into what the agent did), explainability (understanding the value received), and pricing flexibility based on customer-specific outcome definitions.
  • Medina contends that the era of 70-80% SaaS gross margins was an anomaly enabled by cheap money and will end, with sustainable software businesses likely operating at 40-60% gross margins in the future.
  • Medina positions Sierra's strategy of maintaining a catalog of pre-defined outcomes—customizable per customer—as a successful model proving outcome-based pricing works, unlike competitors who get paralyzed debating what outcomes mean.
  • Medina claims that seat-based pricing contraction problems (procurement reducing purchased seats at renewal) will disappear when pricing is tied to agent-delivered work rather than user seats.
  • Foundation model companies like OpenAI and Anthropic operate like hardware firms, continuously releasing expensive models while improving efficiency, making it mathematically difficult for open-source distilled models to ever catch up.
  • Medina argues that culture is the exhaust of winning, not a separate priority—when organizations win together, culture naturally improves, but focusing on culture without results is futile.
  • Medina regrets not hiring a product marketer at Outreach early and made the opposite mistake at Paid, now realizing it's one of the highest-ROI hires because it compounds over time.
  • Paid takes a percentage of customer revenue (outcomes-based compensation) rather than selling software licenses, aligning Manny's incentives with customer success and revenue growth.
  • Medina observes that enterprise AI adoption is still sub-1% penetration and most companies are still using AI to make old workflows faster rather than completely reimagining roles for full autonomy.
  • Medina emphasizes that life is path-dependent—regretting the last round of Outreach fundraising wouldn't change where he is now, so processing regrets productively is more useful than dwelling on them.

Topics

AI agents and agentic softwarePricing models and monetizationShift from seat-based to outcome-based pricingTransparency and metering in AI workSoftware margins and profitabilityLessons from OutreachFoundation models vs. open-source modelsEnterprise AI adoption and ROIVertical agent businessesSystem of record vs. workflowsCost of compute and token efficiencyProfessional services and forward-deployed models

Transcript

The world will be fully agentic. So there is no software that will not be agentic in five years. And you just have to figure out how to make money from it. That's Manny Medina, CEO of Paid, a platform which helps companies with AI agents get paid for the work those agents do. He explains why every software company is racing to build their own agents before someone else's agents plug into their platform, do the work and take the money. And ultimately, why it's the beginning of the end for seat-based pricing models. So like, what the f*** do you get for the $100,000 you bought nobody knew? Like, you want value. So all of a sudden, we're…

Full transcript available for MurmurCast members

Sign Up to Access
View original source →

More from Topline

Get AI summaries like this delivered to your inbox daily

Get AI summaries delivered to your inbox

MurmurCast summarizes your YouTube channels, podcasts, and newsletters into one daily email digest.