Christina Cacioppo says Vanta avoids marking up model tokens because users quickly compare it to Claude and ask why the markup is better. Vanta is absorbing token costs, accepting lower margin, and focusing on getting more customers into agent workflows before sorting out pricing later.
Cacioppo says an eight-year-old codebase overwhelms coding agents because they lack context and do silly things. Vanta is modularizing parts so agents can fully understand smaller pieces, following Notion, instead of needing high-context engineers to babysit every change.
Cacioppo says Vanta's product team kept pitching onboarding wizards because they knew wizards work and doubted the LLM drag-and-drop flow could. Leadership had to explicitly say they would bet on the LLM and allow hill climbing; once that fear was removed, the team shipped the new experience.
Cacioppo says inference providers admit post-training is about cost, not quality. She sees companies doing it for investor and customer marketing, and doubts most users want to be handed a custom model; the deployment and form factor are still unsolved.
Cacioppo calls FDE a talent arbitrage, like chief of staff, but often just professional services. Palantir-style eight-figure contracts can support it, while five- and six-figure contracts leave the economics upside down in an era of 30% growth and margin.
Cacioppo says models persist where people give up and can chain many individually benign findings into a real exploit. That breaks normal 90-day prioritization, pushing security toward fixing issues as fast as they are found, eventually with constant penetration-test and patch agents.
Cacioppo says she does not trust big companies with her email or chats and would rather Google or Apple hold that data. She sees B2B trust as measurable through data and contracts, but consumer trust as vibes that are much harder to turn back.
Cacioppo says Vanta is building an agent harness instead of SaaS UI; reporting comes from an agent making charts on demand, not Tableau inside the product. She accepts the timing risk and refuses to build an AI off switch for enterprise procurement.
Cacioppo says the top person at Vanta earns roughly 10x what that role paid five years ago, and she thinks comp dispersion should be even wider. High-context people who make reasonable decisions, iterate on bad ones, and move fast are worth millions.
Cacioppo says an 80% growth business with 80% margins can return more cash over time than a 160% growth business with 30% margins, yet the market values the latter several times higher. Founders must keep doing what makes business sense and live for the fullness of time.
Cacioppo expected French software companies to use Chinese models out of nationalism, but they told her the switch came after access to an American model was cut off for non-Americans. They saw relying on one US provider as a business-continuity risk.
Cacioppo says governments already knew their processes were outdated, but AI has made modernization feel urgent. Officials and CISOs are using AI risk to ask for budget they might not have gotten for an enterprise risk overhaul three years earlier.
Cacioppo says Vanta connects to 400 tools and pulls data into a trust graph, or security data lake, that is broader than competitors'. That data foundation powers automation and agents and lets Vanta build new products faster.
Cacioppo admits the billboard was a placement order, not the multi-year contract she had joked about. A competitor paid more and the seller refused even triple because the company was cash-flow positive, leaving Vanta without the spot and its founder worried.
![Christina Cacioppo - Building an Enduring Software Company in the Age of AI - [Invest Like the Best, EP.496]](https://megaphone.imgix.net/podcasts/ef669774-cccd-11ed-889b-c36caad6646f/image/158efdddfb983d2678b3530d484e8aa2.jpg?ixlib=rails-4.3.1&max-w=3000&max-h=3000&fit=crop&auto=format,compress)