The software teams we seeing getting the most out of AI are putting AI agents on their org charts, prioritizing customer outcomes over faster shipping, and adopting unconventional ideas about how to measure success within their UX.
Nick Olsen spent 20 years as an engineer and CTO building and scaling vertical software before joining us as head of Mainsail AI Labs. He talks through what these moves look like in practice with former software operator Greg Head on the Practical Founders podcast.
In the episode, Nick draws a useful baseline for what founders should expect of their product & tech teams in mid-2026 and explains the framework AI-forward development orgs are building around agentic coding — tools, talent, CTO leadership, and even rewrites.
Worth a listen if your team has the tools but you’re still figuring out whether you’re solving the right customer problems with them.
Key Takeaways
- Table Stakes — AI-first engineering is no longer appears to be optional for serious SaaS companies. It’s becoming the minimum operating standard.
- Roles Are Collapsing — Product, design, and engineering are moving closer together as AI removes slow handoffs and bottlenecks.
- Velocity Misleads — Shipping 3x more features doesn’t matter if customers don’t want them. Outcome beats output.
- Judgment Wins — AI makes building cheaper and faster. Strategic product judgment becomes more valuable, not less.
- Agents Matter — Strong teams increasingly treat AI agents like teammates with jobs, accountability, and measurable performance.
- UX Is Changing — Customers may increasingly stop logging into SaaS apps and interact through agents instead.
Learn more about how Nick’s team works alongside our portfolio companies to commit code, ship features, and shift to treating AI as the way product and engineering teams operate: Mainsail AI Labs.
Quotes from Nick Olsen, Head of AI Innovation at Mainsail
“We are well past any point of skepticism or any engineer that says, Hey, this isn’t gonna be here forever, or this isn’t my thing. Engineers no longer should be starting in the traditional code editor, the IDE. Engineering has moved to a much more agentic approach where you’re starting in the terminal or some sort of Claude Code or Codex app.
How we’re building software has completely and radically shifted. Companies are starting to move past treating AI as a tool. They’re starting to treat AI as the way that they actually operate. It’s not a tool like Excel is for a financial analyst, but that’s actually where I start.
We’re trying to see people transition to treat AI as an actual teammate and actually affecting change within their organization.. And that’s from the product and management side as well as to the engineering side.“
