AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

Tenex, a firm that builds AI systems for large companies, keeps plans and decisions in Git beside the code so that AI agents can read why work was done.

Tenex is a 75-person New York firm that helps large companies adopt AI and builds AI systems for them. On the Latent Space podcast, cofounder Alex Lieberman and CJ Hess, the firm's first hire, described how its engineers work with AI agents and how the firm pays them. Engineers are paid partly by story points, and the firm keeps its plans and decisions in the code repository.

• Tenex's shared toolkit, which it calls MetaHarness, combines agent skills, project documents and a command-line tool, and plans, decisions and AI agents' logs are committed to Git beside the code.
• Hess said most engineers use one AI model to write code and a second AI model from another maker to review it, passing the work back and forth.
• Hess estimated that output per engineer roughly doubled this year, measured in story points, and the firm recalibrates the value of a point about once a quarter.
• Lieberman said inflated estimates have not been the main problem. The firm's client-facing technical consultants are rewarded for customer retention, so they push back on work that does not serve the client.

An engineering team can commit plans and decision records to the repository so that colleagues and AI agents can both read the reasons for past choices. This is the firm's own account of its practice, given in an interview, and its productivity figure rests on story points it sets itself.

Derived by Working Surface from the article; more in the Playbook. Source line: When an agent comes back to a task, it can read not just what was built, but why the team chose to build it that way.

22 October 2026: as AI agents make building cheap, teams are moving effort to deciding what to build and writing down the reasons where people and AI agents can reuse them.