A glossary for AI-first product teams

Plain-language definitions for the terms behind AI-first product work: intent lead time, the Space Agent, the DVF lens, decision logs, and more.

Intent lead time

The gap between a decision being made and the first commit that acts on it landing. As coding agents collapse implementation time, this upstream gap becomes the real bottleneck, and it is the metric Stoa is built to reduce.

Space Agent

Stoa's in-room AI agent. It listens to a live meeting and captures decisions, drafts PRDs and user stories, tracks open questions, and can write and run code, all while the team talks.

DVF lens

Desirable, Viable, Feasible, from IDEO's design thinking. Naming which of the three a meeting is about clarifies who needs to be there, what decision is on the table, and what artifact should come out.

Last 10% problem

The edge cases and deferred decisions that evaporate when a call ends but resurface as rework later. Capturing them live, with their context, is cheaper than rediscovering them in QA.

Decision log

A structured, searchable record of what was decided, by whom, and why. It stops teams from re-litigating settled decisions and helps new members get up to speed by reading rather than asking.

Spec-driven development

An AI-first workflow that keeps the intent and specification next to the generated code, on the principle that understanding the why behind code matters as much as the code itself. SpecStory captures that intent automatically.

Local-first context

The idea that the context you accumulate, your decisions, questions, and artifacts, belongs to you in a queryable format you can sync to disk, version with your code, or feed to other agents, rather than being locked in a vendor.

Frequently asked questions