Tessl and Pathmode both work with AI agents, but they operate at different layers of the stack.
Tessl is an agent execution framework. It defines how agents should implement code — Usage Specs that describe APIs, data models, and technical contracts. Tessl's world starts when someone has already decided what to build. It answers "how should the agent build this?" with rigorous, machine-readable implementation specifications. Think of it as a technical blueprint layer that sits between a feature decision and the generated code.
Pathmode is an intent definition layer. It answers the question that comes before Tessl: "what should we build, for whom, and how do we know it worked?" Pathmode starts from user evidence — friction signals, support tickets, interview quotes, behavioral data — and compiles that into IntentSpecs. These specs define the objective, expected outcomes, edge cases, and verification criteria. They're the strategic layer that feeds execution layers like Tessl.
The relationship is complementary, not competitive. Pathmode defines why something should be built and what success looks like. Tessl defines how an agent should implement it at the code level. In a mature AI-native team, Pathmode produces the IntentSpec, and Tessl (or a similar agent framework) consumes it as upstream context for generating implementation specs.
Where the tools diverge is audience. Tessl is built for engineering teams — developers who think in APIs and code contracts. Pathmode is built for cross-functional product teams — PMs, designers, and engineers who need to align on user intent before any code is written. Tessl assumes the "what" is settled. Pathmode exists because it rarely is.