Research

Research on AI-native go-to-market

Long-form research and shorter notes from building go-to-market context systems in production - the theory the work runs on, and what broke along the way.

Why AI Stalls

Three-Piece Salvo · I of III · Diagnose → Architect → Operate

The Jagged Go-To-Market Frontier: Why AI Is Bad at GTM Tasks, and How to Fix It

AI won a Nobel Prize and surfaced 10,000 critical software vulnerabilities in a month, yet it still can't write a prospecting email your team would send. Two structural conditions explain why, and this page is the demonstration of the fix.

Read →

On Context

On Context · Why a revenue org needs a context system

Go-to-Market Context Is a System, Not a Document

Your team's AI output is only as good as the system that feeds it context. Most go-to-market orgs have a brain and no way to deliver it. Why go-to-market context has to be a system — with owners, cadences, and curation — not a document.

Read →Three-Piece Salvo · II of III · Diagnose → Architect → Operate

Go-to-Market Context: One Architecture, Many Systems

CLAUDE.md is an on-ramp, not a moat. Go-to-market context, structured as one architecture, many systems — conceptual, canonical, deterministic — for GTM teams that want AI to compound.

Read →On Context · The benchmark · Architecture has consequences

The Go-to-Market Context Benchmark: Which Architecture Produces the Best Output

We ran the same go-to-market tasks through six context architectures and graded the output blind. The shape of the context matters more than the amount — and the best output doesn't come from the most tokens.

Read →