"Context: Moat or Wall?" Test
When you need this method
A company presents its AI foundation, context repository, playbooks, agent-readable knowledge base, as its central competitive advantage. Whether that is true hinges on a question rarely asked: if every competitor must and can build the same foundation, it is not an edge but the ticket required just to keep up.
Approach
- 1Name the claimed AI/context foundation concretely: which data, playbooks, or workflows are meant?
- 2Check whether every company in the competitive set is building or could build the same foundation, then it is a wall: necessary, but not differentiating.
- 3Test against moat criteria: is the underlying data proprietary? Does the context compound with usage? Do real switching costs arise?
- 4Check the effect on metrics: does the claimed edge show up in retention or expansion or only in the self-description?
- 5Report the result separately: wall components as mandatory work, moat components as the defensible core.
Typical application
A typical case: a B2B software vendor argues to investors that its internal AI context system is its central edge. The test dissects the claim: competitors are building comparable playbook repositories and agent workflows, a wall that must be maintained just to keep pace. What remains as a genuine moat candidate is a dataset from years of customer usage that measurably improves output quality and would be lost in a vendor switch. The investor story becomes narrower, but robust.
Limits and counter-indications
The test is a Convios-internal extension of the regulatory density test for AI foundations and has not been independently validated. The line between wall and moat shifts with the technology, what looks proprietary today may be standard tomorrow. The test assesses defensibility, not usefulness: a wall still has to be built.
How to measure impact
A suitable observation: whether the context foundation shows up in switching costs and retention, for instance, measurably worse results from a competitor setup lacking your data.
Related methods
Tools for this
Sources
- 1.The Red Queen in Organizational Evolution (opens in a new tab) · Strategic Management Journal 17, via Stanford Graduate School of Business (William P. Barnett, Morten T. Hansen) · 1996 · academic and scholarly literature · supports the underlying mechanismShows, in a longitudinal analysis of a single industry spanning roughly ninety years, that competitive experience raises the probability of survival and that one firm's learning triggers learning among its rivals, while the study also documents competency traps and diminishing effects. That the relative gap stays constant is an interpretation made for this method and is not measured by the study.
- 2.The New Business of AI (and How It's Different From Traditional Software) (opens in a new tab) · Andreessen Horowitz (Martin Casado, Matt Bornstein) · 2020 · investment, consulting and analyst firms, industry bodies and public agencies · provides the contextArgues from a venture investor's portfolio experience that models and data on their own rarely defend a position, because architectures are developed in the open and data often belongs to the customer or becomes table stakes over time. The assertion rests on industry observation and portfolio metrics, not on a controlled study.
Origin: Convios