Named frameworks

Four concepts, with canonical sources.

The named ideas Plan B's writing introduces for thinking about brand in an AI-shaped landscape. Short definitions here; the arguments are in the articles.

Brand Sentinel

A framework for keeping brand distinctiveness intact when generative AI is in the production loop. Brand Sentinel treats brand governance as a set of precise, machine-legible inputs rather than a style guide humans can approximate.

Canonical article · Brand inside the machine

Model Drift

The gradual dilution of brand distinctiveness that occurs when AI systems are given just enough direction to be plausible, but not enough to be true. Model drift is cumulative — each generation lands close to the average, and the average does not sell.

Canonical article · Brand inside the machine

The Encode Layer

The layer where brand is translated into the concrete inputs that teams, partners, and machines actually operate on — prompts, rules, references, examples. If the encode layer is vague, everything downstream drifts.

Canonical article · The encode layer

Judgement

The argument that the commercial defensibility of a brand is increasingly the judgement behind its decisions, not the decisions themselves. Machines can reproduce decisions. They cannot yet reproduce the taste that chose which decisions to make.

Canonical article · On judgement