Recall Debt Compounding

Recall debt is the accumulated cost of context an agency never wrote down: every client architecture decision, brand constraint, or debugging fix that lives only in a closed chat session must be re-derived the next time an agent touches the account. The debt compounds because each re-prompt consumes senior review time and each re-derivation risks a slightly different answer, so delivery drifts while the retainer stays flat. The framework asks agencies to price recall debt explicitly: count the prompts per project that exist only to restate what was already known, then compare that against the cost of a memory layer. A coding agent that re-learns a client's deployment constraints every Monday is the clearest case, which is why session-recall tools such as Skillsync and project-note stores like Knownbase exist. Forrester's separation of memory from prediction and decision functions gives the budget line a name, and the 105 of 175 AI tools that changed monetization terms between June and October 2026 show why that memory should not sit inside a single vendor's walls.