# Issues and playbooks

> How benchmark gaps become persistent issues, how playbooks turn them into content actions, and how later full benchmarks measure change.

- Section: Core concepts
- Updated: 2026-07-13
- Canonical: https://knitknot.ai/docs/issues-and-playbooks/
- Publisher: KnitKnot, the AI Presence Management platform (https://knitknot.ai)

---

The report tells you where you stand; issues and playbooks are how you change it. This is the "fix" half of the benchmark → report → fix → re-run loop.

## Issues

Material gaps can become **issues** — persistent records with evidence attached. The main kinds include:

- **Factual errors** — AI states something that conflicts with available company facts and supporting receipts.
- **Outdated claims** — a statement conflicts with dated evidence or a newer canonical fact.
- **Missed features** — a capability you have that AI says you lack, often while crediting a competitor with it.
- **Losing comparisons** — head-to-heads where AI consistently picks a competitor, with the cited reasons.
- **Visibility gaps** — categories where competitors surface organically and you don't.

Issues are prioritized using evidence such as reach, severity, demand, concentration, and source ownership. Each issue resolves to the right evidence grain: a claim and quote, captured evaluation, cited or owned page, coverage record, or site check.

Issues keep their identity and history across qualifying full benchmarks. Reach can fall to zero, and an issue can become dormant after repeated zero observations, without erasing the record. If the same signature returns, KnitKnot redetects the existing issue instead of creating a disconnected alert.

## Playbooks

A **playbook** turns one or more issues into a concrete create, revise, or repair action. Page-touching playbooks provide a ready-to-edit brief with relevant evidence, buyer vocabulary, structure, and verified proof points where available. Technical playbooks provide a remediation plan instead of article copy.

The source view can provide the bridge: an issue describes *what* is happening, reliable citation and content evidence show which pages are involved, and the playbook describes *what to ship*. A claim without a reliable citation signal remains unattributed.

## Measurement after shipping

When you mark a playbook shipped, KnitKnot records the before-state when a completed full benchmark is available. After a later full benchmark completes, the impact view compares observed score, answer, citation, and linked-issue changes. Direct citation pickup is shown when it exists; other movement remains descriptive rather than being presented as proof of causality.

That loop — gap, intervention, later comparable measurement — is the difference between monitoring your AI presence and managing it.