Knowledge Base Gaps

Knowledge-base gaps are the places where customers ask questions but Woes does not have enough trustworthy context to answer. Fixing gaps improves agent quality, reduces handoffs, and makes operator work more repeatable.

Gap Signals

Signal	Meaning
Repeated handoffs	Customers ask something the agent cannot safely answer.
Low-confidence replies	The agent found related context but not enough evidence.
Operator corrections	A human had to replace or rewrite the answer.
Survey complaints	Customers indicate the answer did not solve their problem.
Issue clusters	Many issues reference the same endpoint, auth flow, or product area.

Fix Workflow

Identify the missing answer

Start from a conversation, issue, survey response, or analytics trend.

Find the source of truth

Prefer product docs, API specs, SDK examples, or internal runbooks over memory.

Add or update source content

Improve the original docs when possible, then rescan the source in Woes.

Test the same question

Ask the agent the original customer question and verify the answer is specific and grounded.

Close the loop

Add tags or issue links so future analytics show whether the gap is improving.

Common Gap Types

Gap	Better source material
Auth confusion	Token examples, scopes, headers, and common 401/403 causes.
Missing request fields	Request schema, required field table, and valid enum values.
Webhook uncertainty	Event names, delivery rules, retry behavior, and payload examples.
SDK mismatch	Current SDK examples and migration notes.
Account-specific answers	Verified customer identity, operator notes, or portal context rather than public docs.

What Not To Do

Do not lower confidence thresholds just to make the agent answer more often. Better coverage should come from better context, not looser safety behavior.

Last modified on June 30, 2026

​Knowledge Base Gaps

​Gap Signals

​Fix Workflow

​Common Gap Types

​What Not To Do

Knowledge Base Gaps

Gap Signals

Fix Workflow

Common Gap Types

What Not To Do