Problem:
The AI-generated company background/description feature works by prompting an LLM directly rather than pulling from a verified source like the company's website URL on the account record. When a company name is shared by multiple organizations, the AI can generate a description for the wrong company entirely — and today there is no way for admins or CSMs to correct it.
Real example:
For our Account A Gainsight support confirmed the AI generated a description for Account B a UK-based job-matching platform - a completely different company that happens to share part of the name. Support confirmed via Engineering that:
- The description is generated from general AI/web knowledge, not from the company URL or internal timeline/account data already stored in Gainsight.
- There is currently no way to manually edit or correct a generated description.
- No way exists to feed the AI the correct company URL to ground its answer.
Why this matters:
- CSMs & leadership rely on this field for quick context before calls/QBRs — an inaccurate description actively misleads rather than helps.
- We already have the correct company URL/website stored on the Company object in Gainsight. It seems reasonable that the AI should be grounded in that data (or at least allow it as an input) rather than guessing from name alone.
- Name collisions aren't rare - many companies share names with unrelated entities (recruitment platforms, consumer apps, etc.), so this isn't an edge case.
Requested solution (either would resolve this):
- Allow admins/CSMs to manually edit or override an AI-generated company description, similar to other editable fields, OR
- Let the AI generation process use the Company's website URL (already stored on the account) as a grounding input instead of/alongside name-based lookup, so it doesn't have to guess.
Impact if implemented:
Improves trust and accuracy of AI-generated content across the CS org, especially for companies with generic or non-unique names.
