Client Knowledge (the client brain)

Client Knowledge is a structured, per-client memory. Every fact, preference, decision and piece of context your team learns about a client lives in one place: the Knowledge tab on the client's page. Over time each client builds a "brain" your whole team (and later, your AI tools) can draw from.

Why not just notes?

Notes are free-form and pile up. Knowledge entries are atomic and categorized, so they stay findable and reusable:

• One entry = one fact ("Invoices go out on the 1st, NET-15, CC the PA")
• Every entry has a category (Overview, Contacts and Roles, Preferences, Decisions, Meeting Notes, Technical, Commercial, History) and optional tags
• Entries carry their author and last-updated date, so you know how fresh a fact is

Your existing client Documents and Notes stay where they are; the Knowledge tab links to both, so everything about a client is reachable from one screen.

Confidentiality tiers

Every entry has a visibility level:

• Team: visible to everyone who has access to that client's record.
• Confidential: visible only to admins and the person who wrote the entry. Confidential entries are filtered out at the database level, not just hidden, and they are excluded from AI exports unless an admin explicitly includes them.

Clients never see the Knowledge tab. It is your internal brain about the relationship.

Working with entries

1. Open a client and go to the Knowledge tab.
2. Click Add entry, give it a title, pick a category and a visibility level, write the fact, and optionally tag it.
3. Filter by category with the chips at the top. Edit or delete your own entries at any time; admins can manage all entries.

The AI-ready brain export

Everything the app knows about a client can be compiled into a single structured payload, ready for AI ingestion:

• In the app: admins see "Brain JSON" and "Brain MD" links on the Knowledge tab.
• From the command line: php artisan vivotics:client-brain {clientId} prints the JSON; add --format=md for markdown or --with-confidential to include confidential entries.

The export contains the client profile, a projects summary, document names, public notes and every knowledge entry as a separate chunk with metadata (category, tags, confidentiality, author, updated date). That chunked structure is what retrieval pipelines need, so when you connect an AI assistant later, no data rework is required.

Private and password-protected notes are never included in any export.

Frequently asked questions

Who can add entries? Anyone who can open the client's record. Editing and deleting is limited to the entry's author and admins.

Can I change an entry from Team to Confidential later? Yes, edit the entry and switch the visibility level. The change applies immediately.

Do confidential entries appear in the brain export? Only when an admin explicitly requests them with the confidential flag. Default exports are team-tier only.