Skip to main content
Octolane’s company brain captures what your team learns as it sells: messages that get replies, recurring objections, and the patterns behind closed deals. It combines shared company knowledge with personal context for each rep and uses that context to recommend next actions. Find it under Knowledge in the app. Company brain is included on every plan.

Review what Octolane learns

Automatically extracted learning suggestions wait for review before they become approved knowledge. Explicit edits you make in the app or authorize through MCP can apply directly. Review a proposed learning, correct it when necessary, and approve what should become part of your team’s knowledge. Changes are version-controlled so you can inspect what Octolane knows and how it changed. This review of new knowledge is distinct from routine CRM updates and the action-specific approval settings.

Ask questions

Ask questions in plain English, such as:
  • What objections do we hear most often?
  • Which follow-up approaches have worked for our team?
  • What have we learned from recent won and lost deals?
AI agents can query the same company brain over MCP, using the connected user’s permissions. The client discovers the currently available tools and schemas, lists blocks with memory_list, and reads their content with memory_get. Use scope: "org" for shared knowledge, "user" for personal knowledge, or "both" for both. Pending suggestions are excluded by default. A block can apply to the workspace or a specific account, contact, or opportunity. External agents must read applicable blocks explicitly; activation tags do not automatically supply knowledge to an MCP client. See the memory tool reference. Knowledge writes use memory_upsert and memory_delete. Shared writes require administrator privileges. A request to answer a question does not authorize editing the company brain.

History and data use

Company brain version history is retained for 90 days on Founder, one year on Pro, two years on Team, and a custom period on Enterprise. See Plans and pricing. Your team’s information is not pooled into a shared model or used to improve other customers’ agents. Learning from your team’s approvals, edits, and outcomes is used for your own team. See Security.