Implementation BlueprintExecution layer

OzBrain Client Onboarding Sprint (5-7 days)

A 5-7 day sprint to stand up a shared knowledge base for a client's AI agents, configuring article structure, agent connections, and governance rules so Claude, ChatGPT, and Cursor work from one current source of truth. Time: 5-7 days.

By InnovaAI ResearchPublished

Blueprint

OzBrain Client Onboarding Sprint (5-7 days)

A 5-7 day sprint to stand up a shared knowledge base for a client's AI agents, configuring article structure, agent connections, and governance rules so Claude, ChatGPT, and Cursor work from one current source of truth.

Prerequisites
  • Client has active accounts for at least one AI agent (Claude, ChatGPT, or Cursor)
  • Access to OzBrain admin panel and API endpoint (https://ozbrain.com/api/mcp)
  • Client's core services, FAQs, and process documentation available in digital form
  • A sandbox OzBrain environment for testing agent connections before production
  • Client stakeholder identified to approve article taxonomy and freshness rules
Execution Timeline
  • 1.Create OzBrain account and configure initial brain structure with article categories matching client's service lines
  • 2.Set up MCP connector for Claude using the API endpoint and test read/write in sandbox
  • 3.Define provenance fields (agent name, timestamp) and freshness thresholds for articles
  • 1.Draft first 10-15 structured articles covering client's core services, FAQs, and internal processes
  • 2.Enable automatic routing so new knowledge lands in the correct article based on keywords
  • 3.Connect ChatGPT via custom connector and verify it can read and write to the same brain
  • 1.Configure conflict surfacing rules to flag when agent writes disagree with existing content
  • 2.Test multi-agent scenario: Claude and ChatGPT update the same article and observe version tracking
  • 3.Document write-time size discipline settings to keep articles within limits
  • 1.Integrate Cursor for code-related knowledge (e.g., API docs, code snippets) and test retrieval
  • 2.Set up sharing permissions for the client's team on brains they own
  • 3.Run end-to-end workflow: agent reads article, updates it, and exports markdown for review
  • 1.Populate remaining articles to reach up to 40 total, covering client's full service catalog
  • 2.Create a freshness review schedule and assign ownership to client staff
  • 3.Train client team on how agents maintain the brain and how to resolve conflicts
  • 1.Perform a live test with client's actual agents to validate routing and conflict handling
  • 2.Adjust article taxonomy based on observed gaps or misrouted knowledge
  • 3.Deliver admin panel walkthrough and document governance rules
  • 1.Finalize markdown exports of all articles for client backup
  • 2.Review usage against Free plan limits (50 articles) and recommend Pro or Max if needed
  • 3.Hand off with a runbook covering agent connections, maintenance, and scaling
$0 to $99 per month for OzBrain subscription (Free up to 50 articles, Pro $20 up to 300, Max $99 for larger bases)5-7 days
ROI Logic

With a $1,800 setup fee and roughly 16 hours of effort, the agency's effective hourly rate exceeds $110, while the client's OzBrain subscription costs between $0 and $99 monthly. The agency can also add a monthly retainer for ongoing brain maintenance, creating recurring revenue from a low-cost tool.

Deliverables
  • Configured OzBrain knowledge base with up to 40 structured articles
  • Connected agents (Claude, ChatGPT, Cursor) with verified read/write access
  • Documented provenance and freshness rules for automatic maintenance
  • Markdown export of all articles for client backup
  • Admin runbook covering agent connections, conflict resolution, and scaling
Definition of Done

The client's AI agents can read and write to a shared OzBrain brain without manual context copying, and all articles meet the agreed freshness and provenance rules.