The Problem We Lived
If you have ever worked in enterprise architecture, you know the drill. A project team submits a design for review. It sits in an email thread for two weeks. Three architects weigh in with conflicting feedback. Someone references a decision made eighteen months ago — but nobody can find the original document. The review eventually passes, with caveats that nobody tracks.
This is architecture governance today. It is manual, inconsistent, and frustrating for everyone involved.
We built cajeX because we lived this problem for years and got tired of watching good architecture intentions die in email threads and spreadsheets.
What We Wanted to Build
We started with a simple question: what if architecture governance was a product, not a process?
Not a wiki page that nobody reads. Not a SharePoint folder of outdated PDFs. A real product with structure, automation, and accountability built in.
That led us to directives — the atomic unit of architecture governance. A directive is a single, structured, approved rule: a principle your organization follows, a decision that was made, or a guardrail that must not be crossed. Every directive has an owner, a lifecycle, and a clear scope.
Once you have directives, everything else follows:
- A knowledge base where you curate the raw material — standards documents, whitepapers, policies, reference architectures
- AI that extracts candidate directives from that knowledge, so you are not starting from scratch
- Architecture sessions where projects are reviewed against approved directives — automatically, by AI
- Findings with severity levels and concrete steps to fix issues
- Reports that give auditors and leadership a clear picture of governance health
Why AI Changes Everything
Manual architecture review does not scale. A senior architect can review maybe two or three projects a week thoroughly. Most organizations have dozens of active projects. The math does not work.
AI does not replace the architect. It handles the repetitive part — checking whether a project follows the 50 or 100 directives your organization has approved. The architect focuses on the judgment calls: the edge cases, the trade-offs, the strategic decisions that require experience and context.
We chose Claude as our AI backbone because architecture governance demands reasoning, not pattern matching. An AI reviewer needs to understand intent, weigh trade-offs, and explain its findings clearly. That is what Claude does well.
What Ships Today
Today we are launching cajeX in soft launch. The platform includes:
- Knowledge base management with file attachments and external links
- AI-generated directives from your existing knowledge
- Full directive lifecycle — draft, review, approve, deprecate
- Architecture sessions with AI-powered review
- Findings management with severity levels
- Dashboard and analytics
- Document management
- Reports and compliance exports
- Multi-tenant workspaces with data isolation
This is the foundation. We will be shipping improvements every week based on what we learn from real usage.
What Comes Next
We are building in the open. The changelog tracks everything we ship. If you are an architecture team that wants to try a different approach to governance, we would love to hear from you.
Start free — no credit card required.