SEE IT IN ACTION
The Challenge
The review board meets monthly. Getting ready takes Monday and Tuesday. Baseline comparison, conformance report, trend analysis, executive summary. The board meeting takes two hours. The preparation takes two days. None of the preparation requires the board's judgment. All of it requires the steward's time.
When the primary steward is available and uninterrupted, the review package is complete. When they are traveling or pulled into a delivery engagement, the cross-element analysis gets abbreviated, the trend table gets skipped, and the board makes decisions from a lighter picture than usual. The governance process is as consistent as the person running it.
A governance violation that appears in three consecutive reviews is a systemic problem, not an isolated incident. When the trend table has to be built by hand from month to month, patterns that would warrant a board decision stay invisible. The finding appears. It gets noted. Nothing changes.
When you use AI Augmented Architecture for Repository Governance
Claude calls list_baselines, finds the previous review baseline, and runs compare_baseline against current state. New elements, modified elements, most-active packages, most-active architects: the delta picture is ready before the first coffee of the day. The steward adds organizational context. The health scorecard is done.
AI Power Tools for EA runs the rules sidecar against the current repository and categorizes every finding. Hard violations have a specific fix and a clear owner. Conventions are deviations from the preferred pattern, with the board deciding whether correction is required. Judgment calls require architectural review. The three-bucket categorization is what makes the report actionable rather than just a list.
Claude compares this month's findings against the previous review. Findings that have appeared in prior reviews carry a trend marker: "second consecutive review," "third consecutive review." The board reads the pattern, not just the instance.
The full review package: executive summary, per-bucket findings with rule citations, trend analysis, remediation roadmap, embedded diagrams. The steward reviews and adjusts. The report goes to the board from the model, not from a manual assembly process.
When the board chair asks which packages have the most governance debt during the meeting, the steward types the question into Claude. The answer comes back in thirty seconds from the model, with counts and trend context. The board trusts an answer from the model differently than an answer from recall.
AI POWER TOOLS FOR ENTERPRISE ARCHITECT
86 tools. Full read/write access to your Sparx EA model. Seven-day free trial. No credit card required to start.
HOW TO GET STARTED
Understand your starting point
Get AI Power Tools for EA running
Build the habit across your team
FREQUENTLY ASKED QUESTIONS
Yes. The review board still meets, still adjudicates findings, still decides the remediation roadmap, and still makes every architectural judgment. AI Power Tools for EA prepares the materials the board needs to have the right conversation. It does not make governance decisions. The board does.
Every conformance finding falls into one of three categories. Hard violations are findings that break a defined rule; they have a specific fix. Conventions are deviations from the preferred pattern; the board decides whether to require correction. Judgment calls require architectural review: a dependency risk, a scope question, a design choice that needs senior input. The categorization tells the steward what to fix before the meeting, what to flag for discussion, and what needs the board's judgment. A flat list of findings does not do that.
That is the typical starting point. The Plan tier includes a governance rules elicitation with your senior architects: reviewing historical findings, working through the rules that exist in the MDG, and encoding conventions that have lived only in your senior architects' heads. Your team reviews and confirms every rule before it is deployed. The process produces written, encoded governance rules as a deliverable in its own right.
Yes. After the Train tier, the governance review prep workflow is documented and any trained architect can run it. Architects can also run the review during their modeling sessions to self-check their work. The quality of the output is consistent because the rules sidecar and the toolchain do not depend on individual knowledge. The process is no longer fragile.
The report includes an executive summary with health scorecard numbers and top-line conformance status, a per-bucket findings section with rule citations and suggested actions, a trend table comparing current findings against the previous review, a remediation roadmap listing hard violations with suggested owners and target dates, and embedded diagrams for the elements most likely to be discussed.
Watch the demo, or schedule a call and we will walk through it against yours.
Schedule a Discovery Call