Visibility
See what was created
Stakeholders must be able to see and interact with the solutions. Not screenshots—real, clickable prototypes.
How This Happens
- Deployed preview URLs
- Live demo sessions
- Storybook component views
Back to EEM Overview
05 Governance Gate
The Governance Gate
Human judgment decides what ships. The gate is required; the ritual is yours.
What is the Governance Gate?
The Governance Gate is a checkpoint between Execute and Production where human stakeholders review AI-generated solutions, make judgment calls, and decide what moves forward. It's not bureaucracy—it's the moment where experience, strategy, and business context guide which experiments become real products.
- Execute
- Multiple solutions
- Governance Gate
- Human judgment
- Production
- Selected solution
Critical Positioning. The Governance Gate sits after Execute, not before. This is intentional: humans judge real options, not hypothetical concepts. You're not approving plans—you're selecting from working solutions.
Why Governance Matters in EEM
- Quality Control. AI can build fast, but not everything built should ship. Governance ensures solutions meet quality standards, brand requirements, and user needs before reaching production.
- Strategic Alignment. Multiple solutions might all "work" but serve different strategic directions. Governance brings business context to decide which direction best serves organizational goals.
- Design System Evolution. When execution creates something new (a pattern, component, or interaction not in the system), governance decides whether to adopt it into the system or treat it as one-off.
- Risk Management. Human judgment catches things AI might miss: brand tone issues, accessibility concerns, competitive implications, regulatory considerations, or cultural sensitivities.
Without governance, you have automation. With governance, you have AI-augmented decision making.
Four Governance Activities
Every governance session involves four key activities. The framework specifies these must happen—how you implement them is up to your team.
Comparison
New vs. existing system
Identify what's new. Compare solutions to existing design system patterns. What follows the system? What deviates? What's genuinely novel?
Key Questions
- Does this follow existing patterns?
- What components are new?
- Are deviations intentional or drift?
Decision Capture
Record what and why
Document the decision: what was approved, what was rejected, and critically, why. This context informs future execution.
What to Capture
- Which solution was selected
- Reasoning behind the choice
- Required modifications before production
System Update
Evolve the design system
If governance approves a new pattern, it should be added to the design system. This is how the system evolves—through validated execution, not theoretical expansion.
System Update Actions
- Add new components to Storybook
- Update token files
- Document new patterns
Possible Governance Outcomes
Not every solution moves to production. Governance can result in four different outcomes, each valid and useful.
Approve for Production
Approve
Solution is ready. Move to production with any minor adjustments. This is the path forward.
- Next Step
- Production Engineer takes over for deployment preparation, optimization, and release.
Iterate Further
Iterate
Direction is right but execution needs refinement. Return to Execute with specific feedback for another pass.
- Next Step
- Document what needs to change. Execute again with refined constraints. Come back to governance.
Reject
Reject
Solution doesn't meet needs and isn't worth iterating. This is valid—not every experiment succeeds. Learn and move on.
- Next Step
- Document why. Archive learnings. Consider if the problem is worth re-approaching with different inputs.
Pivot
Pivot
Solutions revealed the problem was wrong or strategy needs to shift. Return to Input stage with new strategic direction.
- Next Step
- Reframe the problem. Adjust strategic inputs. Start a new execution cycle with different constraints.
Who Participates in Governance?
Product Strategist
Evaluates strategic fit and business alignment
Execution Designer
Presents solutions and explains tradeoffs
Design System Steward
Assesses system compliance and evolution
Key Stakeholders
Business owners who can make go/no-go decisions
Keep governance lean. Too many participants slows decisions. Include only those who can evaluate and decide. Others can be informed after.
Governance Checklist
Use this checklist to guide governance discussions. Not every item applies to every solution—use judgment.
Strategic Fit
- Does this solve the stated problem?
- Does it align with product strategy?
- Is this the right problem to solve now?
- What's the expected business impact?
Design System Compliance
- Does it use existing components correctly?
- Are deviations intentional and justified?
- Should new patterns be added to the system?
- Is brand consistency maintained?
User Experience
- Is the flow intuitive and clear?
- Are accessibility requirements met?
- Does performance feel acceptable?
- What did stakeholder testing reveal?
Production Readiness
- Is the code quality sufficient for production?
- What production hardening is needed?
- Are there security or privacy concerns?
- Can this be deployed incrementally?
Common Governance Pitfalls
Anti-Patterns
- Governance as Bottleneck. Making teams wait days or weeks for governance decisions. Speed is the point—don't let governance become the slowest part of the process.
- Design by Committee. Too many voices without clear decision authority. Governance needs a decision maker, not a democracy.
- Perfect vs. Good Enough. Demanding perfection before production. Remember: production is deliberate but not permanent. Ship and iterate.
- Skipping Documentation. Making decisions without capturing rationale. Future you (and your team) will need to understand why.
Best Practices
- Time-Boxed Sessions. 30-60 minutes maximum. Come prepared, make decisions, move on. Extended discussions signal unclear criteria.
- Clear Decision Authority. One person who can say "yes" or "no." Others provide input, but someone owns the decision.
- "Good Enough for Now". Bias toward shipping. Can this be improved in production? Then ship it and improve it.
- Async When Possible. Not every decision needs a meeting. Async review with clear deadlines for simple cases.
The Feedback Loop
Governance doesn't end when a solution ships. Production generates data that becomes input for future cycles.
- Production Deploys
- Solution goes live to users
- Data Collected
- Usage, feedback, performance metrics
- Becomes New Input
- Informs next execution cycle
This continuous feedback ensures the system keeps learning and improving. What works? What doesn't? The data tells you, and the next Execute cycle responds.
Ready to Implement Governance?
Governance is where strategy meets execution. Let's design a governance process that maintains quality without sacrificing the speed EEM enables.
Continue exploring: