Problem
- Brand compliance was becoming difficult to manage manually as the number of clients, industries and template variations increased.
- Teams manually checked marketing assets against brand rules and industry requirements.
What I did
- Designed an AI-based Brand Compliance Template System using classification models and rule-based validation.
- Supported classification across Insurance, Finance, Automotive, Marketing and Food.
- Worked with the technical team to separate asset classification from compliance validation.
- Designed confidence-based routing so low-confidence cases could go to human review.
Key decision
- We did not want the model making the final compliance decision.
- We started with industries where the rules were well understood instead of trying to support everything at once.
Outcome
- Moved compliance validation from a manual review activity into a product capability that could scale with the platform.
- Enabled compliant assets and templates to move directly into downstream marketing workflows.
Learning
- In enterprise AI, the model is only part of the product. The workflow around confidence, review, explainability and human involvement matters just as much.