Bold Crow AI Systems
Beyond client work, Bold Crow builds its own internal AI systems to run the agency itself, testing new AI capabilities on ourselves before offering them externally. As internal product tester and feedback lead, I shaped four systems from early builds through revision, working directly with the founder on requirements and testing.
Challenge / Problem
A small AI-forward agency needs to prove out AI-driven workflows internally — cheaply and quickly — before it can credibly sell them to clients, without a dedicated QA or product team to stress-test each new system. Our goal is also to stay as lean as possible, leveraging agents to complete tasks on our behalf that we then review and approve.
What We Did
Served as the primary internal tester and feedback lead across four proprietary AI systems:
- Blog AI System — Led research, requirements, and testing for a pipeline that synthesizes multiple source materials into publish-ready articles with SEO metadata, structured headers, and automated internal/external cross-linking pulled from Bold Crow's knowledge graph (internal tool, no public site).
- BamBam — AI-powered creative and social content system generating on-brand marketing assets from a client's existing design system (internal tool, no public site).
- Docalign → — AI knowledge-alignment platform centralizing scattered documentation into a searchable, conversational hub.
- RookCMS → — MCP-native CMS embedding Claude directly into website publishing workflows.
Impact / Results
- Blog AI System cut production time from 2 days to 2 hours
- All four systems moved from internal tools to offerings available to agency partners and clients
- Testing and feedback shaped multiple rounds of revisions ahead of each system's release