Ilustration courtesy of ChatGPT
This four-part blog series from 27Global provides an overview of an internal project to plan, design, and implement an AI-driven solution. Follow along to learn about our use case, our challenges, and our successes.
Part 3: Real Progress, Real Friction
At the 75% mark, Dr. Doc was producing real outputs. But limitations in tools and scope forced some hard tradeoffs.
Jesse completed the Cosmos DB plugin, tested it with a sample portal DB, and integrated it into Rivet. This milestone enabled:
– Real vector retrieval
– True embedding-to-diagram updates
– Local PR-triggered validation cycles
Prompt testing also improved with more structure and some exploration of Claude and Anthropic models for future work.
Patrick created a Rivet setup guide and launched a decision log to track architecture choices. Kelley finalized and uploaded the infrastructure diagram, and tickets were created to track test coverage and edge cases.
The diagram update process was measured at 2–3 minutes per cycle—solid for an MVP.
Due to tool limitations and timeline pressure, we formally:
– Dropped new diagram generation (updates only)
– Froze model switching experimentation
– Deferred any CI/CD or pipeline automation
We weren’t shipping production code, but we were proving that diagram updating could be agentized, even if clunky.
Check back next week for Part 4: Dr. Doc Delivers (and Reflects)
About the Author
John Marney is a visionary director at 27Global, boasting a rich tapestry of experience in steering businesses towards digital excellence. With a keen eye for innovative solutions, John has played a pivotal role in leveraging cutting-edge technologies to empower clients in their digital transformation journey. His passion for bridging the gap between business objectives and technological advancements continues to drive 27Global’s commitment to delivering unparalleled software, cloud, and data solutions.
Want to learn more? Contact us at info@27global.com
