Spot-on about starting with what's already there! The sparring approach is underrated, AI as a thinking mirror is way more valuablethan template generation. I've found similar patterns in solutoin architecture where articulating the problem upfront saves hours of rework downstream. The distinction between definitional and factual content is clever too, keeps expectations realistic when dealing with existing landscape data vs conceptual structures.
Really interesting read. I put together a script to integrate "Archi" and each of the LLM engines so that Architects can push models directly to the chat bot and retrieve responses directly as "documentation" or "analysis" on the model. At the end of the day I guess it all comes down to the context and prompt you're using to deliver "value" from the model.
That sounds like a great experiment, thanks for sharing! Integrating Archi with LLMs so architects can push models and get structured documentation or analysis back is exactly the kind of practical AI use that moves EA work forward.
Spot-on about starting with what's already there! The sparring approach is underrated, AI as a thinking mirror is way more valuablethan template generation. I've found similar patterns in solutoin architecture where articulating the problem upfront saves hours of rework downstream. The distinction between definitional and factual content is clever too, keeps expectations realistic when dealing with existing landscape data vs conceptual structures.
Thanks, glad it resonated. I like your point about solution architecture as well: articulating the problem early usually saves time.
Really interesting read. I put together a script to integrate "Archi" and each of the LLM engines so that Architects can push models directly to the chat bot and retrieve responses directly as "documentation" or "analysis" on the model. At the end of the day I guess it all comes down to the context and prompt you're using to deliver "value" from the model.
https://smileham.co.uk/2025/07/02/ai-support-for-archi-models-using-gemini-and-jarchi/
That sounds like a great experiment, thanks for sharing! Integrating Archi with LLMs so architects can push models and get structured documentation or analysis back is exactly the kind of practical AI use that moves EA work forward.