AI implementation cases: how I thought through the problems
Not a catalog. How I thought through the tasks.
Two AI implementation cases and one demo by Andrew Maryasov. Zoho Migration: a client's financial calculator moved from Zoho Creator to a custom stack in 8 days instead of 4–6 months. AI Cross-sell Stylist: an AI stylist for a chain of 30+ clothing stores. SmileClinic (demo): an AI pipeline builds a full proposal package in 6 minutes. The strategy is mine; implementation runs through Auspex and Grow2.ai.
Every case is about the decision, not the tool: where money was leaking, what I decided and what the takeaway was.
Zoho → a modern stack
A client's financial calculator: from Zoho Creator to a custom stack in a few days.
Case →An AI stylist for a retail chain
Cross-sell as a stylist’s advice, not «buy more»: collections, accessories, the right moment.
Case →Demos
A demo on real data, not a client project: it shows how the pipeline goes from a brief to a finished package.
Cases with numbers — at Grow2.ai and Auspex.
I build the strategy; the teams implement it. Full write-ups with numbers are published on their sites:
Apart from client projects, 8 proprietary systems of my own run every day, including a multi-agent orchestrator, an autonomous AI sysadmin, a fleet of production agents, a personal knowledge OS, verbatim conversation memory, a remote AI assistant and an AI code assistant.
Want to work through your own task?
30 minutes of conversation — and we see where the lever is hidden in your process. Behind me — 28 AI projects across 14 industries. We build the strategy together, implementation — through Auspex or Grow2.