
I use this space like a public diary — a place where I drop thoughts, observations, and whatever I’m reflecting on at the moment.

LLM evals look great on dashboards but miss where AI assistants really fail: in conversations. Learn why model-centric metrics are incomplete and what to measure instead

Introducing Cipher: The conversation intelligence engine that turns unstructured dialogue from your AI assistant into clear signals for product improvement.

Build a cost-effective Customer Intelligence System. Master data aggregation, build a product taxonomy, and avoid the LLM cost trap with a smart ML hybrid analysis approach.

A conversational walkthrough of how customer intelligence directly drives revenue, reduces support costs and saves teams countless hours by turning scattered feedback into clear, actionable insights.

In today’s product world, feedback is endless—tickets, reviews, calls, social mentions. The real challenge isn’t gathering it, but turning it into clarity. At Lexsis AI, we don’t just collect feedback; we translate it into actionable customer insight.

How building the “wrong” product taught me what the right one should be.