Blog
Field notes on adoption, agentic engineering, governance, and the Karnataka GCC market.
Free AI training is now genuinely good and reaches agent orchestration and MCP. What it structurally cannot do is change your workflow, on your data, in your systems, under your governance.
Logins, completions and prompt counts measure activity, not outcome. Six metrics that hold up under scrutiny, how to baseline a workflow in a week, and how to instrument without surveilling.
Ask a vendor to name the workflow before the engagement starts, to say what they refuse to automate, and to show the harness they leave behind. Including the questions we would fail.
The policy targets 1,000 GCCs by 2029 and uniquely backs nano GCCs of 5–50 people. What that changes for who buys AI adoption, and what a 20-person centre can build in-house.
AI business cases fail because they cannot be falsified. Name one workflow, state a measured pre-baseline, and write the kill criteria in advance. A one-page template.
The market is quote-gated, so buyers cannot benchmark. What actually drives the number, what each engagement type buys, and our real ladder ranges published in full.
India has no standalone AI statute. It governs AI through existing law, coordinated by a new AI Governance Group. The practical obligations for a GCC, and a framework you can stand up in 90 days.
Enterprise RAG fails on real documents because retrieval, not generation, is the weak link. Document-aware ingestion, retrieval evaluated separately, and the recall metric that comes first.
Model benchmarks do not transfer to your workflow. Score every run against versioned cases on four axes, build a golden set from real tickets, and catch regressions between model versions.
