Home / Insights
Insights
Original writing by Darshan R Krishnan on what actually works when technology meets a real business — and on backing the founders building it.
What actually matters in an early-stage founder — clarity, judgement, a bias toward building, and a real business underneath the technology.
Depth in one market, reach from another — and why building across borders is a stress test for real systems.
Most automations get measured on whether they run, not whether they helped. The two are very different, and the gap is where disappointment lives.
When the tools change every few months, hiring for current skills is a trap. What to select for instead, from someone running an AI services company.
The build is the visible cost of an AI automation, and usually the smallest one. A breakdown of the line items that never make it into the proposal.
Somewhere around 30 people, a services company stops being able to run on the founder's judgement. The failure is quiet and it's structural, not a hiring problem.
Some processes get worse when you automate them — not because the technology fails, but because of what the manual version was quietly doing. Five categories to leave alone.
Most failed AI automation projects weren't technical failures. They were correctly built solutions to processes nobody had examined. A view from the delivery side.