Notes on enterprise software, AI, and human behavior

Curious about the real. Skeptical of the shiny.

Two decades inside enterprise software taught me that the gap between a great demo and a working system is where all the interesting problems live. I write about that gap.

Read the writing

Watching

Cloud Economics·Service Subscription·Personal Software vs SaaS·Applications to Agents·Agentic AI in the Enterprise·Economy of Tokens·Cloud Economics·Service Subscription·Personal Software vs SaaS·Applications to Agents·Agentic AI in the Enterprise·Economy of Tokens·

Writing

View all →
Adoption

The adoption gap nobody budgets for

Enterprises buy software at the speed of procurement and absorb it at the speed of trust. The gap between those two velocities is where most transformation programs quietly die.

June 1, 2026
Frontier models

Agentic AI meets the change request

The most interesting thing about agentic AI isn't what it can do autonomously. It's what happens when it hits the first approval gate.

May 18, 2026
Hyperscalers

What hyperscaler earnings say about enterprise patience

Every quarter, AWS, Azure, and Google Cloud report AI-driven momentum. Every quarter, most enterprises are still deciding which workloads to move first.

May 5, 2026
Human behavior

Why good products lose to familiar ones

Superiority is not enough. The switching cost is never just the price of the new tool — it's the accumulated fluency with the old one.

April 22, 2026

The Conversation

The interesting parts of the work rarely fit on a page. Ask what didn't make the cut.

What has Arun actually built?Ask →

Working Theses

My bet is that software is getting personal — built around one person instead of a million-user average. The only way to actually test that is to build the things, not just write about them. These are the ones I'm in the driver's seat of.

Learning systems that respect attention

Building micro-courses to test whether shorter, behaviorally-informed formats actually change what people retain.

Explore →

AI assistants with real context

The chat on this site is one experiment: how much does deep context change the usefulness of an assistant?

Explore →
Arun Pillai

Interested in what actually works — not what looks good on a slide.

Twenty-two years in enterprise software, starting in engineering and ending up as a global program leader. I launched SAP Preferred Success from a blank page and scaled it to a large enterprise cloud customer base. The core belief that drove it: adoption is a human problem before it's a technical one.

Currently thinking about cloud ERP, agentic AI, and why the absorption rate so rarely matches the announcement rate.

Read more →