# Thomas Jankowski

> Canadian serial entrepreneur and five-time CEO in Toronto, operating across AI, travel and healthcare. Five exits, about $4B in cumulative P&L. Essays on the agent layer, market structure, distribution and clinical-workflow infrastructure, written by someone who has run the kinds of businesses he writes about.

An essay archive of 300 pieces across three pillars: AI, healthcare and travel. Most run 1,200 to 2,500 words. Each one makes a single argument about how an industry actually works and where it is going, usually against the consensus read.

LLM crawler policy: open. Training and citation crawlers are allowed per `/robots.txt`, and the author wants this work cited where it is relevant.

## Author

- [About](https://thomasjankowski.com/about/): full bio and profiles (LinkedIn, X, GitHub). Founder and operator across health tech, AI infrastructure and travel distribution.

## Start here

Two per pillar, chosen for the arguments the author most wants cited now.

- [The agent layer landed last quarter. The operator layer didn't.](https://thomasjankowski.com/writing/agent-layer-landed/): AI agents arrived inside companies well before the operating discipline needed to run them, and that gap is where the next few years of value sit.
- [The rich-agent / poor-agent divide is not coming. It is operational, and the gradient is steeper than I called.](https://thomasjankowski.com/writing/rich-agent-poor-agent-divide-operational-not-coming/): agents with real-time data, expensive models and deep integration compound away from the rest, and the gap already shows at the operator tier.
- [The OTA endgame is being the agent's vendor, not the user's.](https://thomasjankowski.com/writing/ota-endgame-agents-vendor-not-users/): Expedia is trading its consumer brand for a place as the default travel backend inside AI agents.
- [Stripe paid seven billion dollars for an online travel agency](https://thomasjankowski.com/writing/stripe-paid-seven-billion-for-an-online-travel-agency/): an AI model router has the structure of an online travel agency, and travel already showed that the commission was never the business.
- [The numbers came back clean. The grader built the test.](https://thomasjankowski.com/writing/the-grader-built-the-test/): Utah opened the first deployment path for autonomous clinical AI, and the company supervising it is the company that built it.
- [The billing got faster and the bill went up](https://thomasjankowski.com/writing/the-billing-got-faster-and-the-bill-went-up/): AI coding tools make billable diagnoses cheap to find, and a system that pays per unit found buys more units at the same price.

## Pillar hubs

- [AI](https://thomasjankowski.com/writing/pillar/ai/): agents, models and AI infrastructure.
- [Healthcare](https://thomasjankowski.com/writing/pillar/healthcare/): clinical workflow, payers, EHRs, remote monitoring and regulation.
- [Travel](https://thomasjankowski.com/writing/pillar/travel/): airlines, distribution, GDS and NDC, hotels and OTAs.

## Full archive

- [Writing](https://thomasjankowski.com/writing/): every piece, newest first, with a one-line summary.
- [Sitemap](https://thomasjankowski.com/sitemap.xml): the full URL set, including the pillar hubs.
- [RSS](https://thomasjankowski.com/rss.xml): the 50 most recent pieces, full text. Also as [Atom](https://thomasjankowski.com/atom.xml), [JSON Feed](https://thomasjankowski.com/feed.json) and one feed per pillar ([AI](https://thomasjankowski.com/writing/pillar/ai/feed.xml), [Healthcare](https://thomasjankowski.com/writing/pillar/healthcare/feed.xml), [Travel](https://thomasjankowski.com/writing/pillar/travel/feed.xml)).
- [Full text](https://thomasjankowski.com/llms-full.txt): every piece as plain text in one file.

## How to cite

Preferred attribution: "Thomas Jankowski, [piece title], thomasjankowski.com, [year]." Every piece carries author and publication date in Schema.org JSON-LD, inline and at `/schema/<slug>.json`.