Why I Still Run a Services Company in the Age of AI Products

Of everything in my portfolio, the venture that raises eyebrows isn't the AI products or the book — it's the services company. IOanyT Innovations: software consulting, cloud architecture, DevOps, platform engineering. Ten years old, 150+ client engagements, AWS Advanced Partner. "Services don't scale," the playbook says, and the playbook isn't wrong. But everyone repeating it is missing what a services company is for in a portfolio like mine. IOanyT is the engine. Every product bet I make runs on fuel this unfashionable machine produces. This essay is my defense of the most boring venture I own — and why I'd build it again first.

The Unfashionable Venture

The startup canon treats services as what you do when you can't build product: linear revenue, headcount-bound margins, no exit multiple. All true, and all beside the point. The canon is written for founders optimizing for a fund's return profile. I'm optimizing for something else — durable independence. A services business converts skill into cash now, at healthy margins, with customers who pay invoices rather than burn runway. It is the one business model where being good at the work is, by itself, the whole go-to-market. That's not a weakness to escape. It's a foundation to build on.

What 150+ Engagements Actually Buy

The obvious answer is revenue. The real answer is the pattern library. A hundred and fifty engagements across North America, Europe, the Middle East, and India — SaaS platforms, financial systems, AI products, cloud migrations — is a decade-long guided tour of how technology actually fails inside real organizations. Not how it fails in blog posts: how it fails on a Tuesday, in production, with a client on the phone. Every thesis I hold, including the dependency layer itself, was extracted from that pattern library. When I claim AI agents break on trust, context, and accountability rather than on intelligence — that's not analysis. That's inventory.

The Engine Model

Here is the portfolio's circulatory system, stated plainly: IOanyT generates cash flow; the cash flow funds the product ventures; the product ventures test the thesis; the thesis makes the services sharper. No venture capital appears anywhere in that loop. That's a choice with a price — I grow slower than a funded competitor and I feel every experiment in my own P&L — and with a payoff: nobody can tell me to kill a venture like The Heartbeat because it doesn't fit a fund's timeline. The unit economics of the whole portfolio only work because one venture in it was never meant to be a rocket. It was meant to be a generator.

How AI Changed the Services Business From Inside

The standard prediction is that AI kills services companies. From inside one, I watch it do something more interesting: it kills the pyramid. The classic services model billed armies of junior engineers doing repeatable work under a few seniors. That layer is evaporating — the repeatable work now belongs to agents. What's left is what was always scarce: senior judgment about what to build, architecture that survives contact with production, and accountability when things break. A services firm built on senior expertise plus agent leverage gets more valuable per person as AI improves, not less. The firms in trouble are the ones whose margin was the pyramid. Ours never was — at peak we ran 50+ engineers, and the direction of travel since has been fewer people, more leverage, deeper work.

Ten Years of Compounding Trust

Services compound in a currency the product world underrates: verifiable trust. AWS Advanced Partner status is AWS examining the work, not us describing it. ISO 9001 and ISO/IEC 27001 are external auditors inside our quality and security systems. Upwork's Expert-Vetted tier — top 1%, with a 100% job success score — is a public ledger of a decade of clients getting what they paid for, a global electronics manufacturer among them. None of this is glamorous. All of it is the exact commodity my whole thesis says becomes precious as intelligence gets cheap: proof that an accountable party did what it said. The services company doesn't just fund the dependency-layer bets. It is one.

When to Hire Us — and When Not To

Hire IOanyT when the work is load-bearing: a platform that must hold real traffic, a cloud architecture that must survive an audit, a codebase that needs rescuing from its own velocity, AI features that have to work past the demo. You get senior engineers with a decade of pattern library and the certifications to prove the process behind them. Don't hire us to be cheap hands — the pyramid firms still sell that, for now. And if you're a founder wondering whether to build your own engine before chasing your own rocket: the book makes the long version of the argument this essay makes in miniature.