Haris Rahi

Everything I know about AI,
I learned by building.

I design AI solutions, ship them, and teach the people who use them to build their own. Three years across multiple enterprise domains. Hands dirty, every day.

300+
Practitioners Trained
120+
Enterprise Clients

The domain changes.
The approach doesn't.

Whatever the function, procurement, operations, legal, sales; the pattern is the same. I take a team's actual work, build AI solutions on it alongside them, and teach them to keep building after I'm gone. Not demos. Not decks. Working tools on real workflows, owned by the people who use them.

Principles earned through building,
not borrowed from theory.

AI must extend your thinking, not replace it. The effort to ensure that is what actually makes you sharper.
Earn your way to agents.

Most organizations skip straight to agentic without doing the reps. Augmentation teaches you how AI actually behaves. Automation teaches you to write tight instructions. Skip steps and you're automating broken processes faster.

Build in context.

AI fluency only develops when the learning and the doing are the same activity. Separate them and the training evaporates.

Retain agency.

Reviewing AI's reasoning feels like thinking. It isn't. Once you see its output, you're anchored to its frame. Think first, then prompt.

No one is an AI expert.

This technology arrived for everyone simultaneously. The gap isn't expertise; it's who's been building and who's been watching.

Haris Rahi

The proof point.

S2C IQ was the first purpose-built AI solution for source-to-contract at Accenture. I designed it, built it, and trained 300+ practitioners across 120+ enterprise clients to use it within a year. Day 2 usage. 30–79% task-level speed gains. Behavioral change that sustained. Cited by Everest Group in Accenture's #1 Leader and Star Performer ranking for Procurement Outsourcing, 2025.

Plot Twist

No engineering degree. No data science background. I built all of this through relentless curiosity, hands-on experimentation, and the same methodology I teach others. That's the point. AI fluency doesn't require technical pedigree. It requires methodology, real-world application, and reps.

I write about what works, what doesn't,
and what I'm still figuring out.

If this resonates, I'd like to hear from you.