About
I'm Ilia Cherepanov. I build OpenAI Workspace Agents for US small and mid-market companies on ChatGPT Business and Enterprise.
My background is go-to-market operations at an AI-native startup, which means I've spent the last few years shipping AI features into production where the buyer is skeptical, the data is messy, and the failure modes are real. Workspace Agents are the cleanest on-ramp yet for team-scoped automation, which is why I moved fast when OpenAI launched them on April 22, 2026.
I work solo, on one client engagement at a time. This is deliberate. Most agent-consulting firms staff an account manager, a delivery lead, and one or two juniors — which means the person talking to you isn't the person building the agent. With me, the person scoping the spec is the person writing the prompts is the person on the Slack thread when something breaks.
I'm opinionated about what makes a good agent and what doesn't. I'll turn down work where I don't think the workflow is a fit — over-scoped agents that try to be everything, vague use cases that couldn't be spec'd cleanly, or requests where the underlying problem is org design rather than automation. Turning down bad work is the clearest signal I can give that I'll deliver on the work I take.
I'm US-timezone-friendly. I answer my own email. I ship on the dates I quote.
How I work
- Fixed-price builds. $1,000 per agent, one-time, with a written spec and revisions. You know what you're paying before I start.
- Real data from day one. I tune prompts against your actual inputs, not a happy-path sample. Most agent failures come from testing with 3 examples and shipping; I test with 30–50.
- Handoff, not lock-in. You get a written runbook, a Loom walkthrough, and a named owner on your team. I walk away with no access by default. If you want me around on retainer, great; if not, you're not stuck.
- Opinionated but not dogmatic. I'll push back on scopes I think won't work. I'll also listen when you know your team better than I do. If we disagree and I'm wrong, I adjust.
Not the right fit if
- You want a big-agency engagement with a kickoff deck and a quarterly-review deck.
- You need on-prem deployment or air-gapped compliance — not what Workspace Agents are built for.
- Your goal is to replace headcount with AI. Agents multiply good teams; they don't rescue broken ones.
- You need custom LLM fine-tuning or a dedicated model — that's a different consultant.
Want to talk?
20-min intro call. I'll tell you whether your workflow is a fit, what it would take, and what I'd do instead if it isn't.