AI & Web3 automation operator
Risk-aware AI automation for Web3, finance and operator workflows
I build automation that stays out of your way — until something needs a human's sign-off. Approval flows, Telegram bots, execution systems and content operations, built to cut the noise without cutting you out of the loop.
Four systems in production, written up in full.
What I help with
I help teams turn fragile manual and AI-assisted workflows into controlled systems — with approvals, monitoring and reconciliation.
See how that works as an engagementRisk analysis
Map what can go wrong before automation goes live.
Approval workflows
A person signs off before a trade, a post, or a payment goes out — not after.
Controlled automation
Agents and bots that assist, not run unchecked.
Alerts, data & digests
Monitoring that tells you what changed, not everything that happened.
Finance & admin bottlenecks
The recurring finance and reporting tasks that quietly eat an afternoon a week.
Case studies
Four systems I designed, built or operate. Each one automates the repetitive work and stops short of the decision a person should still be making.
Offers
Three focused engagements. Scoped, practical, no open-ended retainers.
Telegram Bot UX Audit
I go through your bot the way a new user would, and flag everywhere it confuses them. You get a short list of fixes, not a redesign.
Best for
Trading bots, alert bots, admin bots, internal tools.
Approval Flow Design
I design the moment where a person has to say yes before something risky happens — clear enough that they actually will.
Best for
AI agents, Web3 bots, trading workflows, internal automations.
Low-noise Alert System Review
I go through what your system alerts on and cut it down to what actually needs a reaction.
Best for
Trading alerts, onchain monitoring, market research, operational dashboards.
Start free
AI Automation Risk Checklist
The pre-launch questions I work through before an agent or bot is allowed near money, data or a public account. Free, and a decent way to find out whether we'd agree on anything.
About
My background is audit and finance, not software — Big4 audit, accounting and ERP implementation. Years of testing other people's processes and finding the step that fails quietly before I started automating my own.
That's still the instinct behind everything I build: trace the whole process, reconcile it against a source of truth outside your own system, and check what happens when one step fails without saying so. Automation should remove the busywork — but someone should still see what happened and sign off on what matters.