You know AI matters for your business. You haven’t figured out where it actually fits.
Readiness audits, workflow builds, AI agents, integration work. For SMBs and ops teams who want AI that fits how they run.
14 years at Amazon and Flexport making big operations actually work. Now I help clients make AI work for them.
Book a 30-min scoping call →What I can build for you
Every engagement moves through three stages — diagnose where AI fits, build it, then operate it. Full scope, timelines, and pricing on Services.
“We have manual processes that should be automated. We don’t know where to start.”
AI Readiness Assessment
In 1 week: a written assessment of your top automation opportunities, ranked by impact and feasibility. One call in, async analysis, one call out. No multi-week discovery sprint.
If you’ve already turned on Claude for Small Business, Copilot, or another pre-built, the assessment tells you what they cover for your case and what needs custom work. If you haven’t, it tells you whether one would fit before you spend on tools.
Next step: a workflow build, AI agent, or back-office automation — quoted at the debrief call.
See details →“We built an MVP with Lovable, Cursor, or Replit. Now it needs to survive production.”
Production Readiness Review
In 1.5–2 weeks: an architecture review of what you actually built. Auth, data handling, scaling, monitoring, security, regulated-shape compliance fit. Multi-stakeholder format: kickoff with founder, tech lead, and ops or security; code walkthrough; written review delivered as a roadmap; 60-min debrief.
Top-5 gaps ranked catastrophic / serious / monitor. What to fix in 30 days, what to queue for 30–90, what to defer with eyes open.
Next step: build the gaps directly, or advisory while your team executes the roadmap.
See details →“One or two workflows are eating my team’s week.”
Workflow Automation Build
In 1–2 weeks: one end-to-end workflow built and deployed on your stack. n8n, Zapier, custom API, whatever you’re running. Failure-mode handling so it doesn’t break silently. Runbook so your team can operate it without me.
If a vendor pre-built handles your case, I’ll tell you and you save the money. Most of the time it doesn’t — the workflow sits between systems they don’t connect, or runs on logic they don’t encode.
Next step: more workflows, an embedded AI agent, or a monitoring retainer to keep what’s running healthy.
See details →“Things break and we don’t find out until a customer tells us.”
Monitoring Retainer
Coverage that tells you a thing broke before your customer does. Health checks across the systems I built you, alerts routed to whoever needs to act, and a monthly tune-up call where we fix the small stuff before it grows.
Everything I build self-monitors from day one. The retainer keeps it that way as your stack shifts underneath. $300–500/mo, depending on how much you’re running.
Next step: expand coverage as you build more, or a full reliability sweep across the stack you’ve already got.
See details →“We’re always building. One-off projects don’t fit how we actually work.”
Implementation Retainer
An embedded AI operator running your build backlog week to week. Automations, agents, integrations — whatever’s next in the queue gets scoped, built, and shipped. No per-project contracting, no re-onboarding every time.
Three tiers: $5,000/mo for one active build per quarter. $8,000/mo for multiple concurrent builds. $12,000–$15,000/mo for weekly cadence and enterprise-scale iteration.
Next step: a scoping call to map your backlog and pick the tier. Most retainers start from a completed workflow or agent build.
See details →How I work
Scope and requirements before build begins. Work ships tested, with documentation your team can use. Regular check-ins throughout. Four frameworks I bring into client work: