02Service
AI agents & workflow automation
Agents that read what comes in, decide what to do, and do it, with a person in the loop where it matters.
The short answer
Updated
An AI agent is software that reads what comes in, like an email, a ticket or a form, decides what to do next, and does it through tools you control. I build agents for one real workflow at a time, with approvals and logs, usually in 4 to 8 weeks for $8k to $30k.
- Typical budget
- $8k to $30k
- Timeline
- 4 to 8 weeks to production
- How it's quoted
- You get a fixed quote after the first call. Model and hosting costs are separate.
Agents that do the boring part of the job, reliably.
An agent is just software that reads a situation, picks a tool and uses it. The hard part isn't the model. It's choosing which steps to hand over, giving the agent tools that can't do damage, and knowing when it got something wrong.
I start with one workflow your team repeats every day: support triage, lead research, document intake, follow-ups. We automate that end to end, measure it against how it's done today, then widen the scope once it's earned trust.
A good fit if
- Your team copies information between tools all day
- You've tried a chatbot and it answered questions but didn't do anything
- Your AI pilot impressed everyone, then stalled before production
- You need audit logs of what the AI did and why
What you get
- 01
A working agent in production
Running on your infrastructure or a managed cloud, connected to your real systems, not a sandbox copy.
- 02
Typed tools with limits
Every action the agent can take is a function with checked inputs and permissions. It can't email a customer unless that tool exists and allows it.
- 03
Human review where it counts
Approval steps for anything expensive or customer-facing, so people stay in control while trust builds.
- 04
Logs and an eval set
A record of every decision, plus a set of real cases we re-run before each change so improvements don't break what already worked.
How a project runs
Week 1
Discovery
I sit with the people who do the work today. We pick one workflow and agree on what success looks like in numbers.
Weeks 2 to 4
First version
A narrow agent that handles the common cases, running alongside your team with every action reviewed.
Weeks 5 to 8
Hand over the wheel
Review steps are removed where the logs show it's safe. Edge cases get their own tools or routes to a person.
Ongoing
Stay on
Monitoring, model upgrades and new workflows on a monthly retainer, if you want it.
- OpenAI
- Claude
- LangGraph
- MCP
- Tool calling
- Python
- Node.js
- PostgreSQL
- Queues & schedulers
Related work
- File 002AI agentsIn beta
Altrun
An AI assistant that turns incoming signals into scheduled actions- Role
- Founder, engineer
- Year
- 2026
- Stack
- LLM tool calling, Agentic workflows, Telegram Bot API
- File 003AI platformLive
Name withheld under NDA
Name withheld under NDA- Role
- Lead engineer
- Year
- 2025
- Stack
- LLM agents, TypeScript, Next.js
- File 005Automation platformLive
Name withheld under NDA
Name withheld under NDA- Role
- Lead engineer
- Year
- 2024
- Stack
- Workflow engine, Laravel, Queues
Questions people ask
Which model do you use?
Whichever does the job at the right cost. I've shipped on OpenAI and Claude, and I test Gemini and open-weight models for each task. The code is written so the model can be swapped without a rewrite.
Will it work with our existing tools?
Usually, yes. If a tool has an API, a database or even a reliable inbox, the agent can work with it. I check this in week one before you commit to a build.
We're in a different timezone. Is that a problem?
No. I keep flexible hours and shift my day to overlap with yours, so standups, reviews and launches happen on your clock.
What happens when the agent gets it wrong?
It will, sometimes. That's why every action is logged, risky actions need approval, and we keep a set of past mistakes that every new version has to pass.
Do you work with in-house teams?
Most of the time. I work as a forward deployed engineer: embedded with your team, in your Slack and your repo, and I leave behind code your engineers can maintain.
How much does an AI agent cost to build?
Most of my agent projects land between $8k to $30k for the first workflow in production. The low end is one workflow with two or three tools. The high end touches several systems and comes with approvals and an eval suite. Model usage is extra, and for most teams it's a small monthly bill.
How long does it take?
4 to 8 weeks to production. Discovery takes the first week, then the agent runs alongside your team with every action reviewed, and the reviews come off as the logs earn it.
What do you need from us to start?
One workflow you want automated, around twenty real examples of it, access to the tools it touches, and one person who knows the job and can answer questions.
Further reading
Contact
Tell me what you're trying to fix.
I read every message myself and usually reply the same day. A rough idea is enough to start.
Or send a short briefHours
Your hours work for me.
San Francisco
--:--
New York
--:--
London
--:--
Berlin
--:--
Dubai
--:--
Singapore
--:--
Sydney
--:--
Surat
--:--
Me
I keep flexible hours and move my day to overlap with yours, whether that's a morning standup in New York or a late release in Sydney.