AI automation consultant. One scoped sprint.
A fixed-scope way to go from “we should be using AI” to a pilot running on your real tools and data, without hiring a team first.
Arrange a free 20-minute scoping call by email. English or Spanish.
1-week diagnostic + 2–4 week pilot
- Launch pricing · from
- US$1,500
- 1-week diagnostic + pilot
Who it's for
- Teams that know AI could save hours every week but don't know where to start.
- Companies with a stalled prototype, or ChatGPT experiments that never reached production.
- Founders who want a senior engineer to scope the work before committing budget.
What you get
- Interviews with the people doing the work today, and a map of the processes worth automating.
- One to three prioritized use cases, each with estimated impact, cost and risks.
- A technical plan: models, integrations, data access, security and monthly running cost.
- In the pilot, the top use case built and connected to your real tools, measured against a success metric agreed up front.
- Handover: documentation, a walkthrough with your team, and the code in your repository.
How it works
1 week
Diagnostic
Interviews, process map and a prioritized roadmap.
2–4 weeks
Pilot
The top use case built on your data and measured against the agreed metric.
Monthly, optional
Operations
Monitoring, improvements and cost control once it's live.
Pricing
- Sprint
- Diagnostic (1 week) + pilot (2–4 weeks), scoped in a free call
- from US$1,500
- Operations
- Optional, month to month · monitoring, improvements and cost control
- Custom
Every engagement is scoped in a free 20-minute call. AI model and API usage is billed at cost, with no markup.
Real work you can inspect
agente-ia.dev
My own platform with free courses and guides for learning AI agent development, built with Next.js.
This is my own product, not a client AI deployment. The sprint applies that development experience to a scoped pilot, not a promise of ROI.
See projects and experience
Questions
What should we bring to the first call?
One repetitive process, the tools your team uses and an example of the data involved. We use the call to define scope, not to request passwords or sensitive customer data.
What if AI isn't the right answer for our problem?
Then the diagnostic says so. A clear “don't build this” is a valid outcome, and it is cheaper to learn it in one week than in three months.
Do we need clean, organized data first?
No. Part of the diagnostic is checking what data you have, where it lives and what is realistic to do with it.
Which AI models do you use?
Whatever fits the job and your constraints: OpenAI, Anthropic Claude, Google Gemini or open models. The accounts and keys are yours.
Who owns the code?
You do. Everything lives in your repository and your cloud accounts.
Tell me what you want to automate.
A 20-minute call is enough to tell whether AI makes sense for your case and what it would take. In English or Spanish.
I reply within 24 hours.
Guatemala (UTC-6): full overlap with US working hours.