How we work
Short phases, clear scope, and no black boxes
Every AI project carries uncertainty. The sensible way to handle it is to move in small phases, measure each one, and decide on the numbers whether to keep going.
The phases
Four steps, and you can stop after any of them
First conversation
You walk us through the problem, the process and the systems involved. You leave with an honest view of whether AI helps in your case and where we would start.
You get
- A concrete recommendation for the next step, in writing
Diagnostic
We talk to the people who do the work, review the data and systems, and measure how the process runs today. It is the foundation for everything else.
Deliverables
- A map of the current process and what it costs in hours
- Use cases ranked by impact and effort
- Estimated return and running costs
- A pilot proposal with defined scope and a quote
Pilot
We build the first use case on your real systems and data, with a small group of users, and measure it against where you started.
Deliverables
- A working solution in a controlled environment
- Measurement against the baseline: hours, errors, turnaround
- A go or no-go recommendation for production, backed by numbers
Production and support
We roll the solution out across the operation, train users and owners, and monitor it. When your team is ready, we hand it over completely.
Deliverables
- The solution in production, documented
- Training for users and owners
- Monthly support, monitoring and improvements, if you want them
Principles
What we do not compromise on
One problem, one metric
Before building, we agree on which number should move and by how much. Without that, we do not start.
Your data, your accounts
The solution lives on your infrastructure or in accounts in your company’s name. We sign an NDA from the first meeting if you need one.
Everything documented
Code, decisions and configuration are written down and yours.
No long contracts
Each phase is contracted separately. If the result does not convince you, you are not bound to continue.
With your people, not around them
The people doing the work today take part in the design. That is why they use it afterwards.
Technical honesty
If a well built spreadsheet solves it better than AI, that is what we will recommend.
Ways to engage
Depending on where your company is
Diagnostic
Defined scope. The starting point for most companies.
Project
Quoted per phase, with deliverables defined before we start.
Monthly retainer
Hours reserved every month for support, improvements and new use cases.
Workshops
Per session or per program, depending on group size.
Frequently asked questions
Before the first conversation
How much does it cost?
It depends on scope, which is why we do not publish a single rate. Every phase is quoted before it starts, with its scope and deliverables in writing.
What size of company do you work with?
Mostly mid-sized companies and departments within larger ones. More than size, what matters is having a process with enough volume to make automating it worthwhile.
Do you work outside Costa Rica?
Yes. We work remotely with companies across Latin America and the United States, in English or Spanish. For on-site workshops outside Costa Rica we arrange travel.
What about the confidentiality of our information?
We sign an NDA before seeing any data. We use AI services configured not to train on your information, and the solution lives in accounts in your company’s name.
Does our data need to be organized before we start?
No. The diagnostic tells us what state it is in. If it needs work, that becomes part of the plan.
Do we need an IT team?
It helps, but it is not required. We coordinate with your IT team if you have one, or with your systems vendor if you do not.
What happens when the project ends?
You get the code, the documentation and the training. Your team can run it, or we can stay on with a monthly plan.
What if AI turns out not to work for us?
That is what the diagnostic and pilot are for: they are the cheapest way to find out. If the numbers do not justify going further, we tell you and you stop spending.
The first step is 30 minutes
No commitment. Leave the call knowing whether AI makes sense for your problem.