AI consulting guide
How to implement AI in your companysteps, timelines and data rules
To implement AI in your company, start with a process that already costs time or money and whose output you can check; the tool comes later. A 2 to 3 week assessment shows where AI pays off, and the first automation reaches production in 30 to 60 days, depending on scope, with a measurable goal, a data rule you approve and one person accountable for the workflow.
The stages, with timelines and exit points
This is the path we follow on AI projects, with the timeline we publish for each stage.
- 1
Assessment call: 30 minutes
Free, over WhatsApp, email or a call. You describe the process and the budget, and we tell you whether AI is the right fix, whether a simple script is, or whether it is not worth solving now.
Exit point: no cost
- 2
Opportunity assessment: 2 to 3 weeks
A map of your processes with the AI opportunities ranked by return and effort. You know where to invest and where not to.
You keep the map
- 3
First automation in production: 30 to 60 days
One process, in your environment and integrated with your systems, with cost and quality metrics from day one. The timeline depends on scope, which we lock with acceptance criteria and price before the first line of code.
- 4
Scale in phases
Every new phase has a measurable goal, and each phase pays for the next. Hands-on workshops leave your team running and evolving the stack without depending on us.
The full offer and how we lock scope and price: AI Consulting · How we work
Data rules before code
AI touches customer, employee and business data. These rules go into the design on day one.
LGPD in the design
Access control and an audit trail are part of the project from day one.
Data in your environment
Data stays in your environment or with providers you approve.
Least access
Each automation reaches only the data it needs.
No real data in tests
No real data goes into testing without the anonymization rule you approve.
Where AI goes in and where it stays out
AI goes where its output can be checked, with one person accountable for each workflow.
Sometimes the answer is a deterministic script, and we will say so. In Care Agent, Experity's LLM agent for patients, scheduling lives in code, outside the model.
Every phase has a measurable goal. In our tech lead's work at Nike, AutoCrop, built with GenAI, cut designers' prep time by 44%, from 6.4 h to 3.6 h.
Frequently asked questions
Where do I start implementing AI in my company?
With a process that already costs time or money and whose output you can check. The assessment starts from your processes, and the technology comes after.
How long does it take to implement AI?
The assessment takes 2 to 3 weeks. The first automation in production usually ships in 30 to 60 days, depending on scope.
My company is not a tech company. Can we implement AI?
Yes, and that is the most common case. Our tech lead has run this assessment in healthcare, accounting, logistics and restaurants.
What does it cost to start?
The 30-minute assessment call is free. The price of the rest depends on scope and is locked before the first line of code.
Start with the 30-minute call
Free. Tell us the process you want to fix, and we will tell you whether AI is the way.