AI and automation

Who it is for

Businesses where operations are the product: a slow dispatch or a missed booking is lost revenue, not a bug report. If your team repeats the same work every day, or a decision depends on data nobody has time to read, a system can take that load. The condition is that it runs in production, not only in a demo.

What it includes

Systems that do real work in production:

• Agents and chatbots. Like the WhatsApp booking bot I have shipped for wellness and service businesses, part of the software that runs the front desk of real locations.
• Computer vision. Like CAMTRACK, an in-store traffic analytics pipeline for a wellness brand: people detection and tracking across three locations, running on real cameras with YOLO, ByteTrack and Python.
• Process automation. Like the 50+ business processes I automated at an assistance group, alongside a cloud migration for 200+ users.

How it works

Each project has a defined scope, a start date and a single point of contact: me. I start from the real problem and build the smallest solution that holds: the least machinery that survives contact with users, with no layers nobody will maintain.

What I build runs like any other part of a production system:

• Nothing ships without tests and a way back.
• Serverless and infrastructure as code, so you pay for what runs and the environment can be rebuilt.
• Agents do the work; the judgment and the responsibility stay with a person.

Every technical decision has a cost and a return. An automation is worth building when it pays for itself in time or revenue, not because the technology is new.

What you have at the end

A system running in production, not a demo, that your team understands and can operate. Every engagement is designed to end, so your team can lead the technology and knows exactly what to ask for when it does need help.