Case studies

What got automated, and what changed after

Four parts each: the situation, what was built, the number before and the number after. No pretty screenshots.

CEMEX

2022–2024 · via NEORIS, Tech Lead

The challenge: A monolithic platform was slowing down feature releases, and support answered everything by hand.

What we did: Migration to microservices, an AI agent embedded in the support flow, and optimization of SQL Server, Kafka, Snowflake, Azure Functions and Redis (CQRS).

Result: A platform that ships faster, with AI-assisted support built into the service flow.

SMB version: In an SMB this is the knowledge copilot: your manuals and price lists answering on their own, without the 40-engineer team.

MicroservicesAI / CopilotAzureKafkaSnowflake

Ternium · CEE Nuevo León

2009–2011

The challenge: Production and distribution decisions relied on manual, country-by-country estimates.

What we did: Designed and deployed a sales forecasting platform across several South American countries.

Result: One forecasting system replacing scattered manual processes, running in production.

SMB version: The SMB version fits on one page: the same numbers you estimate by hand today, computed on their own every Monday.

ForecastingMulti-countryProduction

Why this page has so few cases

The two you see are CEMEX and Ternium: big projects, big teams. I reframe them at SMB scale because the mechanism is the same even when the budget isn't — and I'd rather tell you that than pretend they were 40-person companies.

The 2026 SMB cases are in flight. They get published when the client approves the number, not before. A page full of results with no name and no verifiable figure proves nothing, and you've already seen plenty of those.

Yours starts with a measurement

Five days, $2,900 USD. If I don't find at least 40 automatable hours a month, you get your money back.

Book your 30-minute call

Or first work out what those hours cost you →