Case study · Zecat
Rolling AI out across a 200-person company
Zecat, one of the region’s larger promotional-products groups, wanted AI to actually change how the company works, not sit in a corner as a demo. We made that adoption safe, governed and real.
Industry
Promotional products
Company
Regional leader, ~200 people
Engagement
AI adoption + automation, ongoing
Stack
n8n, Microsoft 365, Python, AWS
The challenge
Almost everyone at Zecat had started using AI tools on their own, including people in finance and administration who had never written code. That energy is exactly what you want, and also a risk: no roles, no secure logins, no review on work that touched finance reports, and a back office still running on long, manual spreadsheets.
Our approach
We did not try to be the bottleneck that builds every automation. The people who know the business rules should build their own, because they iterate faster and debug better. So our job moved to the center: make it safe and easy for each area to build, instead of doing it all for them.
What we built
We automated the loudest pain first, finance and administration: long manual Excel processes that now run on a schedule. The teams train the system themselves by editing a plain rules sheet, no deploy and no engineer in the loop. Underneath, we built the shared, governed foundation: secure logins and roles, MCP servers that expose company systems to AI safely, API integrations and repo templates so any area can ship something connected to their stack without wiring auth and secrets from scratch.
What’s next
The same playbook is moving into marketing and operations, area by area. Next up is a partner-facing WhatsApp assistant: Zecat sells through partners, so it would help them brainstorm product ideas, drop their logo on an item and preview how it looks, straight from a chat.
What changed
- Long manual Excel processes (tax cross-checks, reconciliations, statement categorization) now run on a schedule instead of by hand.
- Finance and admin own and train their own automations through an editable rules sheet, no engineer required.
- A secure, governed foundation (auth, roles, MCP servers, repo templates) so every area can build safely.
- Company-wide AI adoption underway, expanding team by team on infrastructure Zecat owns.
