An AI support agent handling 68% of tickets
We built a retrieval-augmented AI support agent for a US e-commerce platform that now resolves 68% of inbound tickets without human involvement.
68%We built a multi-tenant compliance workflow platform for a German regulatory startup, taking it from spreadsheet-based manual review to a product serving 40+ enterprise customers.

A compliance consultancy was running client audits through a maze of shared spreadsheets and email threads. Each audit took two weeks of manual work, errors were common, and the model could not scale past the founders' own capacity. They had deep regulatory expertise but no engineering team.
We sat through three live audits before designing anything, documenting every decision point and exception the spreadsheets encoded implicitly.
Regulatory frameworks became versioned data rather than hardcoded logic, so new frameworks could be added without a release.
Row-level isolation and per-tenant audit trails were part of the initial schema, avoiding the painful retrofit most early SaaS products face.
The deliverable clients actually pay for — a formatted audit report — is generated from platform data instead of assembled by hand.
The platform became the company's entire commercial offering. They raised a seed round on the back of it, and we continue as their engineering function.
We came in with regulatory knowledge and a folder of spreadsheets. They turned it into a product we could actually sell.
Have a similar problem?
Book a free 30-minute consultation and we'll tell you how we'd approach it.
Book a Free ConsultationBook a free 30-minute consultation with our team — no sales pitch, just a straight answer on scope, timeline, and approach.
info@devluma.io
Response within 1 business day · UK / US / EU hours covered