Case studies

Concrete walkthroughs of AI Work Flow for Business deployed against a real workflow. Each case study follows the same structure: the team, the problem, the chosen workflow, what shipped, and what was learned.

Available case studies

Case study	Industry	Workflow	Status
ISP support triage	ISP / telco	Complaint classification + ticket routing	In production
Bank IT helpdesk	Bank	Internal ticket triage + RAG over SOPs	In production
Factory IT operations	Factory	Shift handover summarization + anomaly flagging	In production

Why these three

The three case studies cover the three audiences locked in ROADMAP.md §3:

• ISP support — primary audience, narrowest scope, easiest to evaluate
• Bank IT — regulated environment, proves the data-residency argument
• Factory IT — non-office environment, 24/7 operations, tests the latency argument

What you should expect

A case study is not a sales brochure. It is a written-up account of what worked, what didn't, and what the team would do differently. Where the actual project has a real customer, the case study is anonymized; where it is a design exercise, the case study says so explicitly.

See also

• Demo — the underlying module behavior
• Adoption journey — how a team gets from "interested" to "in production"