Problem
NPLabs was running on spreadsheets, phone calls, and fragmented coordination. They needed a digital workflow that could support regulated pharmacy operations without piling up compliance debt along the way.
Case study
NPLabs: from fragmented operations to a live regulated platform.
NPLabs is a compounding pharmacy in Athens. This is the clearest public example of how I work: take a messy regulated workflow, make the hard calls, and ship a real system without pretending regulation disappears.
This is the main public case study behind Arvanu. I would rather show one real project properly than pad the page with broad claims nobody can verify.
Snapshot
Timeline
87 days to production
Context
EU compounding pharmacy in Athens
Catalog complexity
1,000+ medication variants
Operating model
Five user roles across the workflow
What the work involved
Constraints
Multiple user roles. Sensitive clinical data. Payments wired into the flow. And the kind of traceability a regulated environment demands before anything goes live.
What I worked on
Architecture, workflow design, platform decisions, and AI-enabled features. The goal was a production system the team could actually use — not another disconnected software project.
Why this matters
This is the shape of work Arvanu exists for: practical AI inside a real operating workflow, with governance pressure and delivery pressure in the room from the start.
Platform highlights
Role-based portals for patients, clinicians, pharmacists, operations, and admin
Workflow and data design built around GDPR and pharmacy regulations from day one
Online payments wired into the operating flow, not bolted on after
AI-assisted treatment recommendation and workflow support embedded in the platform
Operational dashboards the team actually uses every day
What this proves
The first useful AI project in a regulated company is almost always a workflow project, not a chatbot project.
Governance has to be designed into the data model and process. It cannot be patched in after launch.
A solo operator can be credible when the scope is tight, the decisions are visible, and the proof is something a buyer can inspect.
Best use of this case study
Use it to judge how I think, not to pretend I am a giant firm.
The point of the NPLabs story is not “look how many logos we have.” It is that you can see the shape of the work: workflow design, production pressure, AI features inside a real business process, and operational judgment under regulation.