About Preclari.
A tool for one job. Built in Basel. Shipped with an open spec.
Preclari is the AI-preflight layer: a tool for one specific question (is this workflow appropriate?) across the EU regulatory wave that touches AI. It produces an audit-defensible record a human reviewer can verify independently of the model that designed the workflow.
The work that became Preclari started as a recurring deliverable inside a Basel-based advisory practice: a structured pre-deployment review of a pharma AI workflow, run before the pilot, defended in front of the regulator. Preclari is the productised version of that deliverable: sharpened, formalised, and shipped with an open specification so the artifacts it produces outlive the tool. v0.1 ships with the pharma GMP corpus; medtech (EU MDR + FDA) and the EU AI Act high-risk regime are next, trigger-gated to engagement signals.
A workflow-first contract. Built for agents, audited by humans. Built in Basel, with a sister practice at practice.pramod.ch.
Photo by Tani Photto For 25 years, Pramod built the compliance and traceability infrastructure that pharma operates on: track-and-trace architectures from manufacturing line to patient, serialisation systems across four continents, and compliance programs supporting $25B+ in pharmaceutical sales across 15+ markets.
He designed the Tatmeen national track-and-trace system for the UAE Ministry of Health from a blank page. He architected Syngenta's global serialisation platform spanning Germany, Vietnam, India, and Brazil, with a live anticounterfeiting detection layer that cut response time from months to near-real-time. At Pfizer, Gilead, AstraZeneca, BASF, and Novo Nordisk, he was the person writing the functional and technical specifications, leading cross-functional teams of 50+, and staying accountable for the outcome.
For 15 years, that work meant designing systems and writing the specifications for other teams to build. Then AI-enabled development changed the equation. Preclari is the first thing he has fully designed and fully built himself.
The thesis behind it is simple: someone who spent 25 years living pharma regulatory compliance understands what breaks when AI enters those workflows. Preclari is the preflight infrastructure for that transition: MCP-native, audit-defensible, built for agents, reviewed by humans.
- ChainAlign Decision & judgment infrastructure
- Open Museum MCP Another shipped MCP server
- pramod.ch Personal site & full story