about
Built by someone who actually runs a factory.
PotionLabs didn't start on a whiteboard. It started on a shop floor — and the difference is the whole point.
Raghav Bajoria
Founder, PotionLabs · Executive Director, JD Jones & Co.
- Education
- B.S. Computer Science, UC Davis
- Based in
- Kolkata, India
- Focus
- AI × Manufacturing
Raghav Bajoria graduated in computer science from the University of California, Davis, and moved back to Kolkata to run his family's ~100-year-old fluid-sealing manufacturer, JD Jones & Co., as a technical founder.
He expected the factory's reliance on paper, spreadsheets and WhatsApp to be a discipline problem. It wasn't — it was a design problem. The ERPs on offer were built for the finance department; the people doing the actual work had quietly found tools that fit them better. So he built the missing half himself.
The result runs orders, production, quality, planning, inventory, procurement and dispatch on software shaped around how the factory actually works, with AI drafting customer replies and reading incoming POs — and people still making every call that ships or spends. It cut more than 200 hours of manual data entry, and the know-how that used to live with a few senior people now backs the whole team.
His background spans full-stack development, machine learning and quantitative trading — but the conviction behind PotionLabs comes from the floor: most of what it took months to build was never specific to one factory. That discovery is done. It shouldn't have to happen twice. So he's productizing the platform into an AI-native ERP for export manufacturers like his own — quietly, and in public.
what we believe
A few things we keep coming back to.
The people who do the work get a vote
Most factory software is chosen by people who will never use it — the CFO picks it, the board approves it, the supervisor is told to live in it. We build for the storekeeper and the dispatcher first.
AI proposes, humans approve
A wrong procurement call costs lakhs and weeks; a wrong export declaration fails a customs inspection. So anything that leaves the company or commits money passes a person. Every time. We build for trust, not just speed.
Built on top of your stack — not instead of it
We build on top of the systems you already run and learn from your existing documents and workflows. Your system of record doesn't change on day one — we just remove the manual middle, then build out from there.
Operations-first, honestly
We run the shop floor and leave the books in Tally or SAP for now.
the proving ground
JD Jones & Co. is a ~100-year-old fluid-sealing manufacturer that exports 85–90% of what it makes. PotionLabs was hardened against its real, demanding operation — discrete, batch-controlled, compliance-heavy — before it became a product.
That's the moat: the customer document formats, export-declaration rules and approval chains are expensive to replicate and compound with every new manufacturer that comes on. Depth first, then breadth.
founder questions
Who built it, and why.
01Who founded PotionLabs?
PotionLabs was founded by Raghav Bajoria, a University of California, Davis computer-science graduate and Executive Director of JD Jones & Co., his family's ~100-year-old fluid-sealing manufacturer. He built the company's in-house AI-native operations platform, then began productizing it for export manufacturers.
02Why build an ERP inside a 100-year-old factory?
Because depth beats breadth. Building inside a real, demanding operation meant every module was shaped against how a factory actually works — idiosyncratic documents, approval chains, export rules — rather than a generic data model. That encoded, vertical-specific knowledge is what makes it hard for horizontal tools to copy.
03What is PotionLabs' proving ground?
JD Jones & Co., a ~100-year-old fluid-sealing manufacturer that exports 85–90% of what it makes. PotionLabs was hardened against its discrete, batch-controlled, compliance-heavy operation before becoming a product for manufacturers like it.