Built by operators, for operators.
We're a small team that's spent too many nights in inventory spreadsheets. Tropix is the tool we wished existed.
The short story
Tropix came out of a simple, frustrating observation: the math behind inventory optimization is well-understood — reorder point, safety stock, EOQ — but most operators are running it in spreadsheets, copying numbers between tabs, asking “which SKUs need attention this week?” and not getting a confident answer.
We started building Tropix in 2025 because the existing answers were either too expensive (six-figure SaaS for warehouses with $50M+ inventory) or too crude (Excel macros that nobody trusts). We wanted something a 12-person ops team could turn on in a morning, run on real data, and ship actions from the same afternoon.
For teams that want a hand, we also offer Consulting + SaaS engagements — we sit alongside your team for the first 90 days, set up data feeds, walk the first runs, and build the executive cadence — then you keep the software. It's not required. The product is built to run solo. But it's there if you want it.
What we believe
The future of inventory tools is not bigger dashboards. It's prioritized lists. Not “here's a chart of your months-of-supply distribution” — that's nice. The answer is: “here are the 12 SKUs to handle this week, in order, with the recommended action.”
We also believe AI belongs in the explanation, not the model. You don't need a transformer to tell you that a SKU with eighteen months of cover is a problem. You need clear statistics + clear language. The AI insights in PRO are there to translate the math into a paragraph your CFO will read — not to be the math itself.
How we work
The three principles that decide what we ship and what we don't.
Operator-first
Every feature was built because a real ops director said the same thing in two consecutive conversations. We design for the spreadsheet you'd open at 2am, not the demo we'd give a VP.
Explainable, not magical
Classical statistics that read like a textbook, not opaque ML weights. Every recommendation includes the reason. You should be able to defend any number to your board.
Ship what's useful, skip what isn't
We'd rather have eight modules that operators use daily than eighteen that look good on a feature comparison. No vanity work.