Reorder points & safety stock
Where each SKU should trigger, and the buffer that absorbs a bad week.
InventoryIQ Engine™
The inventory expertise behind every answer.
Filo isn’t a general-purpose chatbot with inventory bolted on. Under every answer is the InventoryIQ Engine™ — a reasoning system trained on how inventory actually works, grounded in your live data, and built to show its work.
// Trained on supply-chain science. Tuned to your operation.
IQ
InventoryIQ Engine™
● Reasoning · SKU-WALNUT-08
Signals in
On-hand14 units
90-day burn rate2.3 / day
Days of cover~6 days
Supplier lead time4d (p90: 6d)
Safety stock12 units
Seasonality (Q3)+18%
Recommendation
Reorder 120 units by Mon, Jun 22 — covers the lead-time gap plus safety stock through the Q3 lift, rounded to a full case.
01 / Why it’s different
A general model can describe inventory. InventoryIQ can run it.
Ask a generic assistant about reorder points and it recites a definition. InventoryIQ applies that definition to your SKUs, with your numbers, and tells you what to do next — the way a seasoned inventory planner would.
02 / Trained on real inventory science
The expertise baked in.
These aren’t features you configure — they’re concepts the engine reasons with by default, the moment you connect a source.
Lead-time variability
Not the promised date — the lead time your suppliers actually deliver on.
Economic order quantity
Order sizes that balance ordering cost against the cost of holding stock.
Demand forecasting
Projects forward from real sales velocity, not last month’s average.
Seasonality & trend
Separates the recurring lift from the genuine change in demand.
ABC / XYZ classification
Knows which SKUs deserve attention and which can run on autopilot.
Days of cover & burn rate
Translates a raw count into the only unit that matters: time.
Dead & slow-moving stock
Flags the cash sitting still on your shelves before it ages out.
MOQs & case-pack rounding
Recommends quantities you can actually order, not theoretical ones.
Service levels & stockout risk
Tunes buffers to the fill rate you want to hold, by SKU.
Carrying & holding cost
Weighs the real cost of over-ordering against the risk of running out.
Multi-source reconciliation
Resolves the same SKU across spreadsheets, Shopify, and counts into one truth.
03 / How it reasons
Every recommendation is computed, not guessed.
STEP 01
Ground
It pulls the real numbers — on-hand, in-transit, velocity, lead times — from your normalized sources. No hallucinated stock.
STEP 02
Compute
It applies inventory science to those numbers: cover, safety stock, order quantity, seasonality — the planner’s math, every time.
STEP 03
Explain
It shows the signals and assumptions behind the call, so you can sanity-check the logic instead of trusting a black box.
STEP 04
Defer
It recommends — you decide. Adjust a number or override the call and the engine recomputes on the spot.
04 / The difference
Generic AI vs. the InventoryIQ Engine™.
A general assistant
Knows about inventory.
- Recites textbook definitions
- Answers from whatever you paste in
- Can’t see your live counts or velocity
- Gives advice you still have to do the math on
InventoryIQ Engine™
Runs your inventory.
- Applies the science to your actual SKUs
- Computes against live, normalized data
- Knows your real burn rates and lead times
- Hands you the quantity, the date, and the why
05 / Tuned to your operation
Expert by training. Specialist by experience.
The science is universal; your operation isn’t. InventoryIQ learns the parts that are specific to you.
Your real velocity
It learns how fast each SKU actually moves — including the slow build and the sudden spike — instead of a flat monthly average.
per-SKU burn · trend · variance
Your suppliers’ behavior
It tracks the lead times your suppliers actually deliver on, not the ones they promise — and buffers for the ones that slip.
observed lead time · reliability · p90
Your seasonality
It recognizes your recurring peaks and lulls and pulls reorders forward before the rush, not after you’ve already sold out.
seasonal index · event lift · recovery
06 / Built to be trusted
Expertise you can inspect.
Sophistication is worthless if you can’t trust it. Every InventoryIQ recommendation exposes the numbers and assumptions underneath — and never moves your stock on its own.
- SHOWS WORKNo black box. Every recommendation lists the signals and math that produced it.
- GROUNDEDLive data only. The engine reasons over your real counts — it doesn’t invent stock to fill a gap.
- OVERRIDEYou set the inputs. Adjust a lead time or service level and every downstream number recomputes.
- PRIVATEYour data is yours. Your operational data is never used to train shared models.
Put inventory expertise to work.
// Connect a source · The InventoryIQ Engine™ starts reasoning in minutes · No card