The retail AI platform for merchandising workflows.

Get straight answers out of your own numbers, then turn them into repeatable workflows.

Book a demoWatch a workflow run
Monday trade pack
RUN 14
Reconcile sales and stock
Flag risk
Draft reorder
Propose markdown
Approval gateAxel
Reads the exports you already have
  • Orders feed
  • Product feeds
  • CSV
  • XML
  • Excel

One place to run the merchandising decisions that span your season. Plan, buy, allocate, trade and mark down across the systems you already use.

THE SEASON
  1. Plan
  2. Buy
  3. Allocate
  4. Trade
  5. Mark down

Straight answers from your own numbers.Ask in plain language.The copilot reads the same governed queries your screens use,so every figure traces back to its source.

Why is WK-401 flagged?
Copilot
WK-401 sold out in week 9 of 26, at 98% full price.
For the 17 weeks after, the demand was there and the stock was not. Its rate of sale over clean weeks puts lost sales at about 180 units.
Built on the canvas
Units sold and lost sales by week
Which sizes went first
Against SS25
WK-401 - units sold and lost sales by week

It sees what you see.

The view, its filters and the focused row travel with the question. Nothing needs restating.

Why is this one flagged?
View-context, sent as data
ViewSeason Hindsight
SeasonAW25
Focused rowWK-401Merino Crew Sweater

It works within your permissions.

It reads what your role can read. A change runs as a command, behind an approval gate.

Query catalog
Named, governed queries
read
Raw SQL
No free-form access to tenant data
never
Commands
The only path that writes
approval gate

Sold out is the most expensive lie. Hindsight corrects for stockouts, so a style that looks like a winner also shows the demand you couldn't fulfil.

AW25
High Rise Straight Jean
Size curves by style
26
Bought9.1%
Realized10.4%
Corrected9.0%
Recommended9.0%
LOST SALES
180units
WK-401Merino Crew Sweater
Sold out, with demand it never met.
MARKDOWN EROSION
31.5%of margin
WB-101Off Shoulder Blouse
Sell-through recovered, on clearance.
SIZE CURVE
3xover-bought tails
WJ-301High Rise Straight Jean
Bought flat. The middle sizes starved.

Turn the answer into a workflow. The Monday trade pack assembles itself: read the week, flag the risks, draft the reorders, then wait for your approval.

Monday trade pack
#14
Triggered by schedule · 3 of 6 steps done · 47s
Trigger
Every Monday 07:00
Start of the trading week
Query
This week on the floor
sales.byStyleWeek
, AW25, all
channels
Agent
Flag the risks
Styles running 20% under their clean-week rate of sale
Condition
Anything flagged?
True
flagged.count > 0
False
Action
Draft the reorders
Asks before it writes
Output
Monday trade pack
To the merchandising team
Output
All clear note
Short summary, nothing to do

Consequential actions wait for you.

A proposed markdown or reorder stops at the approval gate until you decide.

Approve 12 items?
Dispatches
replenishment.draftReorder.
This action is consequential and will be audited.

Every run is on the record.

Each step is a registered command: validated, permissioned, audited. Open any run and replay it.

Workflow
Run
Status
Monday trade pack
#14
Waiting for approval
Markdown watch
#63
Completed
Monday trade pack
#13
Completed
Monday trade pack
#12
Failed

Built around the systems you already run. One intelligence layer across your retail stack.

SOURCES
Orders & sales
Inventory data
Product data
Forecast data
Mountin.
INTELLIGENCE + WORKFLOW LAYER
Understand
Signals
Ask
Explain
Find the risks and opportunities in your data
Decide
Propose
Review
Approve
Turn what you find into the next move
Run
Scheduled
Triggered
Automatic
Make the decision repeatable across your systems
YOUR SYSTEMS
ERP / WMS
Commerce platform
Marketplace systems
Campaign tools
ERP / OMS
CHANNELS
Stores
E-commerce
Marketplaces
CRM / Paid
Wholesale

Answers stay tied to source data.

Every figure traces back to the query and the rows that produced it.

Actions follow your permissions.

Each command needs a permission. Roles decide who holds it.

Every workflow run is recorded.

Each run leaves a trace you can open and replay.

Bring the exports you already have. AI drafts the mapping. You confirm it.

  • Reads a sample and proposes every binding and transform
  • States its confidence and validates the sample rows
  • Writes config, never data. Nothing loads until you confirm
Mapping
orders
draft pending review - shape 100% of sample rows valid (high); meaning 83% of mapped fields derived with confidence
Canonical field
Binding
Transform
Sample values
order_external_id
*
Bestellnr
B-2001 · B-2002 · …
line_id
*
Positionsnr
1 · 2
date
*
Bestelldatum
date
04.03.2026 · 05.03…
sku_ref
*
Artikelnummer
WB-101-BLK-M · TS-…
quantity
*
Menge
2 · 1 · 3
unit_price_paid
* (unsure)
Einzelpreis
number
79,90 · 29,90 · 11…
Decimal separator
,
Date format
dd.MM.yyyy

Turn merchandising decisions into workflows.

See what is happening in the season, decide what to do next, and make the decision repeatable.

Book a demo