AI reorders every store, every supplier, every night.
Stokk is the AI replenishment and purchasing platform for multi-location retailers. It sits on top of your ERP and weighs every supplier and every store’s demand overnight, then surfaces the orders worth placing now: on each supplier’s cadence, or the moment a stockout risk says step in. The modern replacement for AGR and the Monday spreadsheet.
Sits on top of your ERPLive in two weeksEU-hosted
Lead time slipped to 21 days last quarter. Order now to keep parity ahead of the May 15 reset.
Cover is healthy and lead time is short. Re-evaluate in 9 days.
Two top movers trending up. Risk of stockout at the Berlin store before next cycle.
Connects directly to your ERP, reads stock and writes transactions back through it. One connector interface, every system below.
Your buyers spend Monday rebuilding the same order they built last Monday.
Across five stores or fifty, replenishment is the decision retail gets wrong most often and most expensively, too much of the wrong stock, not enough of the right. Most teams still make that call in a spreadsheet, by hand, by gut. Legacy planners automated the math but never the judgement. Stokk does both.
The old game
- Buyers rebuild order proposals from scratch every week
- One forecast model per SKU, blind to promotions and seasonality
- National averages hide which store is actually dry
- Stockouts and overstock, both, at the same time
- Supplier quirks live in one buyer's head, lost when they leave
- Every decision made by gut, justified after the fact
The new game
- AI drafts every store's order overnight; buyers approve in the morning
- Best-fit model per SKU per location, promotions and seasonality included
- Each store sized for its own demand, with lateral transfers between them
- Less cash on the shelf and fewer empty ones, at the same time
- Supplier knowledge written down in files the AI reads every night
- Every order explainable, by SKU, by store, by week
The fastest payback is the order you're already getting wrong.
Most teams start with the planning layer, forecasting, replenishment and purchasing, because that's where the cash and the hours are. It's also the cleanest swap: same suppliers, same ERP, same store managers. You just stop building order proposals from scratch.
Forecasting
Demand, per SKU per location
A best-fit model chosen automatically for every SKU in every store, seasonality, promotions and lifecycle included. New items fall back gracefully until they have history.
Replenishment
Each store, sized for itself
A draft order for every store every week, sized against the demand that store actually sees, with lateral transfers when one location is dry and another is overstocked.
Purchasing
The Monday Morning Brief
A ranked list of which suppliers to order from, with MOQ, freight and lead-time economics weighed and the reasoning written out. Approve, edit a few lines, send.
AI doesn’t dashboard. AI decides.
Legacy planners show you charts and wait. Stokk’s brain runs every night and makes the call. It evaluates every supplier against your forecast, your live stock, MOQs, lead times, freight terms, seasonality and the supplier intelligence files your team writes. By 09:00 it has drafted the orders and the transfers, explained, in plain language, ready for one human click.
- Reasoning attached to every order, every store, every week
- Supplier knowledge captured in human-written files, not buried in someone’s head
- Your data is yours, never used to train shared models
- 01 · Reads26 weeks of sales, current stock, open POs, supplier files.
- 02 · ForecastsPer SKU, per location. Best-fit model picked automatically.
- 03 · DecidesOrder now / wait / watch, per supplier, with reasoning.
- 04 · DraftsPOs, transfers, picks, pre-filled, ready for human approval.
Most planners hand you a number and walk away.
AGR and the Shopify forecasting apps stop at the purchase order. Stokk keeps going, to the stockroom, the shelf and the till. Counting, picking, transfers, labels and loyalty, all on the same data, all under one login. The plan you approve in the morning is the same plan your floor staff pick against that afternoon.
Counting
Continuous, multi-counter
Guided and blind counts on a cheap handheld or any phone, with several staff on one session at once. Posts adjustments back to your ERP. Trade the once-a-year shutdown for rolling weekly counts.
Fulfilment
Order picking + light WMS
Warehouse and sales-order picking, store transfers, inbound, put-away and shelf labels, a lightweight WMS on the same login. Shopify in-store pickup included.
Marketing
Loyalty, wallet passes, SMS
Run a loyalty programme on Apple & Google Wallet, then sync consented segments to your SMS or email tool, wired to the same product and stock data. No second integration to build.
Three layers, one brain, on top of the ERP you already run.
Plan covers the thinking. Operate covers the doing. Engage covers the customer. The AI brain wires it together. Adopt the layers in any order, they share data and decisions from day one, so the next module is a switch, not a second project.
Less cash on the shelf. Fewer empty ones. Hours back in the week.
After the first full ordering cycle, lost-sale events on A-classified items roughly halve.
Buyers stop building proposals from scratch. The Brief lands; they review and approve.
Stokk's per-store sizing and lateral transfers free cash that was sitting on a pallet.
Continuous counts replace the once-a-year shutdown count.
Directional outcomes from live deployments in pet supplies, beauty and design & homeware, stated as what the operating model changes rather than audited figures. Impact varies with category mix, lead-time variability and data quality.
Built for retailers who outgrew spreadsheets and never warmed up to legacy planners.
| Capability | Spreadsheets | Legacy planners | Stokk |
|---|---|---|---|
| AI supplier evaluation | Static rules | Every night | |
| Reasoning behind every order | In someone's head | Buried in config | Written in plain language |
| Per-store replenishment | Add-on | ||
| Lateral transfer suggestions | |||
| Counting, picking & labels on the same data | |||
| Loyalty + wallet passes + SMS | |||
| CRM wired to your stock | |||
| Retail reporting (no separate BI tool) | Manual | ||
| Setup style | – | Configuration project | Calibrated for you |
Migrating off AGR or a legacy planner?
We’ve moved teams across in two weeks without breaking the order rhythm. Same forecasting concepts, better tools, and a Monday Brief your buyers will actually open.
Three retailers run Stokk now. They tell the same story.
Real deployments in pet supplies, beauty and design & homeware, described by industry and anonymised until each clears use of its name. More verticals on the use-cases page.
“Migration took weeks, not months, and the store managers actually trust the draft now. They edit a handful of lines instead of rebuilding the whole order every week.”
“We swapped the once-a-year stock-take panic for rolling weekly counts the staff actually like, and the forecasts are right often enough that we trust them, with the reasoning attached when we don't.”
“Each showroom gets its own draft sized for what it actually sells, instead of a national average that's wrong everywhere at once.”
From overnight reasoning to Monday approval to floor pickup.
Four scenes, sixty seconds. The same operating system from the brain to the back office to the customer.
- 1Reads26 weeks of sales, current stock at every store, open POs, supplier knowledge files (SIFs).
- 2ForecastsPer SKU per location. Picks the best-fitting model and writes down why.
- 3DecidesOrder now / wait / watch per supplier, with the reasoning, the cover horizon and the freight flag.
- 4DraftsPre-fills POs and store transfers ready for human approval at 09:00.
Built so security review takes one meeting, not three.
Your data stays in your tenant
Multi-tenant by design with row-level security. Sales, stock and supplier data is logically isolated and encrypted at rest.
Honest about what we have
Supabase backbone with SSO, audit logs and per-role access controls. We don't claim certifications we don't yet have, security questionnaire and architecture diagram on request.
GDPR-aligned, EU-hosted
Pick your region. EU residency is the default for European customers. Data subject requests handled within 30 days.
AI without the data leak
Your data is never used to train shared models. Claude prompts are scoped per request and not persisted by the model provider.
Independent security review available on request.
Common questions before the call.
Is Stokk just an AGR replacement?
That's the most common reason teams switch, the AI replenishment and purchasing is where the fastest payback is. But unlike AGR, Stokk doesn't stop at the purchase order: counting, picking, transfers, labels and loyalty run on the same data when you're ready. Start with planning; grow into the rest.
Does Stokk replace our ERP?
No. Stokk sits on top of your ERP. Your finance team keeps the source of truth. We pull data, run the planning and the operations, and write transactions back through the same connector.
Which ERPs does Stokk connect to?
DK Plus, NetSuite, Dynamics 365 Business Central and Shopify all connect through one connector interface. If you're on something else we can usually start from a nightly file import while we add the connector, which is typically a few days of work.
How long does setup take?
Two weeks is typical, regardless of how many modules you adopt. Week one: connect the ERP and import history. Week two: calibration and team onboarding. You're using the system in week three.
What does the AI actually do?
Each night, Claude evaluates every supplier against your forecast, live stock, lead times, MOQs, freight terms and any open orders. It writes a recommendation per supplier and pre-fills the orders. A human always approves before anything is sent.
Do we have to adopt every module?
No. Most customers start with the planning layer, forecasting, replenishment, purchasing, and add counting, fulfilment and marketing as they want. They share data from day one, so adopting a second module is a switch, not a project.
How do we get started?
Book a demo. We show you the system running live on a sample, you show us your ERP and suppliers, and we plan a two-week onboarding, where your first brief runs on your own data, from there.
See the Monday Brief live, then we map it to you.
A 20-minute demo on a working sample, so you see exactly how Stokk drafts and explains a full week of orders. We learn your ERP and suppliers; your first brief on your own data lands in week two of onboarding.