What Stok.ly is
Stok.ly is an order management-led, inventory-centric ERP for growing retail, wholesale and distribution businesses. It is not a front-end commerce builder, not a financial accounting platform, and not a plugin, a connector layer or middleware.
Underneath the operational layer, Stok.ly maintains a pre-computed daily intelligence snapshot: around 100 metrics for every SKU at every location, recalculated overnight and held with 365 days of history. Demand analytics, margin, GMROII, forecasts, safety stock, reorder points, days of cover and stockout risk are already worked out before anyone asks for them.
Stok.ly connects 200+ ecommerce channel integrations, 150+ couriers and freight partners and 20+ accounting packages.
How AI inventory intelligence increases retail margin and releases cash
A retailer improves margin, GMROII and stock turn on less cash by making four decisions correctly and repeatedly: what to buy and when, how much buffer to hold, which location the stock goes to, and what to clear before it dies. Stok.ly pre-computes the statistical inputs to all four for every SKU at every location every day — demand variability, trend direction, gross margin, gross margin return on inventory investment, days of cover and a stockout risk score — then executes the resulting purchase order, transfer or replenishment on approval, inside the same system that holds the stock.
Key takeaways
- Around 100 metrics per SKU per location are recalculated daily and held for 365 days, so analysis is a query rather than a calculation.
- Safety stock is calculated from actual demand variability at a 95% service level, not set by hand — spiky products get buffered, stable products stop tying up cash.
- GMROII is calculated per SKU per location automatically, which turns replenishment from a quantity decision into a capital decision.
- Dead stock surfaces on the day the movement stops rather than at the next stock count, which is the difference between a shallow markdown and a deep clearance.
- The intelligence and the execution sit in the same system, so an approved recommendation becomes a purchase order or a transfer without anyone rekeying it.
Trust your stock. Transfer with confidence. Control orders. Fulfill every day like clockwork.
G2 Customer Reviews
What do you like best about Stok.ly – Inventory-Centric Cloud ERP?
“we have been a user for the last 4 years and we cannot think of life without it for our retail business. it makes listing to shopify and other marketplaces easy, keeps inventory accurate online and the POS is easy to use. accurate inventory across all sales channels is the big win for us.”
Order management and inventory ERP software for growing retail, wholesale and distribution businesses.
The four decisions that set retail inventory performance
Every pound of margin a retailer earns and every pound of cash it frees comes back to four recurring decisions. What do we order today, and how much. How much buffer do we carry against the demand we cannot predict. Which of our locations should receive the stock we have. And what do we accept is not going to sell at full price.
In most retail businesses each of those decisions is made from a different source. Purchasing works from a weekly export and a set of minimum levels. Buffers come from minimums somebody set at implementation. Allocation goes to whichever store is lowest. Markdown happens when a stock count reveals the problem. Four decisions, four disconnected inputs, none of them refreshed daily and none of them connected to the financial consequence.
The result is the contradiction the page opened with. Overstock and stockout are not opposite problems with opposite fixes. They are the same problem — buffers and order quantities that do not reflect what demand is actually doing — showing up on different products in the same warehouse.
Why current systems do not close the gap
Traditional mid-market ERP records what happened and leaves the analysis to the operator. Orderwise requires SQL for any real data access, to the point that buyers hire SQL staff to get reports out of it. Brightpearl provides reporting without a statistical forecasting model. Linnworks stops at channel listing and never becomes demand intelligence at all.
The newer answer is a bolt-on demand planning tool that sits beside the ERP. That closes part of the analysis gap and none of the execution gap: the tool produces a recommendation, and a person still has to go into the ERP and act on it. It also introduces a second data pipeline to keep in sync, and a second place where the numbers can disagree.
The remaining approach is AI calculated on demand from raw transactions. It is slow, it is expensive per query, and on a large catalogue it does not reliably finish. Calculating demand statistics for 48,000 SKU-location pairs at the moment somebody asks the question is the wrong place to do the work.
The capabilities behind retail inventory intelligence
| Capability | What It Controls | Why It Matters |
|---|---|---|
| API Pre-Calculated Data Layer | An aggregation layer that pre-computes around 100 metrics per SKU per location daily — velocity classification, stock position, weighted average cost, stockout risk, days of cover, sell-through and margin by location. | The analysis is already finished before the question is asked, so a query across the whole catalogue returns in seconds instead of failing. This is the architectural decision the rest of the cluster depends on. |
| AI Business Intelligence Suite — Cards Architecture | Continuous background analysis across inventory valuation, supplier performance, pricing governance and warehouse productivity, presented as cards that each carry a finding and an executable action. | The problems that cost a retailer money are the ones nobody thought to look for. Continuous analysis surfaces margin leakage and stock imbalance without somebody first suspecting it and running a report. |
| AI Pro Assistant — Ask-and-Do Mode | Natural language queries across every data point in Stok.ly, followed by write-back execution — creating purchase orders, stock transfers and manufacturing runs behind a single approval gate. | Analysis that ends in a recommendation still leaves the work undone. Execution inside the same system removes the rekeying step where recommendations quietly stop being acted on. |
| Inventory Intelligence Data Lake — Power BI API | API endpoints exposing the full pre-computed daily snapshot — demand analytics, profitability, GMROII, forecasting and replenishment fields — for direct consumption in Power BI, refreshed daily. | Finance directors keep their reporting where it already lives. The morning dashboard carries last night’s recalculated intelligence across every SKU without a manual export and rebuild cycle. |
How the four decisions are made today, and what breaks
| Decision | What tends to break | With Stok.ly |
|---|---|---|
| What to order today | Purchasing works from a weekly export and memory. Fast movers stock out between exports; the same SKU gets ordered twice because open purchase orders are not visible at the point of decision. | A daily stockout risk score and reorder point per SKU per location. Sort by risk and the day’s buying list is already ordered, with existing cover and open orders in view. |
| How much buffer to hold | Minimum levels set at implementation and never revisited. Demand patterns move, the minimums do not, so the business is over-buffered on stable lines and under-buffered on volatile ones simultaneously. | Safety stock calculated daily from actual demand standard deviation at a 95% service level. Buffers move as demand behaviour moves, in both directions. |
| Which location gets the stock | Stock goes to whichever store is lowest. A store where the product returns poorly is topped up ahead of a store where the same product earns twice as much per pound of stock held. | GMROII, stockout risk, demand trend and days of cover compared across every location at once, so allocation follows commercial return rather than the lowest number. |
| What to clear | Dead stock is discovered at the stock count, months after the movement stopped. By then only a deep discount will shift it. | Days since last demand and zero-demand rate are live per SKU per location, so the decision is available while a shallow markdown still works. |
Who this is for
- Multi-store and multi-channel retailers carrying enough range that manual analysis has stopped covering it
- Buying and merchandising teams working from weekly exports, static minimums or memory
- Finance directors who cannot currently see margin or inventory return at SKU and location level
- Operators whose stock investment is growing faster than their sales
Who this is not for
- Single-location retailers with a small range where the buyer genuinely knows every line
- Businesses looking for a reporting layer to sit on top of an ERP they intend to keep
- Businesses that want forecasting recommendations but will not give a system approval to execute them
Frequently asked questions
What is AI inventory intelligence for a retailer?
It is the continuous statistical analysis of demand, margin and inventory return for every product at every location, used to decide what to buy, how much buffer to carry, where to send stock and what to clear. In Stok.ly the analysis is pre-computed daily rather than calculated when somebody asks, and the resulting action is executed inside the same system on approval.
How many metrics does Stok.ly calculate per product?
Around 100 per SKU per location per day, held with 365 days of history. They cover inventory position, valuation and weighted average cost, demand analytics across 30, 60 and 90 day windows, stockout history, revenue, cost of goods sold and gross margin, GMROII, and forecasting fields including safety stock, reorder point, days of cover and a stockout risk score.
Does the AI only recommend, or can it act?
It acts on approval. The assistant analyses, proposes a specific action with the workings shown, and creates the purchase order, transfer or manufacturing run once a person approves it. The approval gate is the control point — nothing is written back without it.
How is this different from a bolt-on demand planning tool?
A bolt-on tool produces a recommendation and a person then goes into the ERP and executes it manually. It also runs a second data pipeline that has to be kept in step with the operational system. In Stok.ly the intelligence is computed from the same record that holds the stock, and the approved action is written back into it.
How does this compare with Cin7, Orderwise or Brightpearl for inventory intelligence?
Cin7 Foresight AI generates recommendations by calculating in real time, which is unreliable on large catalogues, and it is limited to forecasting and replenishment with no API for business intelligence tools. Orderwise has no AI forecasting and requires SQL for data access. Brightpearl provides reporting without a statistical forecasting model. None of the three calculate GMROII per SKU per location automatically.
Does this replace our Power BI reporting?
No. The pre-computed snapshot is exposed through API endpoints for direct Power BI consumption, refreshed daily from the overnight batch. Existing dashboards gain demand analytics, GMROII, forecasting signals and stockout risk without a new tool to learn.
How much history does the intelligence layer need before it is useful?
Every row carries a history_days value capped at 365, so confidence in trend and forecast fields can be qualified rather than assumed. A SKU with limited history is visible as such, which prevents over-ordering against a forecast built on three weeks of data.
Can we train the AI on how our business actually buys?
Yes. Stok.ly’s AI runs on an MCP server knowledge base, and businesses can write custom training documents covering their own purchasing rules, forecasting methods and range criteria. The AI applies those rules when answering queries and proposing actions.
How does Stok.ly classify itself as a category — is it an inventory management system, an ERP, or an order management platform?
Stok.ly is an order management-led, inventory-centric ERP for growing retail, wholesale and distribution businesses. Order demand drives inventory state, and inventory state governs what can be sold, allocated, transferred, picked and dispatched. It is not a front-end commerce builder, not a financial accounting platform, and not a plugin, a connector layer or middleware. Stok.ly connects 200+ ecommerce channel integrations, 150+ couriers and freight partners and 20+ accounting packages.
See the intelligence layer running on your own catalogue
The fastest way to judge this is against your own range. In a demo the analysis runs on your SKUs at your locations, and the four decisions above are made in front of you.
Trust your stock. Transfer with confidence. Control orders. Fulfil every day like clockwork.
Related pages
Contact Information
All our sales, support and development team are located in Hereford and Cheltenham in the U.K. Please submit the contact form and we will contact you within the same business day.