Product

One plan, built from your own numbers.

plnnd plans from data extracts of your own systems: materials, stock, open orders, demand history, suppliers and more. The planning engine computes every number from that data. An AI model can explain the results, but never decides them.

The screenshots are the real app, on its three fictional sample companies.

Control Tower

What needs attention, at a glance.

The overview of every material at every location, by state and by location. It counts what is at immediate risk, which order windows are closing, and which actions protect against shortages or free up cash.

A line at the top says which items changed state since the last planning run.

The Control Tower for the Linea Apparel sample: 17 materials at immediate risk, 31 order windows due soon, the criticality distribution over eight states, priority replenishment actions and the largest excess stock.The Control Tower for the Linea Apparel sample: 17 materials at immediate risk, 31 order windows due soon, the criticality distribution over eight states, priority replenishment actions and the largest excess stock.
Linea Apparel sample
Materials

Every material at every location, in one list.

Each row shows the state and how long the item has been in it, stock, safety stock and reorder point, cover, forecast and value class, in the item’s own unit of measure.

Filter by state, planner, supplier, value class, VED class or demand pattern, and export the list to Excel or CSV.

The Materials list for the Holmvaart Water Services sample: spare parts at the central warehouse and depots, with value class, demand pattern, state and how long it has lasted, stock, safety stock and reorder point.The Materials list for the Holmvaart Water Services sample: spare parts at the central warehouse and depots, with value class, demand pattern, state and how long it has lasted, stock, safety stock and reorder point.
Holmvaart Water Services sample
Material Detail

The history and plan of one item.

A weekly and monthly chart in two panels: stock and incoming supply on top, demand below. The forecast is an outline that customer orders and work orders fill, with the range actual demand falls in about 4 weeks in 5. The projected stock has a likely range too, and a marker where a stockout becomes possible before a new order could arrive.

Real dates and calendar weeks run along the axis. Hover over a week to see its purchase order lines and orders; click a demand bar to open them. You can exclude an unusual demand day from the forecast, and export a one-slide management summary as an editable PowerPoint.

The Material Detail chart for the End-suction Pump ESP 65-160 at Venlo DC: stock and supply on top with the likely stock range, safety stock 23, reorder point 62, the order-by date, a possible stockout on 15 October and the stockout on 26 October, before an open order arrives at the end of November; demand per week below, with the forecast as outlines.The Material Detail chart for the End-suction Pump ESP 65-160 at Venlo DC: stock and supply on top with the likely stock range, safety stock 23, reorder point 62, the order-by date, a possible stockout on 15 October and the stockout on 26 October, before an open order arrives at the end of November; demand per week below, with the forecast as outlines.
Norvik Pump Systems sample
Simulation

See what the recommendations would do.

Next to the current plan, Simulation shows the item’s recommended actions and what following them would do: the stock line, and the plan and the recommendations side by side on state, stockout, order-by date, fill rate and cover.

Here, pulling in one purchase order makes the pump Reorder Due instead of At Risk, and moves the stockout 25 days later.

Simulation for the End-suction Pump ESP 65-160 at Venlo DC: the recommendation Pull in PO-41056 from 29 November to 9 October, a table of plan against with recommendations (At Risk becomes Reorder Due, stockout 26 October becomes 20 November) and the stock with the recommendation as a dashed line.Simulation for the End-suction Pump ESP 65-160 at Venlo DC: the recommendation Pull in PO-41056 from 29 November to 9 October, a table of plan against with recommendations (At Risk becomes Reorder Due, stockout 26 October becomes 20 November) and the stock with the recommendation as a dashed line.
Norvik Pump Systems sample
Manual simulation

Test your own changes before you make them.

Change demand up or down over a period, add a one-off order, move a purchase order line, change its quantity or take it out, or add extra supply. The engine plans the item again with your changes in seconds and shows the result next to the plan.

Nothing is saved: the plan stays as it is. A calculated simulation adds a slide to the PowerPoint export.

Manual simulation for the same pump: the purchase order moved to 23 October, Calculate, the plan against the manual simulation, and the simulated stock as a brown line on the chart.Manual simulation for the same pump: the purchase order moved to 23 October, Calculate, the plan against the manual simulation, and the simulated stock as a brown line on the chart.
Norvik Pump Systems sample
Orders

Late orders are checked, not planned blindly.

Each item lists its customer orders and work orders with how late they are. An order long past its date may be stale: check it to keep planning it, or ignore it with a reason to take it out of the plan until the ERP changes it.

Where your ERP sends them, delivered quantities and order status count too: only what is still open is planned, orders on hold are flagged, and closed or cancelled ones are shown greyed.

The orders list of the Circulator Pump CRC 32-80 at Venlo DC: a customer order 33 days late marked Possibly stale, with the buttons Checked, keep planning and Ignore this order, and a planned order below it.The orders list of the Circulator Pump CRC 32-80 at Venlo DC: a customer order 33 days late marked Possibly stale, with the buttons Checked, keep planning and Ignore this order, and a planned order below it.
Norvik Pump Systems sample
Forecast simulation

Compare forecast methods, then choose.

On Material Detail, every eligible method’s plan side by side, with what it would do to safety stock, the reorder point, the stockout date, orders and the item’s state.

One method is recommended, with the reason in plain words. You can simulate another on the chart and activate it.

Forecast simulation for the Sludge Pump Rotor of Holmvaart Water Services: Croston SBA recommended and active, Moving average (13 weeks) passed over for its bias, and per method the forecast, safety stock, reorder point, stockout and order-by date, fill rate and state.Forecast simulation for the Sludge Pump Rotor of Holmvaart Water Services: Croston SBA recommended and active, Moving average (13 weeks) passed over for its bias, and per method the forecast, safety stock, reorder point, stockout and order-by date, fill rate and state.
Holmvaart Water Services sample
Actions

What to do, per material, with the reason.

Seven action types: Expedite needed, Pull in, Push out, Cancel candidate, Place new order, Confirmation missing and Review unusable stock. Each comes with its reason.

Accept or dismiss each one: decisions are saved, and every screen shows their effect after the next planning run. Export the planned changes to Excel or CSV.

The Actions screen for the Norvik Pump Systems sample: open actions counted per type, and a list of Place new order actions with the planner and the reason for each.The Actions screen for the Norvik Pump Systems sample: open actions counted per type, and a list of Place new order actions with the planner and the reason for each.
Norvik Pump Systems sample
Supplier overview

One overview and one e-mail per supplier.

Choose a supplier in the top bar and every screen follows. Actions then shows that supplier’s open order lines, confirmations, delivery reliability (actual against planned lead time) and the open requests.

One e-mail draft holds all open requests. It opens in your own e-mail program: plnnd sends nothing itself.

The supplier overview for Norvik Deventer Plant, the sample company’s own plant that supplies its distribution centres: 20 open order lines, 2 unconfirmed, delivery reliability per distribution centre, and the button E-mail all open requests.The supplier overview for Norvik Deventer Plant, the sample company’s own plant that supplies its distribution centres: 20 open order lines, 2 unconfirmed, delivery reliability per distribution centre, and the button E-mail all open requests.
Norvik Pump Systems sample
Excess

Stock beyond what the plan needs.

Items whose stock above safety stock covers more than six months of forecast demand, and obsolete items that are discontinued, with their stock and cover.

Open orders that are no longer needed show up in Actions as Push out or Cancel candidate.

The Excess screen for the Linea Apparel sample: 34 excess and obsolete items such as insulated parkas and selvedge jeans, with stock, safety stock and reorder point.The Excess screen for the Linea Apparel sample: 34 excess and obsolete items such as insulated parkas and selvedge jeans, with stock, safety stock and reorder point.
Linea Apparel sample
Data

Your data in, checked before it is used.

The input files of the last planning run, with their row counts, and what the data check found: Fatal, Error and Warning, plus notices such as unusual demand days for the planner to review.

There is no direct ERP connection and no upload screen yet. For a pilot on your own data, we arrange the extracts with you.

The Data screen for the Holmvaart Water Services sample: input files with rows read and used, a note that uploading in the app is planned, and the data check with 0 fatal, 0 errors, 2 warnings and 27 notices, such as a work order past its date and demand days for the planner to review.The Data screen for the Holmvaart Water Services sample: input files with rows read and used, a note that uploading in the app is planned, and the data check with 0 fatal, 0 errors, 2 warnings and 27 notices, such as a work order past its date and demand days for the planner to review.
Holmvaart Water Services sample
Ask PLNND

A chat on every screen.

Ask in plain language, for example which items need action first, or which are past their order-by date. The answers come from the engine’s own numbers: the AI finds and words them, and every answer is checked before it is shown.

In the demo, plnnd’s own AI (Claude Sonnet 5) is included for 30 questions. It runs at Anthropic, outside the EU, so please don’t type personal details.

Ask PLNND opened on the Control Tower of the Norvik Pump Systems sample, before any question: suggested questions such as Which items need action first?, and the note PLNND’s demo AI (Claude Sonnet 5), 30 of 30 questions left, questions go to Anthropic outside the EU.Ask PLNND opened on the Control Tower of the Norvik Pump Systems sample, before any question: suggested questions such as Which items need action first?, and the note PLNND’s demo AI (Claude Sonnet 5), 30 of 30 questions left, questions go to Anthropic outside the EU.
Norvik Pump Systems sample
Settings

Your parameters, your layout, your AI.

Planning parameters, the target service levels in use with how often demand really ran past safety stock, locations, CSV format, table layout and the AI provider: plnnd’s demo AI, or a key of your own, which is then used instead. In the demo also the sample company and Start over.

The top bar sets the location, the scope (everyone, one planner or one supplier: every screen follows), weekly or monthly view, with or without unconfirmed supply, and search. Light and dark mode, and a layout for phones and tablets. A Help page explains every screen and term with worked examples, and every info tip links to it.

Settings, AI provider: PLNND’s demo AI (Claude Sonnet 5), 30 of 30 questions left, with the note that questions go to Anthropic outside the EU; below it, connect your own provider instead (Anthropic Claude, OpenAI, Google Gemini or Microsoft Azure OpenAI), and what is sent to the provider: a token instead of the material and location, never names, prices or planner notes.Settings, AI provider: PLNND’s demo AI (Claude Sonnet 5), 30 of 30 questions left, with the note that questions go to Anthropic outside the EU; below it, connect your own provider instead (Anthropic Claude, OpenAI, Google Gemini or Microsoft Azure OpenAI), and what is sent to the provider: a token instead of the material and location, never names, prices or planner notes.
Settings > AI provider
Under the hood

How the engine plans.

Seven forecast methods, plus a launch plan

Each method is backtested on the item’s own history, and only methods that fit its demand pattern (smooth, erratic, intermittent or lumpy) are eligible. Real accuracy is measured on the forecasts of earlier runs; an item whose forecast keeps missing is flagged Forecast needs review.

  • Moving average (13 weeks)Every demand pattern. The simple reference.
  • Exponential smoothingRegular demand, steady or erratic.
  • Croston SBAOccasional demand, such as spare parts with many weeks without use.
  • TSBOccasional demand that may be dying out: it notices when use fades.
  • Damped trendRegular demand that grows or declines, without extending the trend forever.
  • SeasonalA repeating yearly pattern, with at least two years of history.
  • Moving average with growthRegular demand that grows or shrinks steadily, with more than a year of history.
  • Launch planItems not sold yet: from a launch date and the expected weekly sales.

Eight states, one per material and location

plnnd shows how long an item has been in its state, for example "At Risk for 14 days".

  • EscalationA planner escalated the item.
  • CriticalThe date to order by has passed: an existing order has to come earlier.
  • At RiskThe date to order by is close. Ordering later cannot prevent the stockout.
  • Reorder DueStock plus open orders reaches the reorder point soon.
  • ExcessStock above safety stock covers more than six months of forecast.
  • ObsoleteThe material is discontinued.
  • Not sold yetNever sold. Planned on its launch plan when it has one.
  • HealthyNone of the above.

Safety stock and service

Safety stock comes from the item’s own forecast errors over a year of backtests, plus the supplier’s delivery spread when it is measured. Each item has its own target service level, and carries two labels: value class (ABC) and criticality (VED: Vital, Essential, Desirable). How ABC and VED work together.

Supplier reliability compares actual with planned lead times and flags suppliers that deliver later than planned. A check replays planning on earlier dates and shows how often demand ran past the safety stock.

Data in, and where it runs

Twelve extracts from your own systems: locations, materials, material-locations, suppliers, stock on hand, demand history, receipts history, open purchase orders, open customer orders, open production requirements, supplier reject rates and planner decisions.

Everything runs at Scaleway in Amsterdam (EU); e-mail goes through Scaleway in Paris (EU). plnnd has its own login service, and the app has no tracking, no advertising and no analytics cookies. One exception, in the demo only: its AI runs at Anthropic, outside the EU, on fictional sample data with names replaced by codes. With your own key, questions go to the provider you chose.

Ask PLNND

The engine decides. The AI explains it.

Every number in plnnd comes from the planning engine. An AI model puts the engine's reasoning into plain words and answers your questions in a chat on every screen. In the free demo that works straight away: plnnd's own AI (Claude Sonnet 5) is included, with 30 questions per demo.

  • Before facts go to the AI, material and supplier names are replaced by codes.
  • Every number in an answer is checked against the engine. If one does not match, you see the engine's own explanation instead.
  • The demo's AI runs at Anthropic, outside the EU, so please don't type personal details into the chat.
  • After 30 questions, or when the day's limit for all visitors is reached, plnnd shows the engine's own explanation.
  • Prefer your own AI? Connect a key for Claude (Anthropic), OpenAI, Google Gemini or Azure OpenAI. It is used instead, without the demo's limit, stored encrypted and never shown again.

Example · Norvik Pump Systems sample data

Why is the ESP 65-160 at Venlo DC at risk?
Stock runs out on 26 Oct 2026, 34 days before the open order PO-41056 arrives on 29 Nov 2026. Delivering by 9 Oct 2026 keeps stock above the safety stock of 23 units, so plnnd suggests pulling that order in.
Material Detail, Why this forecast method: the planning engine's own reason (Damped trend had the lowest lead-time error over 13 backtests), the line PLNND's demo AI (Claude Sonnet 5), 30 of 30 questions left, and the button Explain with AI.Material Detail, Why this forecast method: the planning engine's own reason (Damped trend had the lowest lead-time error over 13 backtests), the line PLNND's demo AI (Claude Sonnet 5), 30 of 30 questions left, and the button Explain with AI.

See it on a sample company like yours.

Want to try it on your own data? Talk to us about a pilot.