Manufacturing
Planr
Demand intelligence for multi-plant manufacturers.
The field report
The planner’s shortest decision is still taking hours.
Import-dependent plants live inside a 45–65 day replenishment window. Inventory, transport, and demand data each live in their own file. So the simplest question (can we keep the line running next month?) becomes a day of phone calls. Planr compresses that answer to minutes.
- Deployment
- Live · 8 plants
- Input files
- 4
- Fields normalised
- 27+
- Decision cycle
- Days → minutes
The crisis
The decision that keeps breaking.
On a Tuesday morning at a filter plant outside Nasik, a planner opens his laptop to the same alarm that greeted him on Monday. Media 007, an imported filter substrate, is down to 8 kilograms against a daily burn of 46. Including material in transit, he has coverage for two and a half days. His supplier’s replenishment lead time is 45 to 65 days. The math leaves him looking at a shortage of 912 kilograms across a nineteen-day window.
Behind him are seven other plants in the same network, each with its own lag report and its own bench of surplus. The fix already exists somewhere in that network: the same part sits at two of the sister plants, a compatible substitute sits at a third, and a master roll in the building can be cut down to the width he needs. The scenario above is real, and Planr surfaced all three paths within seconds of the data loading.
“From lag report to actionable shortage resolution options, in one session.”
The gap
Why planning stays manual.
The problem with manufacturing planning is cohesion, not raw data. Four operational reports carry almost everything a planner needs: a Lag Master for stock and buffers, a Daily Status Report for open POs and ETAs, an Inventory Value Report for plant-level coverage, and a Transport Master for inter-plant cost and transit. Between them, twenty-seven fields describe the full state of the network. But the reports arrive in separate files, at different cadences, owned by different teams.
Before a planner can make a call, someone has to stitch the files into one view. That stitching is usually a spreadsheet, sometimes a phone call, often both. Planners end up spending most of the day consolidating and almost none of it deciding. Every at-risk part becomes a small coordination project, and the institutional knowledge (who substituted what, with what consequence) lives in the heads of whoever happens to still be in the room.
What Planr is
Introducing Planr.
Planr is a demand-intelligence platform for multi-plant manufacturers. It ingests the four operational reports, normalises them against a stable internal schema, and runs an operations-research engine over the result. What comes out the other side is a menu of resolution options, each already ranked and costed so the planner can act on them directly.
Underneath, the product is three layers. A data layer accepts any ERP’s native exports and maps their columns to a standard vocabulary. An analytics engine, built on operations-research methods rather than LLMs, computes shortage windows and the full set of resolution pathways. A decision interface mirrors the tabular conventions of Excel, so planners can read the output without retraining. On top of those, a Phase-2 AI layer turns hundreds of options into three ranked plans.
The four layers
- Accepts four structured report types. A configurable field-mapping interface aligns any ERP’s column names to the platform’s 27+ standard fields without code changes.
Dashboard & drill-down
Shortage intelligence.
Once the four files are in, Planr runs a shortage computation across every plant and every raw material at once, which is the thing a spreadsheet can’t do. A plant-level dashboard quantifies the network’s position in a single view: which parts are covered, which are at risk, and the production exposure expressed in both kilograms and days. Click any at-risk part and the analytical context opens: the RLT window, on-hand stock, material-in-transit against confirmed ETAs, the exact stockout window, the shortage quantity, and the replenishment quantity required to restore the green-level buffer.
The same analysis runs in two planning modes. In MTS/MTO mode, planners with clean BOM and demand data tie shortages directly to production commitments. In Lag-Based mode, planners in stock-level environments use consumption rates as the demand proxy. Both modes feed the same options engine downstream.
Resolution pathways
The Options Engine.
For every at-risk part, the engine computes a structured set of resolution pathways, organised into four families. Finding the same part at a sister plant is the simplest case; the engine just needs the surplus at the source plant above its own green level, the transport cost, and the transit days. A compatible substitute at another plant is harder, because the engine has to quantify wastage from width mismatch on top of cost and transit. And when the required width isn’t held anywhere in the network, it computes cutting combinations from available master rolls, minimising waste inside the machine-constraint bounds.
Each option lands in a single row: source plant, source part, width and grade match, transferable quantity, cutting wastage where applicable, transit days, transport cost, total delivered cost. Planners work through the rows in a tabular view, pick an action, and record their commentary for the audit trail. Every number is visible and every number is auditable.
Same part · different plant
Plants within the network that hold surplus stock of the exact part number. Computes transferable quantity (above the source plant’s own green level), transport cost, and transit days.
Substitute · different plant
Compatible substitute parts at other plants (same material class, close specifications). Presents wastage from width mismatch, transferable quantity, transport cost, and total cost.
Substitute · same plant
Parts already held at the requesting plant that can be converted or re-allocated. Minimises transport cost at the expense of potentially higher cutting waste.
Master roller cutting
Feasible cutting combinations from available master rolls. LP-optimised to minimise waste inside the 3–12% machine-constraint window.
AI decisioning
From options to plans.
The options engine handles the feasibility problem; it enumerates every pathway the laws of the network allow. The AI layer handles the optimisation problem, picking the right subset from that enumeration. In Phase 2, the OR engine’s structured output goes to a reasoning agent that sees every at-risk part and every plant in a single pass. Rather than another list of options, the agent returns three complete plans, each covering every shortage part with per-action reasoning the planner can validate.
Each plan is built against a different optimisation goal. One minimises total spend and leans on transfers; another minimises time-to-resolve and leans on the shortest transit; the balanced plan makes a deliberate trade across cost, speed, and waste. The agent accounts for cross-plant dependencies (surplus used for one plant is no longer available for another), so each plan is internally consistent. Click any action and the underlying combination detail opens alongside it, with the same calculation tables, wastage breakdowns, and formula tooltips from the options view.
“Hundreds of unranked options to three plans with reasoning, in a minute instead of a morning.”
For commercial teams
What-If, a decision layer for sales.
The other person who lives inside a shortage story is the sales owner. They’re the face of every commitment to a customer, but they don’t control the plant, and they can’t answer “can you deliver 2,500 units by the 18th?” without pinging Planning and waiting for a reply. What-If Analysis is Planr’s answer for them. Sales describes the order in free text (part number, quantity, customer need-by date, plant), the assistant confirms the ask, and then runs through three qualifiers: can delivery be split, is air freight acceptable, would an alternate equivalent part be accepted.
Behind the scenes the feature runs the same cross-plant analysis a planner would: BOM explosion, inventory check at the requested plant, a review of open POs and the MIT pipeline, a cross-plant network scan for transfer availability, lag calculation per component, and air-freight feasibility for import parts. The output is a feasibility answer, a fulfilment timeline, a list of highlighted BOM blockers, and an action plan the sales owner can take back to the customer with options instead of a callback.
Commit with options, not guesswork.
In production
Proof & deployment.
Planr is live at Fleetguard Filters, an eight-plant manufacturing operation in India, covering all raw-material planning for import-dependent media and roll materials. The current deployment runs on manual file upload. Enterprise rollout transitions to direct ERP integration, with Oracle as the first connector in Phase 2.
The platform has no plant-count limit. Any organisation can begin with their existing lag report, DSR, inventory value report, and transport master in whatever format they already produce. The field-mapping metadata layer handles the schema alignment; no code changes required.
Bring the clock back under control
Bring the decision back under a clock you control.
Planr is in production today at a live, eight-plant operation. It adapts to your ERP through configuration rather than a rebuild. Start with the lag report you already produce, and the rest follows.