SupplyChainIntel

Connects to the systems you already run

Know where your supply chain stands before the first meeting of the day.

SupplyChainIntel pulls from your ERP, procurement, warehouse, dealer, and freight systems, reconciles them into one model, and has twelve operating metrics computed and alerted by six in the morning. No spreadsheet consolidation, no waiting on a weekly report.

Book a walkthrough See how it works

Runs on realistic demo data from day one, so you can evaluate it before connecting a single production system.

6:00 AMmetrics ready before the team is
7 feedsreconciled into one view
14 daysof forward demand visibility
5 alert rulesraised automatically, not hunted for
supplychain.internal / dailyDaily operating view92.8%Fill ratebelow 95% target91.4%Supplier on-timeabove target96.2%Production attainmentabove target3Freight exceptionswithin toleranceAlert raised automaticallyFill rate 92.8 percent, below the 95 percent targetWarranty claims on part E-21403 are 3.2 deviations above that part's own baseline

The morning view. Every number computed from your own systems, with breached thresholds already flagged.

The problem

The data exists. The answer does not.

Every number a supply chain team needs is already in a system somewhere. The cost is in getting it out, lining it up, and noticing when it moves.

Answers arrive too late

By the time a weekly report shows the fill rate slipping, the shortage has already worked its way through to the line and out to the dealer. The reporting cycle is slower than the problem.

Every system tells a different story

The ERP, the warehouse system, and the freight portal each hold part of the picture in a different shape. Reconciling them is a person with a spreadsheet, repeated every week.

Problems are found, not flagged

A warranty spike on one component, a carrier quietly slipping on delivery times, a supplier drifting below target. These show up in cost totals months after they were visible in the data.

How it works

Connect, reconcile, and watch.

The platform runs on a schedule and does the same work every day without being asked.

1

Connect your systems

Each source system plugs in behind a common interface, so ERP, procurement, warehouse, dealer, freight, warranty, and demand data all land in the same shape. Adding or swapping a system does not change anything downstream.

2

Let it reconcile

Data is collected every fifteen minutes and normalised into one model. Duplicate and updated records are resolved automatically, so the picture stays current through the day without manual intervention.

3

Read the morning view

Twelve operating metrics are computed at six each morning with any breached thresholds attached. Weekly rollups follow on Monday, covering inventory turns, forecast accuracy, and cost.

1Collectevery 15 minutes2Reconcileone common model3Compute12 metrics at 6:00 AM4Alertthresholds attachedThe same sequence runs every day without anyone asking it to.

What changes for you

The work that stops being work.

Most of the value is not a new number. It is the same numbers, earlier, without anyone assembling them.

TaskHow it works todayWith SupplyChainIntel
Preparing the morning reviewSomeone pulls exports from three systems and reconciles them in a spreadsheet before the meeting.The view is already computed and waiting at six, with alerts attached.
Spotting a supplier slippingNoticed when a shortage hits the line, then traced backwards through purchase orders.On-time delivery tracked per supplier on a rolling window, flagged when it crosses your threshold.
Catching a component failureVisible in warranty cost totals a quarter later, once the field population is large.Claims compared against each part's own history, so a genuine spike is flagged while it is still small.
Answering what demand will beA planner's judgement, defended in a meeting, revisited every week.A fourteen day forward number per part and region that everyone is looking at together.
Onboarding a new source systemA new report, a new extract, and a new spreadsheet tab that someone has to maintain.One connector behind the same interface. Nothing downstream changes.
Demand actuals3,750Supplier orders150Inventory snapshots125Dealer inventory120Production output84Freight shipments80Warranty claims60Rows pulled per collection cycle from the built-in dataset, linear scale. 4,369 rows in total, and demand history is most of it. A cycle runs every fifteen minutes.

A morning with it

What the first hour looks like.

The example below uses the built-in demo dataset, which is what you see on day one.

06:00

Overnight data is in and the day's metrics are computed. Fill rate has dropped to 92.8 percent, below the 95 percent target, and the alert is already attached to the record.

07:15

The planner opens the dashboard rather than a spreadsheet. Two suppliers are below their on-time target and both are visible without anyone going looking.

08:30

Warranty detection has flagged one turbocharger part running well above its own claim baseline. Quality is looking at a component issue rather than a cost line.

09:00

The operations review starts with everyone reading the same numbers, computed the same way, three hours before the meeting began.

What you get

The questions your operations team asks every week.

Each one answered on a schedule, with the alert raised before anyone has to go looking.

Are we filling orders?

Fill rate, open and critical backorders, and active purchase orders tracked daily, with a warning the moment fill rate drops below your target.

Are suppliers holding to date?

On-time delivery tracked by supplier across a rolling thirty days, with on-time and late counts, so a supplier drifting off target is visible before it becomes a shortage.

Is freight moving?

On-time percentage and exception counts by carrier, so a carrier having a bad month is a number on a dashboard rather than a surprise in a cost review.

Is a component failing?

Warranty claims are compared against each part's own history, so a genuine spike on one component is flagged as a systemic issue instead of being averaged away in the total.

What will demand be?

A fourteen day forward forecast per part and region, blending trend and recent average with a weekend adjustment, so planning has a number rather than an instinct.

Is inventory working?

Inventory value by location, in-stock counts, and inventory turns computed weekly, so capital tied up in the wrong place is visible while it can still be moved.

Who it is for

Built for the people who own the numbers.

Manufacturing and distribution operations where parts, suppliers, and freight all have to line up.

Supply chain leadership

One reconciled view of fill rate, delivery performance, and inventory position, refreshed before the morning stand-up.

  • Daily operating metrics
  • Weekly performance rollups
  • Threshold alerts by exception

Planning and procurement

Supplier performance and forward demand in the same place, so expediting decisions are made on data rather than on the loudest email.

  • Supplier on-time tracking
  • Fourteen day demand view
  • Backorder and PO status

Quality and warranty

Statistical detection on claim history per component, surfacing a systemic failure while the field population is still small.

  • Per-part claim baselines
  • Warning and critical tiers
  • Claim cost tracking

Under the hood

For the team that has to approve it.

Deployed as a single service. Your data stays in your database.

Source systemsInventory, supplier orders, production output, dealer inventory, freight, warranty, and demand
Refresh cycleCollection every 15 minutes, metrics at 6:00 AM, rollups Monday at 7:00 AM
Metrics12 daily operating metrics and 11 weekly aggregates
Alerting5 configurable thresholds, evaluated automatically and stored with the day's numbers
Bulk loadingInventory snapshots accepted by CSV upload with column validation
InterfaceReact dashboard with time series, anomaly tables, and supplier and carrier performance
DeploymentPython and FastAPI with PostgreSQL, deployable as one service
Evaluation modeFull demo dataset built in, so every metric and alert works before any system is connected

Questions

What a buyer usually asks first.

How long before we see our own numbers?

The platform runs immediately on built-in demo data, so you can evaluate every metric, alert, and forecast on day one. Time to your own numbers depends on access to your source systems rather than on the platform.

Do we have to replace anything we already run?

No. It reads from the systems you have and adds a reporting layer on top. Nothing is migrated and no system of record changes.

What if we use a system you do not support?

Every source sits behind the same interface, so adding one is a self-contained piece of work that does not touch the rest of the platform. The demo connector is a working reference implementation.

Where does our data live?

In your own database, in your own environment. The platform is deployed as a single service that you host.

How does the warranty detection avoid false alarms?

Each part is compared against its own claim history rather than a global average, with a warning tier and a higher critical tier. A part that always generates claims does not trip the alert; a part that suddenly changes behaviour does.

What stage is the product at?

A working MVP, deployed and running. Stated openly: there is no automated test suite yet, the ERP connector covers inventory and orders but not yet freight, warranty, or demand, and there is no user authentication, so it should sit behind your own access control for now. Alert email and Slack delivery is configured but not yet wired up.

See your own operating picture, not a slide deck.

Book a walkthrough on the built-in dataset, then compare it against the report your team runs today. Bring the report. It is the fastest way to tell whether this is worth your time.

Book a walkthrough How it works

Self-hosted. Your data stays in your environment.