- HS-203983 feeds agree
- HS-204023 feeds agree
- HS-20417WMS ≠ EDI
- HS-204333 feeds agree
- HS-204513 feeds agree
- HS-20476ID missing
- HS-204883 feeds agree
AI readiness assessment · two weeks
Evidence before the pilot
We examine six areas of your inventory data and return one verdict, with the evidence a pilot recommendation needs.
- Identifier consistency8% issues
- Feed ownership1 unowned
- History depthPass
- Demand labelsPass
- Refresh latencyWatch
- Decision ownerNamed
Six areas we examine before any pilot
Each area ends in a finding, an owner and evidence you can check. Sampled from your own feeds.
Identifier consistency
Does one SKU carry one ID across WMS, ERP and supplier EDI? A 500-record sample measures it.
Feed ownership
Who answers when a feed breaks. Every feed needs one named owner.
History depth for the forecast
Enough clean weekly demand per warehouse to train a forecast and hold weeks back to test it.
Label quality of demand
Sales, transfers and returns labelled apart, so the forecast learns demand, not stock moves.
Latency and refresh
How old each feed is when the forecast runs, and whether refresh fits the planning cycle.
The decision owner
One named person approves the evaluation scope and calls go or stop.
Three verdicts, and the evidence each needs
Every assessment ends on one rung of this ladder. Each rung names the proof it needs, so nobody argues from a hunch.
- Hollis Supply sits here
01 Repair first
Sampled issue rate8%gate 1%10%The sampled issue rate is above 1%, or a feed has no owner. A pilot now would test the data, not the model.
Evidence it rests on
- Issue rate per feed, from the sample
- Ownership register with every gap marked
- Repairs ranked by how many records they fix
Moves up when owners sign the plan
-
02 Conditional
Issue rate, repair agreed8% to ≤1%gate 1%10%Issues are known and bounded, and a repair plan has owners and dates. The pilot scope can be drafted, not started.
Evidence it rests on
- Repair plan signed by each feed owner
- A re-sample date in the calendar
- Draft evaluation scope and baseline method
Moves up when a re-sample shows ≤1%
-
03 Pilot-ready
Re-sampled issue rate≤1%gate 1%10%The gate is passed and the baseline is agreed before work starts. Now a pilot can be measured fairly.
Evidence it rests on
- Re-sample at or below the 1% gate
- Baseline measured on held-out weeks
- Evaluation scope approved by your owner
Leads to a scoped pilot
Sample report
What you hold at the end of week two
A readiness report, a feed ownership register and a prioritized repair plan. €6,000, fixed before it starts.
- 1
Readiness report
Findings per area, the evidence behind each and one verdict your owner can act on.
- 2
Feed ownership register
Every feed, its owner and its refresh, signed by the people who answer for it.
- 3
Prioritized repair plan
What to fix first, who owns it and when the re-sample runs against the gate.
Before you request one
The questions operations owners ask us in the first call, answered plainly.
How long does the assessment take?
Two weeks from kickoff. Week one samples your feeds and interviews their owners. Week two writes up the findings, the register and the repair plan, and walks your owner through the verdict.
What access do you need to our data?
Read-only exports or read access to each inventory feed, such as your WMS, ERP and supplier EDI, plus your weekly demand history. We sample 500 records and never need write access.
Who must be involved from our side?
The operations owner accountable for the pilot decision, one contact per feed and someone who knows your planning cycle. Short interviews in week one, then the walkthrough at the end.
What if we are not ready for a pilot?
Then the verdict is Repair first, and you keep a ranked repair plan with an owner on every item. You can run the repairs yourselves or with us. A pilot is scoped only after a re-sample passes the 1% gate.
Do you build the forecasting model?
Not in the assessment. When the verdict is Pilot-ready, we scope a pilot with a baseline, an evaluation period and acceptance conditions, quoted separately. No pilot outcome is promised before discovery.
Start with a two-week assessment
€6,000, fixed before it starts. A readiness report, a feed ownership register and a repair plan your owner can act on.