Issue rate 8%
Pilot gate ≤1%
- VerdictRepair first
AI pilot readiness for midmarket operations
Know what your pilot must prove.
Datril checks whether your inventory data can carry an AI forecasting pilot, repairs what blocks it and sets the test first.
- ERP stock levelsPlanning
- Warehouse WMSWarehouse ops
- Supplier cataloguePurchasing
One named owner per feed
500 records sampled
40 conflicting SKU IDs
8% issue rate
- 2 weeksReadiness assessment
- €6,000Fixed fee, agreed up front
- 3Deliverables you keep
- ≤1%Issue rate before any pilot
Is your data ready to carry a pilot?
Five checks we run before budget goes to forecasting.
Feeds disagree on the same item
At Hollis Supply, the WMS, ERP and supplier EDI name one SKU two different ways.
No named feed owner
A feed without an owner never gets reconciled.
No agreed baseline
Without 20% fixed as the baseline, a 16% result proves nothing.
Issue rate never measured
We sample 500 records. A pilot starts only at 1% or below.
The decision has no owner yet
Your operations owner decides whether to proceed.
Request an assessmentTwo weeks, three deliverables
The €6,000 assessment hands over three documents, and they end in one decision: repair the feeds first, or scope a pilot.
What we sampled, what failed and why, and whether the use case is ready for a pilot.
- Records sampled
- 500
- Inventory feeds
- 3
- Failed ID checks
- 40 8%
Every feed, its accountable owner, its refresh cadence and the checks that make its quality observable.
| Feed and owner | Refresh | Checks |
|---|---|---|
| WMS stock levelsInventory control lead | 15 min | SKU matchFreshness |
| Supplier catalogueProcurement manager | Daily | ID formatDuplicates |
| ERP order historySales operations analyst | Hourly | SKU matchNulls |
The fixes in order of impact, each with an owner, ending in one decision your operations owner takes.
- 1Reconcile SKU identifiers across feedsInventory control
- 2Validate supplier IDs at importProcurement
- 3Resample until issues reach 1% or lessOperations owner
Where it leads
Repair plan
Fix the feeds first, then resample.
Scoped pilot
Baseline, evaluation period and acceptance conditions agreed.
No pilot outcome is promised before discovery.
Use caseWeekly demand forecast12,000 SKUs, three warehouses
VerdictRepair firstA pilot waits for a 1% issue rate
FeedWMS stock levelsOwner: inventory control lead
Register3 feeds, 3 ownersRefresh: 15 min, hourly, daily
Pilot gate: ID issue rate≤1%Sampled before the pilot starts
Decision ownerMarta LindqvistHead of Operations, Hollis Supply
Every phase ends at an agreed gate
Each phase has a measure, a threshold and an owner who signs it off. Nothing moves forward on a hunch.
The same 500-record sample, measured before and after repair. Amber marks a missing or conflicting SKU identifier.
- 01
Assess 2 weeks
Sample each inventory feed, measure the issue rate, name every feed owner.
- 02
Repair from the plan
Reconcile identifiers in priority order and re-sample until the rate holds.
Gate · sampled issue rate at or below 1%
- 03
Pilot 6 weeks
Forecast on the same held-out weekly demand data the baseline was set on.
Gate · weighted error from 20% to at most 16%
- 04
Decide your owner
Your operations owner reviews the evidence and calls go or stop.
One distributor, one agreed test
Hollis Supply runs three warehouses with 12,000 active SKUs.
We agreed what the pilot has to beat before we spent anything on it. That made the go or stop call mine, and simple.
Proposed acceptance conditions
Agreed with Hollis Supply before the pilot. Thresholds, not results.
Weighted forecast error
Same held-out weekly demand dataset, lower is better
- Sampled issue rate
- 8%40 of 500 records
- ≤1% before the pilot
- Pilot length
- 6 weeksForecasting pilot
- Starts after the gate
- Decision
- Go or stopOperations owner
- Approves the scope
Fixed fee
Know before you pilot
A two-week assessment shows whether your feeds can carry a forecasting pilot, and what to repair first.
Readiness assessment
Start hereFor your operations owner, before any pilot.
€6,000fixed fee
two weeks
two weeks
- Readiness report
- Feed ownership register
- Prioritized repair plan
- Records sampled per feed
- Findings walkthrough
Repair and pilot
Measured against a baseline you agree first.
Scoped and quoted from your repair plan, once the assessment shows a pilot is viable.
- Feed reconciliation
- Ownership per feed
- Six-week forecasting pilot
- Owner decides go or stop
- Pilot starts only when the sampled issue rate is ≤1%
No pilot outcome is promised before discovery. The assessment tells you whether a pilot is viable.
Distributors who measured first
Midmarket operations teams that assessed their feeds before they funded a pilot.
Nordvik Cold Chain
Frozen food distribution
- Feeds assessed3 stock feeds, 2 zones
- EngagementAssessment, 2 weeks
- StatusRepair plan agreed
Move across the cards to browseSwipe the row, tap a card for its record
See the Hollis case-
Nordvik Two temperature zones, three stock feeds reconciled before a replenishment pilot.
Find out what your data can prove
Tell us about the feeds, the decision you want to support and who owns it. We reply with a scoping call, not a pilot pitch.
- Within two working days a 30-minute scoping call with your operations owner.
- Fixed fee, €6,000 for the two-week assessment, agreed before it starts.
- Your owner decides whether a repair plan or a pilot comes next.