Go-live day gets the attention. The mechanical handover, the punch list, the moment the cranes or the shuttles or the conveyor line run their first live cycle — that's the date on the project plan, and it's the date everyone remembers. The maintenance programme that has to run behind all of it for the next fifteen years doesn't get a go-live date. It gets built quietly, in the weeks after, usually starting from a shelf of OEM manuals nobody on the maintenance team has had time to read cover to cover yet.
That gap — between "the equipment is installed" and "we have a maintenance programme that actually matches what the manufacturer specified" — is where the first ninety days of an automated line actually go. It's rarely where anyone budgeted the time.
What actually happens in the first 90 days
The commissioning punch list absorbs the first two to three weeks almost everywhere — snagging, calibration, integration issues between the new equipment and the WMS. That part is expected, and it's usually budgeted for.
What isn't always budgeted for is what happens next: someone on the maintenance team is handed the OEM manuals, often several of them, one per major subsystem, and asked to turn them into a working PM schedule. That work competes for the same hours as everything else already on their desk, and unlike commissioning, nobody enforces a deadline on it.
The manual is the bottleneck, and it's not close
An OEM manual for a piece of automated material handling equipment routinely runs several hundred pages. Service intervals are scattered across sections rather than collected in one table. Torque specs sit next to procedures that reference part numbers defined in a completely different appendix. None of that is a criticism of the manufacturer — it's a service manual, written to be read start to finish by a trained technician, not skimmed for the fifteen intervals that actually need to go into a CMMS.
Turning that into a working schedule is fundamentally transcription work, not maintenance engineering — and transcription work at that scale takes months, not days. Which is exactly why it usually isn't finished by the time it needs to be.
What normally fills the gap while the real programme gets built
None of what fills that gap is anyone doing a bad job. It's the predictable result of a deadline nobody set colliding with a task nobody has time for:
- A first-pass spreadsheet covering the intervals that were easy to find — usually the ones on the first few pages, not necessarily the ones that matter most.
- Reactive maintenance quietly covering for the tasks nobody has transcribed yet, which looks like normal breakdown response until someone traces a failure back to a missed interval.
- Warranty terms nobody is actually positioned to prove they met, because the paperwork trail for "we performed the manufacturer-specified service" doesn't exist yet.
- A six-monthly or annual task buried on a page nobody reached, which surfaces at the exact moment it becomes expensive — an early failure, or a warranty claim the OEM has grounds to challenge.
How we approach it
This is the specific ninety days Reliabilytics was built to remove. Upload the OEM manual for a crane, shuttle, conveyor or robot on day one, and the AI reads it in full the same day — every service interval, torque spec and safety-critical check — and turns it into a live PM schedule tied to the actual asset, not a spreadsheet someone has to remember to keep updating.
That doesn't replace the ninety days of real work a new line still needs — commissioning issues don't go away, and neither does the judgement a maintenance engineer brings once the equipment is running. What it removes is the transcription bottleneck sitting in front of that work, so the maintenance programme is complete against the manual from week one — and the ninety days go into running the equipment well, not typing it into a system.