Manufacturing Autonomy: Ready, Not Ready, or Almost There?

Someone can manually check a couple of systems (MES, SCADA) and piece it together

There's a known troubleshooting path across 3+ systems, even if manual

Data from MES, SCADA, and the historian is already correlated somewhere

Some maintenance follows a fixed preventive schedule

Most maintenance is scheduled, with some condition-based exceptions

Maintenance is condition-based, using real sensor/machine data

At least two systems (e.g. MES and a paper log), handed off manually

Three or more systems (MES, SCADA, ERP, historian), loosely connected

Three or more systems that already need active reconciliation

It's run the same way informally, without a written standard

It diverges in small ways nobody's tracked down

The divergence is significant enough that supervisors flag it as a known problem

Some data exists, even if undocumented or requiring manual export

Data is accessible via OPC-UA/APIs/middleware with moderate effort

Well-documented, API-accessible data is already in active use

It's collected, even if scattered across systems

It's collected and partially unified, with some gaps

It's collected and unified enough to build on today

One could be set up, even if it'd take real effort

A test line exists but isn't regularly used

A staging/test line is ready to use right now

Feasible within a few months, given the right sponsor

Feasible within a few weeks, with a controls engineer involved

We've run non-disruptive pilots before — days to a couple weeks

There's at least some track record to point to

Leadership is frustrated — expected more ROI than what's been delivered

Leadership is actively pushing for something that moves OEE/downtime numbers

Budget exists for exploration (PoC-level)

Budget is approved pending a credible plan

Budget is approved for a real operational improvement initiative

There's rough agreement, even if not made explicit

One clear front-runner bottleneck exists

Full alignment across plant and corporate on the #1 bottleneck

There's cautious openness, even without prior exposure

There's exposure to assisted tools already, with reasonable buy-in

Success can be tracked with hard metrics (OEE, downtime hours, scrap rate, MTTR)