Generate maintenance plans with AI based on equipment data and manufacturer recommendations, optimize plan frequencies based on execution history, and adjust counter thresholds adaptively.

AI Optimization

Once your maintenance program has been running for a few months, you have something valuable: execution data. The AI Optimization module uses that data to answer two questions that would otherwise require hours of manual analysis: "Are my maintenance intervals right for this equipment?" and "What maintenance should I be doing that I'm not?"

What is it for?

Maintenance intervals are usually set at implementation based on manufacturer recommendations or experience — and then never revisited. But equipment behavior is not static: a compressor running 24 hours a day in a dusty, high-temperature environment deteriorates faster than the same model running 8 hours a day in controlled conditions. Static intervals ignore this reality.

AI Optimization allows you to:

Generate initial maintenance plans for new assets based on manufacturer data and industry best practices
Analyze whether existing plan intervals should be increased or reduced based on actual execution history
Identify when an asset is wearing faster than expected and trigger more frequent maintenance automatically
Extend intervals for assets that are performing well, reducing unnecessary maintenance costs

How does it work?

The system analyzes multiple data sources for each asset:

Execution history of its maintenance plans (completion times, findings rate)
Machine events and alarms linked to the asset (how many critical events between maintenance cycles)
Sensor data trends (vibration, temperature, pressure patterns)
Asset metadata (manufacturer, model, operating conditions)

It then generates specific recommendations with the reasoning behind each one — not just "increase frequency" but "increase frequency because the asset had 3 critical events in the last 60 days, suggesting premature wear."

How to use it?

Generate plans for a new asset

When you add a new asset and need to build its maintenance program from scratch:

Go to Maintenance > Optimization in the sidebar.
Select the asset.
Click Generate Suggestions.
The AI analyzes the asset's technical data and searches manufacturer recommendations.

Each suggestion includes:

Field	Description
Plan name	Suggested descriptive name
Frequency	Recommended interval (e.g., every 90 days)
Checklist	Verification list based on manufacturer specifications and best practices
Priority	Suggested priority based on asset criticality
Justification	Plain-language explanation of why this plan is recommended

For each suggestion, you can Accept (creates the plan immediately, editable afterward) or Dismiss (ignores the suggestion without creating anything). Dismissed suggestions can be regenerated at any time.

Optimize an existing plan's frequency

After a plan has at least 5 executions, you have enough history for meaningful analysis:

Go to Maintenance > Optimization and select an existing plan.
Click Analyze Frequency.
Review the recommendation:

Recommendation	When it appears	What to do
Increase frequency	Asset shows premature wear or frequent incidents between cycles	Reduce the interval — maintenance was happening too late
Decrease frequency	Asset consistently in good condition, no incidents between cycles	Extend the interval — save cost without increasing risk
Maintain	Current frequency is well-matched to equipment behavior	No change needed

If you agree with the recommendation, click Apply to update the plan's interval.

Always review the reasoning before applying frequency changes on critical assets. AI recommendations are based on historical patterns — if something has recently changed (new operating conditions, different shift patterns), factor that in before accepting.

Condition-based maintenance triggers

Beyond scheduled maintenance, you can configure rules that trigger additional maintenance tasks when specific conditions are detected:

Condition detected	Automatic action
Temperature exceeds critical threshold	Generate urgent inspection task
Abnormal vibration pattern detected	Reduce maintenance interval by half
Equipment efficiency drops below 80%	Increase maintenance frequency
Same critical alarm repeats 3 times	Schedule extraordinary maintenance

These triggers work alongside the regular plan schedule — they add maintenance when conditions require it without replacing the baseline frequency.

Key benefits

Initial plan generation for new assets in minutes rather than days of manual research
Data-driven frequency recommendations based on actual equipment behavior
Plain-language justifications that explain why each change is recommended
Condition-based triggers for equipment that needs more attention during certain periods
Adaptive counter thresholds for hours-based plans that match real wear patterns

Common use cases

Scenario 1: Building a maintenance program for new equipment A new centrifugal pump arrives with only the manufacturer's basic maintenance guide. Instead of manually researching industry standards and building plans from scratch, the maintenance engineer selects the pump in Optimization and generates suggestions. The AI creates 3 plan proposals: monthly (general inspection and lubrication), quarterly (coupling alignment check), and annual (complete disassembly and bearing inspection). Each has a pre-built checklist. The engineer reviews, adjusts the checklist items based on local conditions, and accepts all three.

Scenario 2: Reducing unnecessary maintenance on a stable asset Centrifugal pump B-04 has a weekly inspection plan. After 8 months of executions, the Optimization analysis shows: 100% of inspections completed without findings, zero incidents between cycles, stable vibration trend. The AI recommends reducing to bi-weekly. The maintenance coordinator accepts. The team saves 52 inspection tasks per year without increasing risk for that asset.

Scenario 3: Identifying a deteriorating compressor Compressor C-02 has a monthly PM plan. After its last 4 monthly cycles, the Optimization analysis shows a pattern: each maintenance found more issues than the previous one (oil consumption increasing, filter clogged faster). The AI recommends increasing to bi-weekly. The reliability engineer accepts, and the more frequent maintenance catches a developing seal failure before it causes a production stop.

AI Optimization

On this page