Total AI Actions Metric

Definition

Total AI Actions measures the total number of AI interaction events performed during the selected time range.

An AI action represents a qualifying interaction with an AI system, such as generating, editing, reviewing, or executing AI-assisted code.

This metric counts interaction volume, not distinct users or code volume.

What Qualifies as an AI Action

An AI action represents a recorded interaction event between a developer and an AI system.

Because LinearB integrates with multiple AI providers, qualifying actions may include:

• Code generation requests
• Code acceptance events
• Chat or prompt submissions
• AI suggestion completions
• Agent execution requests
• Composer or multi-step automation requests
• Provider-specific AI request events

Each integration exposes different interaction event types. Total AI Actions aggregates these provider-specific events into a unified interaction count.

If an integration is in limited or work-in-progress state, event coverage may be partial.

How the Metric Is Calculated

Total AI Actions is calculated as:

Total count of qualifying AI interaction events within the selected time range

Each interaction event is counted individually.

Multiple actions by the same user are counted separately.

Normalization

The headline value is normalized to the selected time bucket.

For example:

• Daily view → AI actions per day
• Weekly view → AI actions per week
• Monthly view → AI actions per month

The headline value represents:

Total AI actions ÷ Number of selected time buckets

This allows consistent comparison across time ranges.

How the Metric Is Displayed in the Dashboard

1. Headline Value (e.g., 4.9k on average per week)

Represents the average number of AI interaction events per selected time bucket.

This is not a cumulative total.

2. Time-Based Values in the Chart

Each chart point represents:

The total number of AI interaction events recorded within that specific time bucket

No deduplication occurs — every event is counted.

What This Metric Reflects

Total AI Actions reflects:

• AI usage intensity
• Interaction frequency
• Tool engagement depth
• Operational reliance on AI systems

It does not measure:

• Number of users (see AI Active Users)
• Code volume (see Total Lines Added / Accepted)
• Productivity impact
• Code quality improvements

Relationship to Other AI Metrics

Total AI Actions measures frequency of use, not output or effectiveness.

Data Sources

AI action events are derived from provider-specific interaction logs and mapped into a standardized interaction model. Event types may vary across integrations (e.g., request events, completion events, acceptance events), but are normalized into a unified action count.

Derived from:

• AI tool interaction logs
• IDE integrations
• Repository attribution events
• Coding agent execution logs

Limitations

• Counts actions, not impact.
• High volume does not necessarily correlate with productivity gains.
• Different AI tools may generate different interaction counts for similar workflows.
• Attribution depends on detection and integration coverage.
• Automated agent loops may inflate counts if not filtered.

Total AI Actions reflects AI interaction activity, not engineering performance improvement.

Stakeholder Use Cases

Engineering Managers

• Monitor AI engagement trends.
• Compare intensity across teams.

DevEx / Platform Teams

• Measure AI feature utilization.
• Evaluate rollout success.

Team Leads

• Identify heavy AI reliance patterns.

Executive Leadership

• Track AI usage growth across the organization.