Skip to main content

User Profiles & Behavioral Insights

Raven Docs includes a sophisticated behavioral analysis system that builds personalized user profiles based on activity patterns. Rather than simple XP or levels, the system evaluates users across seven core behavioral traits.

Overview

The profile system analyzes your activity signals to create a comprehensive behavioral profile:

The Seven Traits

Your profile is evaluated across seven behavioral dimensions:

Focus

Ability to maintain concentrated attention on tasks and resist distractions.

Behavioral Indicators:

  • Deep work sessions (>30 minutes on single tasks)
  • Single-project focus over task-switching
  • Minimal context-switching between unrelated items
  • Sustained engagement with content

Score Guide:

ScoreDescription
9-10Exceptional focus, extended deep work sessions
7-8Strong focus with occasional distractions
5-6Moderate focus, some task-switching
3-4Difficulty maintaining attention
0-2Frequent context-switching

Execution

Ability to complete tasks and deliver results.

Behavioral Indicators:

  • High task completion rate (>80%)
  • Meeting deadlines consistently
  • Following through on commitments
  • Translating plans into action

Score Guide:

ScoreDescription
9-10Excellent completion rate, consistently delivers
7-8Strong execution with minor gaps
5-6Moderate completion, some overdue tasks
3-4Struggles with completion
0-2Many incomplete or overdue tasks

Creativity

Ability to generate novel ideas and approach problems innovatively.

Behavioral Indicators:

  • Creating original content
  • Making cross-domain connections
  • Using creative tools (diagrams, brainstorming)
  • Diverse content types

Score Guide:

ScoreDescription
9-10Highly innovative, original approaches
7-8Creative with consistent new ideas
5-6Some creative output
3-4Mostly follows established patterns
0-2Limited creative activity

Communication

Ability to express ideas clearly and collaborate effectively.

Behavioral Indicators:

  • Clear documentation practices
  • Active in comments and discussions
  • Providing constructive feedback
  • Collaborative editing

Score Guide:

ScoreDescription
9-10Excellent communicator, highly collaborative
7-8Strong communication skills
5-6Adequate documentation and collaboration
3-4Minimal communication
0-2Limited collaborative activity

Leadership

Ability to guide initiatives and make decisions.

Behavioral Indicators:

  • Initiating new projects
  • Setting priorities and direction
  • Documenting decisions
  • Guiding team activities

Score Guide:

ScoreDescription
9-10Strong leader, initiates and guides effectively
7-8Takes initiative regularly
5-6Occasional leadership activity
3-4Follows more than leads
0-2Minimal initiative

Learning

Ability to acquire knowledge and develop skills.

Behavioral Indicators:

  • Exploring new topics
  • Using research features
  • Applying newly learned information
  • Expanding knowledge areas

Score Guide:

ScoreDescription
9-10Constant learner, rapid skill development
7-8Active learning and growth
5-6Moderate learning activity
3-4Limited exploration
0-2Minimal learning engagement

Resilience

Ability to persist through challenges and maintain momentum.

Behavioral Indicators:

  • Continuing despite obstacles
  • Re-engaging after gaps
  • Iterating on work
  • Long-term persistence on goals

Score Guide:

ScoreDescription
9-10Exceptional persistence, overcomes all obstacles
7-8Strong resilience with sustained effort
5-6Moderate persistence
3-4Gives up when facing challenges
0-2Limited persistence

How Scoring Works

Signal-to-Trait Mapping

Your activities generate signals that contribute to trait scores:

ActivityPrimary TraitsWeight
Create projectLeadership, CreativityHigh
Update projectExecution, FocusMedium
Create pageCreativity, ExecutionMedium
Add commentCommunication, LeadershipMedium
Agent chatLearning, CreativityLow
Research jobLearning, Focus, CreativityHigh
Complete taskExecution, ResilienceHigh

Behavioral Patterns

Four additional patterns influence your scores:

Completion Rate

completed_projects / total_projects

Boosts: Execution

Consistency Score

unique_active_days / date_range

Boosts: Focus, Resilience

Diversity Score

variety_of_activities / max_possible

Boosts: Creativity

Collaboration Score

communication_signals / total_signals

Boosts: Communication

Normalization

Raw scores are normalized using a logarithmic scale to prevent gaming:

normalized = (log(1 + raw) / log(1 + max_expected)) × 10

This provides diminishing returns for high activity volumes.

Profile Distillation

Requirements

For a full profile to be generated, you need:

RequirementMinimum
Behavioral signals12+
Time span30+ days
Active days10+ distinct days

For preferences and goals analysis, you need:

  • 5+ journal entries

Process

  1. Signal Collection - Gathers 90 days of activity (extends to 365 if needed)
  2. Signal Analysis - Maps activities to trait contributions
  3. Pattern Calculation - Computes behavioral patterns
  4. Trend Analysis - Compares to previous profile
  5. AI Enhancement - Qualitative analysis and recommendations
  6. Profile Generation - Creates/updates your profile page

Triggering Distillation

Profiles are automatically distilled:

  • When you access Memory Insights
  • On a scheduled basis (if configured)
  • Manually via the agent
You: Update my user profile

Agent: I'll distill your profile based on recent activity...

Trend Tracking

The system tracks how your traits change over time:

TrendIndicatorMeaning
Improving↑ (+0.5 or more)Score increased significantly
Stable→ (< ±0.5)Score remained steady
Declining↓ (-0.5 or more)Score decreased significantly

Trends help identify areas of growth and areas needing attention.

Viewing Your Profile

Memory Insights Page

Access your profile visualization:

  1. Open the sidebar
  2. Click Memory Insights (or access via agent settings)
  3. View the radar chart and trait breakdown

Profile Components

Radar Chart

  • Visual representation of all 7 traits
  • Scaled 0-10 on each axis
  • Shows your behavioral shape at a glance

Trait Scores

  • Individual score for each trait (0-10)
  • Trend indicator (↑ → ↓)
  • Confidence level (low/medium/high)

Strengths & Challenges

  • AI-identified strong areas
  • Areas with opportunity for growth
  • Based on trait scores and patterns

Evidence

  • Specific activities supporting each score
  • Links to related tasks, pages, projects

AI-Enhanced Insights

When available, the system uses AI to:

  • Analyze qualitative patterns beyond raw signals
  • Adjust scores by ±2 points based on evidence
  • Generate contextual insights
  • Provide personalized recommendations
  • Identify non-obvious patterns

Example Insights

Strengths:
- Strong execution with 85% task completion rate
- Consistent daily engagement over the past month
- Active documentation contributor

Challenges:
- Focus could improve - frequent context switching detected
- Limited creative exploration in recent weeks

Recommendations:
- Try time-blocking for deep work sessions
- Explore the research feature for learning opportunities
- Consider journaling to boost reflection

Privacy & Control

Enabling/Disabling

Control activity tracking in your preferences:

  1. Go to SettingsPreferences
  2. Toggle Enable Activity Tracking

When disabled:

  • No new activity signals are recorded
  • Existing profile remains but won't update
  • You can still use the agent normally

Workspace Settings

Admins can configure at the workspace level:

SettingDescription
enableMemoryInsightsEnable profile system workspace-wide
enableActivityTrackingTrack user activity sessions

Data Retention

  • Activity signals: Configurable retention period (default 90 days)
  • Profile snapshots: Kept for trend analysis
  • Raw signals can be deleted without affecting profile

Best Practices

Improving Your Profile

  1. Consistency - Regular activity beats sporadic bursts
  2. Completion - Finish what you start
  3. Documentation - Write clear notes and comments
  4. Exploration - Try new features and topics
  5. Reflection - Use journaling for self-awareness

Understanding Scores

  • Low scores aren't bad - They indicate areas with less activity, not failure
  • Context matters - A researcher may have high Learning, low Leadership
  • Trends matter more - Growth trajectory is more meaningful than absolute scores
  • It's personal - Compare to yourself, not others

API Access

Distill Profile

await client.agent.memory.distillProfile({
workspaceId: 'ws_123',
userId: 'user_456', // optional, defaults to current user
spaceId: 'space_789', // optional, scope to specific space
});

Query Profile

const profile = await client.agent.memory.query({
workspaceId: 'ws_123',
tags: ['user-profile', 'user:user_456'],
limit: 1,
});