> For the complete documentation index, see [llms.txt](https://docs.pharos.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.pharos.ai/platform-overview/pharos-platform-architecture/ai-analysis/org-genie/org-level.md).

# Org Level

Use Org Genie at the org level to ask broad questions about your org's observability data, spot trends, and prioritize which Issues need attention first.

Access Org Genie from the main navigation to ask broad questions about your entire org. Org-level queries allow you to explore patterns, trends, and insights across all your observability data—perfect for health checks, weekly reviews, and understanding what's happening across your Salesforce environment.

{% hint style="success" %}
**Big Picture Intelligence**

Org-level Org Genie provides the 30,000-foot view of your org's health, helping you identify which Issues deserve attention, spot emerging trends, and understand overall error patterns without building dashboards.
{% endhint %}

***

## Overview

Org-level Org Genie provides a conversational interface for analyzing your entire Salesforce org's logs, Issues, and metadata changes. Unlike issue-level context (which focuses on a single Issue), org-level context enables exploration across all your observability data simultaneously.

### Key Capabilities

Org-level queries support a range of exploratory tasks across your entire org's data.

<table><thead><tr><th width="200">Capability</th><th>What You Can Do</th></tr></thead><tbody><tr><td><strong>📊 Trend Analysis</strong></td><td>Compare time periods, identify spikes and anomalies, understand if things are getting better or worse</td></tr><tr><td><strong>🎯 Prioritization</strong></td><td>Discover which Issues warrant immediate attention based on severity, impact, and frequency</td></tr><tr><td><strong>📈 Visualization</strong></td><td>Generate charts on demand without building dashboards—see patterns emerge visually</td></tr><tr><td><strong>🔍 Pattern Recognition</strong></td><td>Identify outliers, anomalies, and emerging Issues before they become critical</td></tr><tr><td><strong>⚙️ Metadata Queries</strong></td><td>Ask about configuration changes across your org: "What changed last week?"</td></tr><tr><td><strong>👥 User Impact</strong></td><td>Understand which users are most affected by errors and why</td></tr></tbody></table>

***

## Accessing Org-Level Org Genie

### How to Access

Navigate to **Org Genie** from the main navigation menu. This opens the org-level context where you can ask questions about any data in your org.

{% hint style="info" %}
**Per-User Chat History**

Each user has their own org-level chat history, allowing you to maintain multiple conversation threads for different investigations or topics. Your conversations are private to you.
{% endhint %}

***

## Example Queries

Org-level Org Genie excels at broad, exploratory questions that span your entire org's observability data.

### Error Analysis and Trends

These queries help you understand how error volumes are changing over time.

{% tabs %}
{% tab title="📊 Comparing Time Periods" %}
**Questions to Ask:**

* *"Can you tell me how many errors I've had this week versus last week?"*
* *"Show me a chart of errors over the last month broken down by category"*
* *"Were there any anomalies in the errors this week versus last week?"*
* *"How do this month's errors compare to last month?"*

**What You Get:**

* Quantitative comparison (e.g., "243 errors this week vs 412 last week - 41% decrease")
* Visual charts showing trends and breakdowns
* Identification of anomalies or unexpected patterns
* Context about whether the trend is positive or negative

**Example Response:**

```
Error Comparison: This Week vs Last Week

This Week (Jan 22-28): 243 errors
Last Week (Jan 15-21): 412 errors
Change: ↓ 169 errors (41% decrease)

Breakdown by Category:
• Apex Errors: 87 (↓ from 156) - 44% decrease
• Flow Errors: 92 (↓ from 178) - 48% decrease  
• Integration Errors: 64 (↓ from 78) - 18% decrease

Notable: Large drop in Flow errors correlates with 
deployment on Jan 17 that fixed null handling issues.

[Chart showing side-by-side comparison]
```

{% endtab %}

{% tab title="🔍 Identifying Anomalies" %}
**Questions to Ask:**

* *"Were there any unusual error patterns this week?"*
* *"Show me any spikes in error rates over the last month"*
* *"Are there any errors occurring at unusual times?"*
* *"Did anything abnormal happen today?"*

**What You Get:**

* Identification of spikes, drops, or irregular patterns
* Timing and severity of anomalies
* Potential causes or correlations
* Recommendations for investigation

**Example Response:**

```
Anomaly Detection: Last 30 Days

✅ Normal Pattern: 50-80 errors per day (weekdays)
✅ Normal Pattern: 10-20 errors per day (weekends)

⚠️ ANOMALY DETECTED:
January 15: 347 errors (4.3x normal)
- Spike started at 2:15 PM
- Primarily Flow errors (234 out of 347)
- Correlated with deployment at 2:00 PM

⚠️ ANOMALY DETECTED:
January 23: 8 errors (unusually LOW)
- Holiday (org had minimal activity)
- Not a concern

Recommendation: Investigate Jan 15 deployment impact
```

{% endtab %}

{% tab title="📈 Long-Term Trends" %}
**Questions to Ask:**

* *"Give me a line chart of all errors over the last quarter broken down by category"*
* *"Show me error trends over the last 6 months"*
* *"Are errors increasing or decreasing over time?"*
* *"What's the overall trend for Integration errors?"*

**What You Get:**

* Long-term visualization showing trends
* Assessment of whether things are improving or degrading
* Identification of seasonal patterns or cycles
* Context for whether current error rates are normal

**Example Response:**

```
Error Trends: Last Quarter (Oct-Dec 2023)

Overall Trend: ↓ IMPROVING (23% decrease)

October: 2,847 errors
November: 2,341 errors (↓ 18%)
December: 2,198 errors (↓ 6%)

Category Breakdown:
• Apex: Stable (~800/month)
• Flow: ↓ IMPROVING (down 45% since October fix)
• Integration: ↑ WORSENING (up 30% - investigate!)

Notable Pattern: Weekly spikes every Thursday 
(correlates with batch job execution schedule)

[Line chart showing trends by category]

Recommendation: Focus on Integration errors—
they're trending in wrong direction.
```

{% endtab %}
{% endtabs %}

### Metadata Changes and Correlation

These queries connect configuration changes to the error patterns they may have caused.

{% tabs %}
{% tab title="⚙️ Recent Changes" %}
**Questions to Ask:**

* *"What's changed in my org over the last 5 days?"*
* *"Show me all Flow changes from last week"*
* *"Were there any Apex deployments in the last month?"*
* *"What validation rules were modified recently?"*

**What You Get:**

* Summary of metadata changes by type
* Timing of changes (when and who)
* High-level description of what changed
* Potential impact assessment

**Example Response:**

```
Metadata Changes: Last 5 Days (Jan 24-28)

Flows (7 changes):
• Case_Assignment_Flow_v3 (Jan 25) - Added null checks
• Lead_Routing_Flow (Jan 26) - Updated queue assignments
• Opportunity_Validation_Flow (Jan 27) - New validation added
• [4 more...]

Apex Classes (3 changes):
• AccountTriggerHandler (Jan 24) - Bug fix for bulk processing
• OpportunityService (Jan 26) - Added logging
• IntegrationHelper (Jan 28) - Timeout adjustment

Validation Rules (2 changes):
• Opportunity_Stage_Required (Jan 25) - Criteria updated
• Account_Industry_Check (Jan 27) - New rule added

Impact Assessment:
⚠️ Opportunity_Stage_Required change correlates with 
Issue #1847 spike—investigate connection
```

{% endtab %}

{% tab title="🔗 Change Correlation" %}
**Questions to Ask:**

* *"Were there any changes around December 15th that could explain the error spike?"*
* *"What changed between last week and this week?"*
* *"Could any recent changes have caused the increase in Integration errors?"*
* *"Show me deployments that happened right before error spikes"*

**What You Get:**

* Metadata changes correlated with error patterns
* Timing analysis (did changes precede errors?)
* Assessment of likelihood that changes caused Issues
* Recommendations for further investigation
  {% endtab %}
  {% endtabs %}

### Issue Prioritization

These queries help you decide which Issues to tackle first.

{% tabs %}
{% tab title="🎯 What Needs Attention" %}
**Questions to Ask:**

* *"Which Issues from the last 30 days warrant immediate attention and why?"*
* *"What are the top 5 most impactful Issues this month?"*
* *"Show me critical Issues that are unresolved"*
* *"Which Issues should I tackle first?"*

**What You Get:**

* Prioritized list of Issues with reasoning
* Severity and impact assessment for each
* Links to specific Issues for investigation
* Recommendations for which to address immediately vs later

**Example Response:**

```
Priority Issues: Last 30 Days

🚨 CRITICAL - Immediate Action Required:

1. Issue #1847: Flow Assignment Failures [Link]
   - 234 occurrences, 47 users affected
   - Blocking case routing (business-critical)
   - Started Jan 15 after deployment
   - PRIORITY: Urgent (fix today)

2. Issue #1892: Integration Timeout with ERP [Link]
   - 156 occurrences, 3x increase from baseline
   - Blocking order creation (revenue impact)
   - No metadata changes detected (external cause?)
   - PRIORITY: Urgent (investigate immediately)

⚠️ HIGH - Address This Week:

3. Issue #1901: Batch Job Failures [Link]
   - Daily batch failing since Jan 20
   - Affects data quality but no user-facing impact
   - PRIORITY: High (fix this sprint)

4 more Issues... [Show All]
```

{% endtab %}

{% tab title="📊 Impact Analysis" %}
**Questions to Ask:**

* *"Which Issues are affecting the most users?"*
* *"Show me Issues with increasing frequency"*
* *"What Issues have the highest occurrence count?"*
* *"Which functional areas have the most problems?"*

**What You Get:**

* Issues ranked by user impact, frequency, or occurrence count
* Analysis of which areas of your org are most problematic
* Identification of Issues that are getting worse
* Context for prioritization decisions
  {% endtab %}
  {% endtabs %}

### User Impact Analysis

These queries reveal which users and profiles are most affected by errors.

{% tabs %}
{% tab title="👥 Affected Users" %}
**Questions to Ask:**

* *"Show me the top 10 impacted users (not system or integration users) over the last 6 months"*
* *"Which users have experienced the most errors in the last week?"*
* *"Are errors concentrated in specific user profiles or roles?"*
* *"Which teams are most affected by errors?"*

**What You Get:**

* List of users ranked by error impact
* Filtering to exclude system/integration users
* Analysis of whether errors affect specific profiles or roles
* Context for whether problems are org-wide or localized

**Example Response:**

```
Top Impacted Users: Last 6 Months
(Excluding system and integration users)

1. Sarah Johnson (Sales Rep) - 89 errors
   - Primarily Opportunity-related Issues
   - Most frequent: Issue #1756 (Case Assignment failures)
   
2. Mike Chen (Support Agent) - 76 errors
   - Case routing and Flow errors
   - Concentrated in last 2 weeks (recent Issue)

3. Emily Rodriguez (Account Manager) - 64 errors
   - Integration timeouts affecting order sync
   - Stable pattern (ongoing Issue)

[7 more users...]

Pattern Analysis:
⚠️ Sales Reps experiencing 2.3x more errors than other profiles
   - Investigate Sales-specific automation and validation rules

[Bar chart showing top 10 users]
```

{% endtab %}
{% endtabs %}

***

## Example Workflow: Weekly Health Check

Here's a typical workflow using org-level Org Genie for weekly org health reviews:

{% tabs %}
{% tab title="1️⃣ Start Broad" %}
**You ask:**

*"Can you tell me how many errors I've had this week versus last week?"*

**Org Genie Response:**

```
Error Comparison:

This Week: 243 errors
Last Week: 412 errors
Change: ↓ 169 errors (41% decrease)

Overall: ✅ IMPROVING

Breakdown:
• Apex: 87 (↓ from 156)
• Flow: 92 (↓ from 178) 
• Integration: 64 (↓ from 78)

Most significant improvement in Flow errors.
```

**What you learned:**\
Overall trend is positive—41% decrease is substantial.
{% endtab %}

{% tab title="2️⃣ Visualize the Pattern" %}
**You ask:**

*"Show me a chart comparing errors this week versus last week broken down by category"*

**Org Genie Response:**

```
[Bar chart generated showing side-by-side comparison]

Chart shows:
- All three categories decreased
- Flow errors had largest absolute decrease (86 fewer)
- Proportionally, Flow decreased most (48%)
```

**What you learned:**\
Visualization confirms Flow errors drove the improvement.
{% endtab %}

{% tab title="3️⃣ Investigate the Cause" %}
**You ask:**

*"I'm seeing a large drop in Flow errors this week versus last. Were there any metadata changes that could've caused this improvement?"*

**Org Genie Response:**

```
Analysis of metadata changes:

✅ Likely Cause of Improvement:

Flow "Case_Assignment_Flow_v3" fixed on Jan 17
- Added null handling for Contact.Department
- This fixed Issue #1756 (was causing 40-60 errors/day)
- Timing: Fix deployed Jan 17, errors dropped Jan 18

No other significant Flow changes detected.

Impact: Issue #1756 represented ~70% of Flow errors
Fixing it caused the observed decrease.
```

**What you learned:**\
The improvement was due to a specific fix, not general trend.
{% endtab %}

{% tab title="4️⃣ Expand Context" %}
**You ask:**

*"Can you give me a line chart of all errors over the last month broken down by category?"*

**Org Genie Response:**

```
[Line chart generated showing 30-day trend]

Chart reveals:
- Flow errors: Sharp drop on Jan 17 (from ~60/day to ~15/day)
- Apex errors: Stable around 15-20/day (no major changes)
- Integration errors: Gradual increase over month (↑30%)

⚠️ Concerning Trend: Integration errors increasing
```

**What you learned:**\
While Flow errors improved, Integration errors are getting worse—needs investigation.
{% endtab %}

{% tab title="5️⃣ Prioritize Action Items" %}
**You ask:**

*"Which Issues from the last 30 days warrant immediate attention and why?"*

**Org Genie Response:**

```
Priority Issues for Action:

🚨 CRITICAL:
1. Issue #1892: ERP Integration Timeouts [Link]
   - 156 occurrences (3x increase)
   - Revenue impact (blocks order creation)
   - INVESTIGATE: External system issue vs config

⚠️ HIGH:
2. Issue #1901: Batch Job "Account Cleanup" Failures [Link]
   - Daily failures since Jan 20
   - Data quality impact
   - FIX: This sprint

3. Issue #1905: Validation Rule Blocking Opportunities [Link]
   - 47 occurrences (new Issue, started Jan 25)
   - Blocking sales team
   - FIX: Adjust validation criteria

Recommendation: Focus on Issue #1892 first (revenue impact)
```

**What you learned:**\
Clear action items with prioritization reasoning—ready for sprint planning.
{% endtab %}

{% tab title="6️⃣ Deep Dive on Priority Issue" %}
**You click through to Issue #1892:**

From the Issue detail page, you open **issue-level Org Genie** and continue:

*"What could be causing these Integration timeouts?"*

\[Conversation continues in issue-level context with focused analysis]

**What happened:**\
Seamlessly transitioned from org-level discovery to issue-level investigation.
{% endtab %}
{% endtabs %}

{% hint style="success" %}
**Iterative Exploration**

Notice the natural flow: Start broad → Visualize → Investigate patterns → Expand context → Prioritize → Deep-dive. Each question builds on the previous response, creating a conversation that guides your investigation.
{% endhint %}

***

## Chat Management

Org-level Org Genie maintains your personal chat history, allowing you to create and manage multiple conversation threads.

### Creating and Organizing Chats

You manage org-level chats through a few common actions.

<table><thead><tr><th width="200">Action</th><th>When to Use</th><th>Benefit</th></tr></thead><tbody><tr><td><strong>📝 Create New Chat</strong></td><td>Starting a completely new investigation or topic</td><td>Keeps contexts separate, prevents conversation mixing</td></tr><tr><td><strong>💬 Continue Existing Chat</strong></td><td>Following up on previous investigation or adding related questions</td><td>Maintains context from previous conversation</td></tr><tr><td><strong>🗑️ Delete Old Chats</strong></td><td>Cleaning up completed investigations</td><td>Keeps workspace organized and relevant</td></tr><tr><td><strong>🏷️ Name Chats (if supported)</strong></td><td>Making chats easy to find later</td><td>Quick reference for returning to specific investigations</td></tr></tbody></table>

### Best Practices for Chat Organization

Organizing your chats keeps investigations easy to follow and easy to return to.

{% tabs %}
{% tab title="📁 Topic Separation" %}
**Create separate chats for:**

* Weekly health checks (one per week)
* Error spike investigations (one per incident)
* Sprint planning research
* Executive reporting preparation
* Exploratory analysis sessions

**Don't mix:**

* Different time periods in the same chat
* Unrelated investigations
* Different stakeholder audiences

**Why it matters:**\
Clear separation prevents context confusion and makes it easy to return to specific investigations later.
{% endtab %}

{% tab title="🔄 Continuity" %}
**Keep related questions together:**

If you're investigating an error spike over multiple days:

* Day 1: "Show me errors from last week"
* Day 2: "Has the pattern changed since we last talked?"
* Day 3: "Were there any new metadata changes that could help?"

All in the same chat to maintain context.

**Why it matters:**\
Org Genie remembers previous responses within a chat, making follow-ups more efficient.
{% endtab %}

{% tab title="🧹 Regular Cleanup" %}
**Chats don't auto-delete:**

Schedule regular cleanup (monthly or quarterly):

* Delete chats from completed investigations
* Archive chat contents if needed for documentation
* Keep only active or reference conversations

**Why it matters:**\
Too many old chats clutter your workspace and make it hard to find relevant conversations.
{% endtab %}
{% endtabs %}

{% hint style="info" %}
**Your History is Private**

Org-level chat history is per-user. Other users in your org don't see your conversations, and you don't see theirs. Each user has their own private workspace for org-level investigations.
{% endhint %}

***

## Best Practices for Org-Level Queries

### Effective Querying Strategies

A few strategies make org-level queries more productive.

{% tabs %}
{% tab title="🌐 Start High-Level" %}
**Recommended Approach:**

Begin by understanding overall trends before drilling into specifics.

**Good First Questions:**

* *"Show me error trends over the last month"*
* *"What's the overall health of my org this week?"*
* *"Are errors increasing or decreasing?"*

**Why it works:**\
You avoid premature focus on specific Issues; let data guide where to investigate next.
{% endtab %}

{% tab title="⏱️ Be Specific About Time" %}
**Good:**

* *"Last 7 days"*
* *"This quarter"*
* *"January 15 through January 20"*
* *"Last month versus this month"*

**Less Effective:**

* *"Recently"*
* *"A while back"*
* *"Last few weeks or so"*

**Why it matters:**\
Precise time ranges produce more accurate comparisons and correlations.
{% endtab %}

{% tab title="📊 Request Visualizations" %}
**When to ask for charts:**

* Comparing time periods
* Looking for trends over time
* Identifying patterns in categories
* Presenting findings to stakeholders

**How to ask:**

* *"Show me a chart of..."*
* *"Visualize the trend for..."*
* *"Can you plot that as a bar chart?"*

**Why it matters:**\
Patterns and anomalies are easier to spot visually than in tables or text.
{% endtab %}

{% tab title="🔍 Follow the Data" %}
**Let responses guide next questions:**

If you see an interesting pattern (spike, anomaly, trend), ask about it:

* *"What caused that spike on January 15?"*
* *"Why are Integration errors increasing?"*
* *"Were there any changes around that date?"*

**Why it matters:**\
Data-driven investigation is more effective than starting with assumptions.
{% endtab %}

{% tab title="🚫 Exclude Noise" %}
**Filter out irrelevant data:**

* *"Not including system users"*
* *"Exclude test accounts"*
* *"Only production errors"*
* *"External users only"*

**Why it matters:**\
Focuses analysis on what matters for your business, not system overhead.
{% endtab %}

{% tab title="🔗 Correlate Events" %}
**Connect observability with changes:**

* *"Were there any changes around December 15th that could explain this spike?"*
* *"Did any deployments happen before the error increase?"*
* *"Show me metadata changes correlated with Issue trends"*

**Why it matters:**\
Understanding cause-and-effect relationships is key to preventing future Issues.
{% endtab %}
{% endtabs %}

### Workflow Patterns

Several common workflows map naturally to org-level queries.

**Weekly Health Check:**

1. Compare this week to last week
2. Visualize trends
3. Identify anomalies
4. Prioritize Issues
5. Deep-dive on top priorities

**Sprint Planning:**

1. Review Issues from last sprint
2. Identify unresolved high-priority Issues
3. Estimate impact of Issues
4. Export findings for planning meeting

**Executive Reporting:**

1. Generate long-term trend visualizations
2. Quantify improvements or degradations
3. Identify top impacted users or areas
4. Export charts for presentation

**Incident Investigation:**

1. Identify when spike started
2. Correlate with recent changes
3. Assess scope and impact
4. Drill into specific Issues
5. Determine root cause

***

## Navigating to Issue-Level Context

When Org Genie identifies specific Issues in its responses, you can seamlessly transition to issue-level investigation.

### Workflow

Moving from org-level discovery to issue-level investigation follows a simple path.

1. **Org-Level Discovery**\
   Ask: *"Which Issues warrant immediate attention?"*
2. **Org Genie Returns Results**

   ```
   Found 3 critical Issues:
   1. [Link] Issue #1847: Flow failures affecting 47 users
   2. [Link] Issue #1892: Integration timeouts (3x increase)
   3. [Link] Issue #1901: Batch job failures
   ```
3. **Click Issue Link**\
   Navigate to Issue #1847 detail page
4. **Open Issue-Level Org Genie**\
   From the Issue detail page, open Org Genie
5. **Continue Investigation**\
   Ask: *"What changed that could have caused this Issue?"*

{% hint style="success" %}
**Seamless Transition**

Moving from org-level discovery to issue-level deep-dive is natural and preserves context. Each maintains its own chat history, so you can return to org-level later without losing your place.

→ [**Learn more about Issue-Level Queries**](https://docs.pharos.ai/platform-overview/pharos-platform-architecture/ai-analysis/org-genie/issue-level.md)
{% endhint %}

***

## Limitations

Understanding org-level query limitations helps set appropriate expectations.

### Scope Constraints

Org-level queries cover observability data and tracked metadata, not your business records.

**Can analyze:**

* ✅ All Pharos logs and Issues across your org
* ✅ Aggregated patterns and trends
* ✅ Metadata changes from Change Tracking
* ✅ Comparative analysis across time periods

**Cannot analyze:**

* ❌ Standard Salesforce objects (Accounts, Contacts, etc.)
* ❌ Business data unrelated to observability
* ❌ Real-time Salesforce configuration (only historical metadata)
* ❌ External systems outside Salesforce

### Performance Considerations

Query speed depends on the time range and the complexity of the analysis.

**Fast queries:**

* Last 7-30 days with standard aggregations
* Simple comparisons and visualizations
* Questions targeting specific time periods

**Slower queries:**

* Very large time ranges (e.g., "last year")
* Complex multi-dimensional analysis
* Queries requiring correlation across many data sources

**Response Time:** Most queries complete in 10-15 seconds; complex queries may take 20-30 seconds.

### Data Requirements

**Prerequisites for full functionality:**

* Logs and Issues must exist (Pharos must be collecting data)
* Historical data must accumulate over time
* Change Tracking must be enabled for metadata correlation
* Sufficient data volume for pattern recognition

**Expected limitations:**

* Recently-installed Pharos instances have limited historical context
* Change Tracking must run over time to build metadata history
* Very quiet orgs may not have enough data for meaningful trend analysis

***

{% hint style="success" %}
**Start Your First Org-Level Investigation**

1. Navigate to Org Genie from the main menu
2. Ask a simple comparison: *"How many errors did I have this week versus last week?"*
3. Request a visualization: *"Show me a chart of error trends over the last month"*
4. Investigate a pattern: *"Were there any anomalies or spikes?"*
5. Correlate with changes: *"What metadata changes happened around that spike?"*
6. Prioritize: *"Which Issues from this period warrant immediate attention?"*
7. Click through to a specific Issue and continue with issue-level Org Genie

Experience how conversational queries transform org health monitoring from a manual dashboard-building exercise into a natural investigation flow.
{% endhint %}
