AI Research and Data Enrichment in Account Studio

AI Research in Account Studio allows you to enrich your account data with insights generated by artificial intelligence. This powerful feature helps you add valuable data points that can be used for filtering, personalization, and asset targeting.

What is AI Research?

AI Research uses artificial intelligence to analyze publicly available information about your accounts and generate custom data points based on your specific questions. This allows you to:

• Add data points that don't exist in your CRM or other integrations
• Answer specific questions about your accounts at scale
• Create custom fields for filtering and personalization
• Generate insights for time-sensitive information

Adding AI Research to Your List

Step-by-Step Process

1. Select Your List: Start with an existing list or create a new one in Account Studio
2. Add AI Research Column: Click "Add AI Research"
3. Configure Your Prompt: Write a clear, specific prompt for the data you want
4. Preview Results: Review the first 5 accounts to ensure quality
5. Run for All Accounts: Execute the research across your entire list

Writing Effective Prompts

The quality of your AI research results depends heavily on how you write your prompts.

Best Practices for Prompts:

Be Specific and Measurable

• Good: "Have they raised funding in the last 6 months?"
• Poor: "Are they growing?"

Use True/False Questions When Possible

• Good: "Do they have a remote work policy?"
• Poor: "What's their work environment like?"

Include Timeframes When Relevant

• Good: "Have they launched new products in 2025?"
• Poor: "Do they launch new products?"

Ask for Specific Information

• Good: "What industry vertical do they primarily serve?"
• Poor: "Tell me about their business"

Example Prompts:

Funding and Growth

• "Have they raised funding in the last 6 months?"
• "Are they in a growth stage or mature stage?"
• "What was their most recent funding round size?"

Technology and Tools

• "Do they use Salesforce as their CRM?"
• "Are they primarily a SaaS company?"
• "What marketing automation platform do they use?"

Market and Competition

• "Who are their main competitors?"
• "What market segment do they serve (SMB, Mid-Market, Enterprise)?"
• "Are they B2B or B2C focused?"

Understanding AI Research Results

What You'll See

Each AI research result includes:

• The Answer: The generated data point
• Rationale: Explanation of why this answer was chosen
• Sources: References used to generate the answer (when available)

Interpreting Results

High-Quality Results: Clear answers with specific rationale and sources Medium-Quality Results: Answers with general reasoning but limited sources Low-Quality Results: Vague answers or "unknown" responses

Note: AI research quality depends on the availability of public information about each account.

Managing AI Research

Handling Incomplete Results

If some accounts don't have AI research data, you have several options:

Run for Missing Accounts

• When to Use: When you've added new accounts to your list
• How: Click "Run for Missing Accounts" in the AI research column
• Result: Only empty cells will be processed

Retry Failures

• When to Use: When some accounts failed to process
• How: Click "Retry Failures" option
• Result: Failed attempts will be re-processed

Updating AI Research

Edit Prompts

• Modify Research Question: Change your prompt to get different insights
• Refine Criteria: Make prompts more specific or broader as needed
• Add Context: Include additional context for better results

Rerun Research

• Update Existing Data: Run updated prompts across all accounts
• Refresh Time-Sensitive Data: Manually update time-based research
• Improve Quality: Rerun with better prompts for higher quality results

Time-Sensitive Research

Some AI research is time-sensitive (e.g., "raised funding in the last 6 months"). Currently, you need to manually rerun this research to keep it current.

Best Practices for Time-Sensitive Research:

• Note the date you ran the research in your prompt or list name
• Set calendar reminders to refresh time-sensitive data
• Consider the refresh frequency based on your use case

AI Research Performance

Processing Speed

AI research processes accounts in batches and may take time for large lists. Current performance factors:

• Batch Processing: Accounts are processed in groups
• Rate Limits: Processing speed may be limited during high usage
• Model Upgrades: Performance improvements are ongoing

Preview Mode

Always use preview mode before running AI research on large lists:

• Preview First 5: See sample results before full execution
• Quality Check: Ensure prompts are working as expected
• Refine Prompts: Make adjustments based on preview results

Using AI Research for you Assets

Filtering with AI Research

Once you've added AI research to your list, you can use it for filtering:

• True/False Fields: Filter for "Yes" or "No" responses
• Text Fields: Filter by specific text values
• Exists/Does Not Exist: Filter accounts that have research data

Personalization with AI Research

AI research data can be used in asset personalization:

• Dynamic Content: Reference AI research fields in your assets
• Targeted Messaging: Use insights to customize your messaging
• Audience Segmentation: Create segments based on AI research results

Advanced AI Research Strategies

Layered Research

Build comprehensive account profiles by adding multiple AI research columns:

1. Basic Information: Industry, size, business model
2. Technology Stack: Tools and platforms they use
3. Market Position: Competitors, market segment
4. Recent Activity: Funding, product launches, hiring

Research for Filtering vs. Personalization

For Filtering: Use true/false or categorical questions

• "Are they in a growth stage?" (Yes/No)
• "What market segment do they serve?" (SMB/Mid-Market/Enterprise)

For Personalization: Use more detailed, narrative responses

• "What are their main business challenges?"
• "What makes them different from competitors?"

Quality Optimization

Improve Research Quality:

• Use specific, focused prompts
• Include relevant context in your questions
• Test prompts with preview mode
• Refine based on results

Monitor Research Performance:

• Review rationale and sources for accuracy
• Check for consistent "unknown" responses
• Adjust prompts based on result quality

Troubleshooting AI Research

Common Issues

Incomplete Results

• Some accounts may not have enough public information
• Try broader or more general prompts
• Use "Run for Missing Accounts" to fill gaps

Inconsistent Results

• Review prompt specificity
• Check if timeframes are too narrow
• Consider the availability of public information

Getting Better Results

Prompt Optimization:

• Start with broad questions and narrow down
• Use industry-specific terminology when relevant
• Include examples in your prompts when helpful

Data Quality:

• Focus on publicly available information
• Avoid questions requiring internal company knowledge
• Use multiple research columns for comprehensive insights

Best Practices Summary

Before Running AI Research:

1. Test with Preview: Always preview results first
2. Refine Prompts: Make sure questions are clear and specific
3. Consider Timeframes: Include relevant time periods
4. Plan for Updates: Consider how often you'll need to refresh data

During AI Research:

1. Monitor Progress: Check processing status
2. Review Quality: Look at rationale and sources
3. Handle Failures: Use retry options for failed accounts

After AI Research:

1. Validate Results: Spot-check accuracy where possible
2. Use for Filtering: Apply research data to refine your lists
3. Plan Maintenance: Schedule updates for time-sensitive data
4. Document Process: Keep notes on effective prompts

For more information on using your enriched lists for assets, see our asset personalization documentation