Overview
AI research helps you uncover useful insights about people and companies without leaving Apollo. Use it to qualify prospects, understand account fit, personalize outreach, and enrich records with details that support your go-to-market strategy.
Start with a research template built by Apollo, or create your own custom prompt for a specific use case. Apollo generates research based on the context and data you provide, then saves the output to an AI research field that you can review and reuse across Apollo.
Check out the following sections to learn more about AI research.
How It Works
AI research follows a simple request-to-reuse model:
- Choose what to research. Run AI research on people or companies from prospecting searches, saved lists, or supported Apollo workflows.
- Define what you want to learn. Start with an Apollo research template for a common use case, or create a custom prompt tailored to your business, ICP, product, or outreach strategy. Custom prompts can reference record data and other AI research fields with dynamic variables.
- Provide context and choose a model. Apollo combines your selected AI context, available Apollo data, information referenced in your prompt, and the AI model you select. Depending on the model and prompt, AI research can also reference information from the web.
- Preview and run the research. Preview results before running AI research at scale, then save and run the research when the output matches your goal. Apollo saves each result to an AI research field on the person or company record.
- Reuse the results. Review, edit, or regenerate research, filter prospects by the results, reference the field in messaging, use it in another AI research prompt, or incorporate it into an automated workflow.
Additionally, when you create custom AI research, you can make the resulting field private or global. Global AI research fields are visible to everyone at your organization and require permission to add or edit custom field definitions.
Supported AI models and their capabilities can change. Review the available models.
What You Can Do
Use AI research throughout your prospecting and outreach process based on what you want to accomplish.
| Your goal... | Try this |
|---|---|
| Research a defined group of prospects | Use AI research on a saved list to uncover insights across a group of people or companies you want to evaluate together. |
| Refine your targeting | Filter by AI research to narrow your prospecting results based on insights such as qualification, company characteristics, or your own custom research criteria. |
| Personalize your outreach | Use AI research in messaging to add relevant prospect or account insights to emails and sequences. |
| Automate research at scale | Use AI research in workflows to research people or companies automatically as they move through your workflow. |
Get Started
Set up your context, test AI research on a small sample, then refine your prompts before you run research at scale.
| Configure AI context | Add information about your company, products, services, value propositions, and other business context to the AI context center so Apollo can use it to make your research more relevant. |
|---|---|
| Run AI research | Start with an Apollo research template or create custom AI research. Then preview the output, and save and run the research on your prospects. |
| Improve your prompts | Learn the basics of AI prompting like how to provide specific instructions, structure your output, use dynamic variables, and add fallbacks so your custom research returns useful results. |
| Review credit usage | Review your AI research credit usage to understand which actions consume credits in your workflow, and monitor usage across your team. |
FAQs
|
Frequently asked questions |
||
| Which model should I use for my AI research? | How do I review AI research for a specific prospect? | What happens to AI research after a run? |
| When does AI research use credits? | How can I monitor and control AI research credit usage across my team? | Should I use an AI research template or create a custom prompt? |
Which model should I use for my AI research?
Choose a model based on the type of research you want Apollo to perform:
| Model | Best for... |
|---|---|
| Perplexity Agent | Researching contact or account details using AI with web access. Perplexity may miss pay-walled sites. |
| OpenAI GPT-4o mini | Targeting, categorizing, and summarizing Apollo data without web access. |
| OpenAI GPT-5 nano | Transforming and analyzing Apollo data with advanced reasoning and improved accuracy. |
| Anthropic Claude Haiku 4.5 | Qualifying leads and creating personalized outreach at scale. |
| Anthropic Claude Sonnet 4.6 | Analyzing customer data and creating targeted sales strategies. |
Preview your AI research on a few records before you run it at scale. If the results don't match your goal, try a different model, field type, or prompt.
Should I use an AI research template or create a custom prompt?
Use an AI research template when your goal matches a common prospecting or research use case and you want to get started quickly.
Run custom AI research when you need research tailored to your business, ICP, product, campaign or outreach strategy, or when you want more control over the criteria and output.
When you create your own prompt, follow AI prompt best practices to give Apollo clear instructions, reference data with dynamic variables, and define the type of output you want.
How do I review AI research for a specific prospect?
After you run AI research, Apollo saves the result to an AI research field on the contact or account. To review the research, open the contact or account, click All fields then find AI custom fields. From there, you can review, edit, or regenerate the AI research result.
If you import via CSV or CRM, manually create a contact or account, or manually delete individual fields on a contact or account, some AI research won't generate.
What happens to AI research after a run?
Apollo saves each result to an AI research field on the person or company you researched. You can review or update the result, filter prospects with AI research, use the field to personalize messaging, or reference it in another AI research prompt.
When you create custom AI research, you can make the field private to you or global for your organization. To create a global AI research field, you need permission to add or edit custom field definitions.
When does AI research use credits?
Each AI research run costs one credit. Apollo counts one AI research run for each record and each column you enrich. If your AI research generates multiple columns, each column counts as a separate run for every contact or account you research.
For example, enriching ten contacts with one column each uses ten AI research runs, or ten credits.
Enriching the same ten contacts with three columns each uses thirty AI research runs, or thirty credits.
AI research can also use credits when it runs automatically as part of a workflow. You can review your credit usage to see which AI research actions consumed credits.
AI research may not run if your team doesn't have enough credits available.
How can I monitor and control AI research credit usage across my team?
Estimate your AI research usage before you run it by multiplying the number of records you want to research by the number of columns the research creates. Each run costs one credit, so researching 100 contacts across one column uses 100 credits. To reduce credit usage, research fewer records or reduce the number of columns you create.
Before running AI research at scale, preview the results to confirm that your prompt, model, and output match your goal. Apollo provides free previews for each AI research prompt you create. You can also review your credit usage to see which AI research actions consumed credits.
You can review credit usage to see which AI research actions are consuming credits across your team. You can also set credit limits per user to establish guardrails that match your organization's credit budget. If your team uses AI research in workflows, you can set processing limits to control how many records and credits the workflow can process.
Learn More
Go beyond the basics with guided examples and training for using AI research in your prospecting workflow.
| How to Excel at AI Research in Three Steps | Follow an end-to-end example to research prospects, refine your targeting, and turn your research into personalized outreach. |
| How to Use AI Research to Accelerate Prospecting | Take a guided Apollo Academy course on using AI research to improve targeting, research leads, and personalize messaging. |
| Apollo AI Glossary | Learn the terminology behind AI research, AI filters, AI SDRs, and other Apollo AI features. |