Your intent data provider tells you “Acme Corp is researching CRM tools.” Great. Acme has 5,000 employees. Who exactly should your SDR call?
The VP of Sales who’s leading the evaluation? The IT Director who’ll own the integration? The marketing intern who Googled “best CRM” for a college project?
You don’t know. Your intent data doesn’t tell you. And so your rep does what every rep does - sprays the org chart with generic outreach, hoping one of those emails lands on the right desk.
Company-level intent is a starting point. Person-level intent is the answer.
This guide breaks down exactly how person-level intent data works, how it differs from the company-level signals most providers sell, and how to use it to put your outreach in front of the actual human who’s in-market right now.
Table of Contents
- The Problem with Company-Level Intent
- What Is Person-Level Intent Data?
- How Person-Level Intent Works Under the Hood
- Provider Comparison: Who Offers What
- Leadpipe’s Intent API Deep Dive
- Use Cases That Actually Move Revenue
- Person-Level Intent + Visitor Identification
- Getting Started with Person-Level Intent
- FAQ
The Problem with Company-Level Intent
The intent data market has exploded. Bombora, 6sense, G2, Demandbase, TrustRadius - all of them promise to tell you which companies are “in-market” for your solution.
And they do. Sort of.
Here’s what a typical company-level intent signal looks like:
“Acme Corp is showing strong intent for ‘CRM software’ - intent score 87/100.”
Sounds useful. But let’s think about what happens next.
Your SDR opens LinkedIn, searches “Acme Corp,” and finds 47 people with titles that could be relevant. VP of Sales. Director of Revenue Operations. Head of IT. Sales Manager, East Region. Three different Marketing VPs.
Which one is actively researching CRM tools? Nobody knows. The intent signal is attached to the company, not a person.
So your SDR picks the VP of Sales because it “seems right,” writes a personalized email about their background, and hits send. Problem is, the person actually doing the research is the Director of RevOps - who never sees the email and won’t for another three weeks until the VP forwards it with “hey, is this relevant?”
By then, they’ve already shortlisted a competitor.
The Spray-and-Pray Tax
This isn’t an edge case. It’s the default workflow for every team using company-level intent:
- SDRs guess who to contact based on title, seniority, and gut feeling
- Sequences target 3-5 people at the company simultaneously, hoping to hit the right one
- Personalization is generic because you’re personalizing to the person’s role, not their actual research behavior
- Response rates stay flat because most of the outreach lands on people who aren’t involved in the buying process
And here’s what makes it worse in 2026: AI SDRs amplify the problem.
When you feed company-level intent into an AI sales agent, the AI picks a contact from your database, generates a beautifully personalized email based on their LinkedIn profile, and sends it. The email looks great. It’s just going to the wrong person.
The AI is doing exactly what you told it to do - reaching out to someone at the company. It just has no way of knowing who at the company is actually in-market.
What Is Person-Level Intent Data?
Person-level intent data ties research signals to individual people, not just companies.
Instead of “Acme Corp is researching CRM tools,” you get:
“Jane Smith, VP of Revenue Operations at Acme Corp, has been researching ‘CRM software,’ ‘Salesforce alternatives,’ and ‘revenue operations platforms’ with an intent score of 85/100 over the last 7 days.”
That’s a fundamentally different signal. You know:
- Who is researching (Jane Smith, not “someone at Acme”)
- What specific topics they’re exploring (not just one category - three related topics)
- When the research activity happened (last 7 days, not “sometime recently”)
- How strong the signal is (85/100, indicating sustained, focused research)
With this data, your SDR doesn’t guess. They email Jane directly, reference the specific problem she’s trying to solve (Salesforce migration, based on her topic research), and offer a relevant resource.
The difference in reply rates between “I noticed your company might be exploring CRM options” and “I saw you’ve been evaluating Salesforce alternatives - here’s how we compare on the specific criteria RevOps teams care about” is night and day.
What Person-Level Intent Data Points Include
| Data Point | Description | Example |
|---|---|---|
| Full name | Verified identity of the researcher | Jane Smith |
| Job title | Current title at their company | VP of Revenue Operations |
| Company | Employer with firmographic data | Acme Corp (500 employees, SaaS) |
| Topics researched | Specific categories they’re consuming content about | CRM software, Salesforce alternatives |
| Intent score | Strength of signal (1-100) | 85/100 |
| Timeframe | When the activity occurred | Last 7 days |
| Business email | Verified work email address | jane.smith@acme.com |
| LinkedIn profile | Profile URL when available | linkedin.com/in/janesmith |
| Phone number | Direct dial when available | (555) 123-4567 |
This isn’t a database lookup. It’s a live signal that tells you someone is actively in-market, right now, for the thing you sell.
How Person-Level Intent Works Under the Hood
Person-level intent isn’t magic. It’s infrastructure. Here’s the technical architecture behind how it works.
1. Cross-Site Pixel Network
The foundation is a network of tracking pixels deployed across thousands of B2B and B2C websites - publishers, review sites, industry blogs, software directories, and content platforms.
When someone visits these sites, their browsing behavior is captured. Not on one site, but across the network. This cross-site visibility is what makes the signal meaningful - it’s the difference between “someone read one article about CRM” and “someone has visited 14 CRM-related pages across 6 different sites in the past week.”
2. Identity Graph Matching
Raw browsing data is anonymous. The critical step is matching those signals to verified identities - real people with names, emails, companies, and titles.
This is where identity graphs come in. A good identity graph uses deterministic matching (verified data connections, not probabilistic guessing) to link browsing activity to specific individuals.
The quality of the identity graph determines the quality of the intent data. If the graph is thin, you get limited coverage. If it relies on probabilistic matching, you get false positives. The best person-level intent providers maintain their own identity graph rather than reselling a third-party one.
3. Topic Classification
Browsing behavior gets classified against a taxonomy of pre-defined topics. The better providers maintain 20,000+ topics covering everything from broad categories (“marketing automation”) to specific products (“HubSpot pricing”) and niche verticals (“dental practice management software”).
The classification engine doesn’t just look at the page URL. It analyzes the page content, the context of the visit, and the pattern of visits across multiple sites to accurately categorize what someone is researching.
4. Scoring
Each person gets an intent score (typically 1-100) based on:
- Signal volume: How many relevant pages they’ve visited
- Recency: How recently the activity occurred (signals from yesterday score higher than signals from two weeks ago)
- Topic specificity: Researching a specific product scores higher than browsing a general category
- Cross-site consistency: Seeing the same topic researched across multiple sites increases confidence
- Engagement depth: Reading a detailed comparison article scores higher than bouncing off a homepage
5. Daily Materialization
Audiences aren’t static exports. They’re refreshed daily as new signals come in. Someone who was showing moderate intent last week might spike to high intent today after visiting three competitor pricing pages. Your audience reflects that in real-time - or at least within 24 hours.
This daily refresh is critical. Intent data has a short shelf life. Research from Bombora’s own studies shows that intent signals lose 50% of their predictive value within 14 days. You need fresh data, not stale exports.
Provider Comparison: Who Offers What
Here’s the honest breakdown. Most “intent data” providers operate at the company level. True person-level intent with individual identities is rare.
| Provider | Level | Topics | Person-Level | API Access | Pricing |
|---|---|---|---|---|---|
| Leadpipe Orbit | Person | 20,000+ | Yes | Full API (18 endpoints) | Included in plans |
| Bombora | Company | 12,000+ | No | Limited | $25K-300K/yr |
| 6sense | Company | Varies | No | Enterprise-gated | $25K-100K+/yr |
| G2 | Company | Software only | No | Limited | Custom |
| Demandbase | Company | Varies | No | Enterprise-gated | Custom |
| TrustRadius | Company | Software only | No | Limited | Custom |
| ZoomInfo | Company | Varies | Some | Enterprise-gated | $15K+/yr |
A few things stand out:
Bombora is the backbone of many intent platforms - 6sense, Demandbase, and others license Bombora’s data. So if you’re paying 6sense enterprise pricing for intent, you’re often getting Bombora data with a different UI on top. And it’s company-level only.
G2 and TrustRadius are limited to software categories. If you’re selling something that doesn’t have a G2 category page, their intent data is irrelevant to you.
ZoomInfo has started offering some person-level signals, but it’s gated behind their enterprise tier and the coverage is inconsistent. Their strength is still the contact database, not intent.
Leadpipe Orbit is the only provider on this list offering true person-level intent with open API access at a price point that isn’t enterprise-only. More on this below.
Leadpipe’s Intent API Deep Dive
Leadpipe’s Orbit API gives you programmatic access to person-level intent data across 20,000+ topics. Here’s what the API actually lets you do.
Browse and Search Topics
You’re not limited to a fixed list. The API lets you browse topics by type (B2B, B2C), industry, and category. There’s an autocomplete search endpoint, so you can find topics like “visitor identification,” “lead generation automation,” or “dental practice management software” without scrolling through a massive taxonomy.
You can also track topic trends over time - see how research volume for a topic has changed over the past 30, 60, or 90 days. Compare trends across up to 10 topics simultaneously. And view the top “movers” - topics growing or declining fastest - to spot emerging demand before your competitors do.
ICP Filtering: The Killer Feature
This is where person-level intent gets powerful. You don’t just pick a topic and get a firehose of names. You define your ideal customer profile using granular filters and get back only the people who match.
| Filter | Options | Example |
|---|---|---|
| seniority | C-suite, VP, Director, Manager, etc. | ”VPs and Directors only” |
| companyIndustry | 100+ industries | ”SaaS, Technology, Marketing” |
| companySize | Employee ranges | ”50-500 employees” |
| department | Sales, Marketing, Engineering, etc. | ”Marketing department” |
| companyRevenueRange | Revenue ranges | ”$10M-100M” |
| state | US states | ”California, New York” |
| hasBusinessEmail | Boolean | Only people with work emails |
| hasLinkedin | Boolean | Only people with LinkedIn profiles |
| hasPhone | Boolean | Only people with phone numbers |
| jobTitle | Free text search | Contains “Marketing” or “Growth” |
So instead of “people researching CRM software,” you get: “VPs and Directors in Marketing or Sales at SaaS companies with 50-500 employees and $10M-100M revenue who are researching ‘CRM software’ with a business email and LinkedIn profile.”
That’s not a list. That’s a target-rich outbound campaign ready to go.
Preview Before You Commit
Before you burn through credits, you can preview any audience. The API returns:
- Total count: How many people match your filters
- 50 masked samples: Anonymized examples showing titles, companies, and industries so you can validate the audience quality
This matters because intent audiences can vary wildly in size depending on your filters. “CMOs researching AI” might return 15,000 people. Add a company size filter of 50-200 employees and it drops to 800. The preview lets you iterate on filters without wasting exports.
Save, Refresh, Export
Once you’ve built an audience you like:
- Save it - the audience definition is stored and refreshed daily with new intent signals
- Query it via API - pull the latest audience programmatically for your workflows
- Export as CSV - get a signed download URL for one-click export into your CRM, enrichment waterfall, or ad platform
The daily refresh is automatic. People who stop showing intent fall out of the audience. New people who start showing intent get added. Your audience is always current.
18 Endpoints
The full Orbit API includes 18 endpoints covering:
- Topic browsing, search, and autocomplete
- Topic trend analysis and comparison
- Top movers (fastest-growing and fastest-declining topics)
- Audience creation with ICP filters
- Audience preview (count + samples)
- Audience save and management
- Audience export (CSV with signed URLs)
- Webhook notifications for audience refreshes
For developers building on top of intent data - whether you’re feeding it into an AI agent, a custom scoring model, or a proprietary outbound engine - the API is designed to be the infrastructure layer, not just a dashboard you log into.
Try Leadpipe free with 500 leads ->
Use Cases That Actually Move Revenue
Person-level intent isn’t just “better data.” It unlocks workflows that are literally impossible with company-level signals.
1. AI SDR Targeting That Actually Works
Here’s the problem with feeding company-level intent into AI agents: the agent picks a contact semi-randomly from your database and generates personalized outreach based on their LinkedIn profile. The email sounds great but has nothing to do with what that person actually cares about.
With person-level intent, you feed the AI the exact person who’s researching relevant topics, along with what they’re researching. The agent can then write outreach that references the buyer’s actual research behavior, not just their job title.
The difference: “Hi Jane, I noticed Acme Corp is exploring CRM options” vs. “Hi Jane, I saw you’ve been evaluating Salesforce alternatives - here’s our migration guide for RevOps teams.”
Read more: The Data Layer AI Sales Agents Are Missing and How to Choose a Data Provider for Your AI SDR
2. ABM Campaigns with Named Targets
Traditional ABM starts with a target account list and then tries to find the right people at those accounts to engage. With person-level intent, you flip the process:
- Pick topics relevant to your solution
- Filter for your ICP (seniority, company size, industry)
- Get back named decision-makers at companies you may not even have on your target list
You’re not just fishing in your existing account list. You’re discovering new accounts through the people who are actively in-market.
3. Sales Prioritization and Lead Scoring
Your inbound leads don’t arrive with equal intent. Some fill out a form because they want the gated content. Others are deep into an evaluation.
Layer person-level intent data on top of your inbound leads. If someone who submitted a demo request has also been researching your category and competitors across the web for the past two weeks, that’s a different conversation than someone who clicked one ad.
This works especially well when combined with visitor identification. You can see that someone visited your pricing page AND has been researching related topics across the web.
4. Competitive Intelligence
Want to know who’s researching your competitors? Create audiences for competitor-related topics:
- “HubSpot alternatives”
- “Salesforce pricing”
- “ZoomInfo competitors”
Filter by your ICP, and you get a daily-refreshed list of decision-makers actively evaluating your competitive set. That’s not a cold list - it’s a warm list of people who are about to make a buying decision.
5. Content Strategy Driven by Real Demand
Most content teams guess at what their audience cares about. With topic trend data, you can see what’s actually being researched - and what’s growing fastest.
If “AI SDR” as a topic grew 340% in the last 90 days, that’s a content opportunity. If “marketing automation” is flat, maybe that blog post can wait.
The movers endpoint literally shows you the topics gaining and losing momentum, updated daily.
6. Ad Targeting with Intent Audiences
Export a person-level intent audience as CSV, upload it to LinkedIn or Google Ads as a custom audience, and run ads specifically to people who are actively researching your category.
This is dramatically more efficient than broad targeting. You’re not showing ads to everyone with a “VP of Marketing” title. You’re showing ads to the VPs of Marketing who researched “lead generation tools” in the last 7 days.
The AI SDR data stack approach combines this with visitor identification and enrichment for a complete pipeline.
Person-Level Intent + Visitor Identification = Complete Picture
Here’s something most teams miss: intent data and visitor identification are complementary signals, not competing ones.
| Signal Type | What It Tells You | Where It Comes From |
|---|---|---|
| Person-level intent | Who’s researching topics relevant to you across the web | Cross-site pixel network |
| Visitor identification | Who’s visiting your website specifically | Your site’s identification pixel |
Intent data tells you who’s in-market. Visitor identification tells you who’s already looking at you.
Together, they create a complete picture:
- High intent + visited your site = hottest leads. They’re researching your category AND evaluating you specifically. Sales should be on the phone today.
- High intent + hasn’t visited your site = outbound opportunity. They’re in-market but haven’t found you yet. Perfect for targeted outreach or ads.
- Low intent + visited your site = early-stage interest. They’re browsing but not deep into research yet. Nurture sequence territory.
- No intent + visited your site = could be a customer, job seeker, investor, or someone who clicked a link by accident. Lower priority for sales.
Leadpipe is the only platform that provides both person-level intent data AND visitor identification through a single API. You don’t need to stitch together Bombora + Clearbit + a visitor ID tool. One integration, one identity graph, one data model.
This matters for AI agents especially. Instead of cobbling together three different data sources with different schemas, refresh rates, and accuracy levels, the agent gets a unified view of every prospect.
Read the full breakdown: Intent Data vs. Visitor Identification
Getting Started with Person-Level Intent
If you’re currently using company-level intent (or no intent data at all), here’s the path to person-level:
Step 1: Browse Topics Relevant to Your Category
Start by searching the topic taxonomy for keywords that match your market. If you sell marketing automation, search for:
- “Marketing automation”
- “Email marketing platforms”
- “Lead nurturing”
- “HubSpot alternatives” (competitor topics)
- “Marketing ops tools”
The autocomplete endpoint makes this fast. You’ll likely find 10-30 topics directly relevant to your business.
Step 2: Set Your ICP Filters
This is where you narrow the firehose. Define who you actually want to reach:
- Seniority: VP and above? Director-level? Managers?
- Company size: SMB (10-200)? Mid-market (200-2000)? Enterprise (2000+)?
- Industry: SaaS? Healthcare? Financial services?
- Department: Marketing? Sales? IT?
- Geography: US only? Specific states?
- Contact data: Must have business email? Must have phone?
Step 3: Preview the Audience
Before committing, preview the audience to validate:
- Is the total count reasonable? (Too small = filters too narrow. Too large = filters too broad.)
- Do the sample profiles look like your actual buyers?
- Are the companies the right type and size?
Iterate on filters until the preview looks right.
Step 4: Save and Activate
Save the audience. It’ll refresh daily with new intent signals. People who stop showing intent will drop out. New people who start showing intent will get added.
Step 5: Export or Query via API
Pull the audience into your workflow:
- Export CSV → upload to CRM, Clay waterfall, or ad platform
- Query via API → feed directly into your AI agent, custom scoring model, or outbound tool
- Set up webhooks → get notified when new high-score prospects enter your audience
For the full API walkthrough, including code examples and endpoint documentation, see: Leadpipe Intent API: 20,000+ Topics
Try Leadpipe free with 500 leads ->
FAQ
How accurate is person-level intent data?
Accuracy depends on two things: the quality of the identity graph and the quality of the topic classification. Person-level intent that uses deterministic matching (verified data connections) is significantly more accurate than probabilistic matching (statistical guessing). Leadpipe uses its own deterministic identity graph - the same one that powers its visitor identification with 30-40%+ match rates. The intent scores are based on observed cross-site browsing behavior, not modeled or inferred data.
How often is the intent data refreshed?
Leadpipe’s intent audiences refresh daily. New signals are processed overnight, and your saved audiences reflect the latest data by morning. This matters because intent signals decay fast - research shows they lose about 50% of their predictive value within 14 days. Daily refresh means you’re always working with current signals, not stale data from last month.
What’s the difference between person-level intent and visitor identification?
Visitor identification tells you who visited your website. Person-level intent tells you who’s researching topics relevant to you across the web (on third-party sites, review platforms, competitor sites, etc.). They’re complementary. Someone might be deep into CRM research across the web but hasn’t found your site yet - intent data catches them. Someone might visit your pricing page but show no cross-web research - visitor ID catches them. Leadpipe offers both through a single platform.
Do I need an enterprise contract to access the intent API?
No. Unlike Bombora ($25K+ minimum), 6sense (enterprise-gated), and Demandbase (custom pricing), Leadpipe’s intent data is available through its standard pricing plans. The Orbit API with all 18 endpoints is accessible to all customers. There’s no separate “intent add-on” or enterprise gate. You can start with a free trial of 500 leads to test both visitor identification and intent data.
Related Articles
- Intent Data vs. Visitor Identification: What’s the Difference?
- Leadpipe Intent API: 20,000+ Topics for Person-Level Buyer Intent
- Visitor Identification API: The Complete Developer Guide
- The Data Layer AI Sales Agents Are Missing
- How to Choose a Data Provider for Your AI SDR
- Best Contact Enrichment APIs in 2026
- How to Easily Identify Anonymous Website Visitors
