How MarketBetter Uses Exa Websets to Build Audiences with Natural Language

Every B2B sales team has experienced the same frustration: you know exactly who you want to sell to, but translating that knowledge into a prospect list takes forever.

You open your data tool. You set filters — industry, employee count, revenue range, job title keywords. You run the search. Half the results are wrong. The CFO you wanted is actually a "Chief Fun Officer" at a 3-person startup. The healthcare companies include veterinary clinics. The 50-200 employee filter caught a company that had 200 employees three years ago but now has 12.

Traditional B2B search forces you to describe your ideal customer through rigid filters that were never designed to capture nuance. MarketBetter's integration with Exa changes that entirely. You describe who you want in plain English, and the system finds them.

The Problem with Filter-Based Prospecting

Conventional audience building tools give you dropdown menus and checkboxes. Industry. Location. Company size. Job title. Maybe some technographic filters if you are paying for the premium tier.

These filters work fine for simple, well-defined segments. "Marketing directors at SaaS companies with 100-500 employees in the US" is a query that dropdowns handle reasonably well.

But most real buying personas are not that clean.

Consider what a sales leader actually says when describing their ideal customer:

• "Series B or C funded companies that recently expanded into Europe and are hiring their first sales ops person"
• "Manufacturing companies that still use spreadsheets for production scheduling and have a plant manager who reports directly to the CEO"
• "Healthcare IT companies that lost their CISO in the past 6 months and have active compliance initiatives"

Try translating any of those into dropdown filters. You cannot. The nuance gets crushed into approximations that return thousands of irrelevant results alongside the handful of companies you actually want.

This is the gap that Exa Websets fills.

What Exa Websets Actually Does

Exa is an AI-native search engine built for machines, not humans scrolling through blue links. Their Websets product takes a natural language description of the entities you want — people, companies, or both — and returns structured, verified results.

Instead of filters, you write a description. Instead of boolean logic, you define criteria in plain English. Instead of hoping your keywords match whatever a company put on their LinkedIn page, the AI understands what you mean.

When you use Exa Websets through MarketBetter, the system:

1. Translates your intent — Your natural language query gets interpreted into a structured search that captures the actual meaning, not just keyword matches
2. Searches across the web — Exa crawls and indexes billions of pages, going far beyond the usual LinkedIn-plus-Crunchbase data sources
3. Applies your criteria — Each potential result gets evaluated against your specific requirements, filtering out false positives that keyword matching would miss
4. Enriches the results — Verified email addresses and phone numbers get appended to matching contacts, so you go from "interesting companies" to "actionable outreach list" in one step
5. Normalizes everything — Results come back in a clean, consistent format ready for your CRM or outreach sequences

The entire pipeline runs asynchronously. You describe what you want, MarketBetter kicks off the search, and results stream in as they are found and verified. No sitting on a loading screen for ten minutes.

How It Works Inside MarketBetter

The integration lives directly inside MarketBetter's audience builder — the same workspace where you manage all your prospecting, signals, and outreach.

Describe Your Audience

Open the audience builder and choose Exa as your data provider. Then describe who you are looking for. You can be as specific or as broad as you want:

"VP of Sales or CRO at B2B SaaS companies with 200-1000 employees that recently raised Series C funding and are expanding into EMEA markets"

Or go narrower:

"Heads of procurement at industrial manufacturing companies in the Midwest that use Oracle ERP and have more than 3 production facilities"

The system parses your description, identifies the entity type (person, company, or both), and builds the search parameters automatically. You do not need to learn a query language or figure out which filters map to which fields.

Choose Your Enrichments

Before the search runs, you decide what contact data you need. Options include:

• Email addresses — Verified business emails attached to the person, not generic info@ addresses
• Phone numbers — Direct dials where available

Each enrichment type has a credit cost that MarketBetter shows you upfront. You see the estimate before committing, so there are no surprises on your bill. The credit system is transparent: you know exactly how many credits a search will consume based on the number of results and the enrichments you select.

Watch Results Stream In

Once you confirm, the search begins. Results do not arrive all at once — they stream in as Exa processes your query and verifies enrichments. MarketBetter shows you progress in real time: how many matches have been found, how many enrichments are complete, and how long the search has been running.

If a search is taking longer than expected (complex queries with multiple enrichments sometimes need a few minutes), you get partial results rather than nothing. The system returns whatever has been verified so far, so you can start reviewing while the rest completes.

From Audience to Action

Results land directly in your MarketBetter workspace as a saved audience. From there, the standard MarketBetter workflow kicks in:

• Route to SDRs based on territory, industry, or round-robin rules
• Enroll in sequences for automated multi-touch outreach
• Layer signals on top — combine your Exa-sourced audience with MarketBetter's intent signals, website visitor identification, and champion tracking to prioritize who to call first
• Push to CRM — sync to Salesforce or HubSpot with full attribution

The audience does not live in a silo. It becomes part of your entire go-to-market motion.

The Waterfall Advantage: Exa as a Coverage Backstop

One of the less obvious but more powerful aspects of the integration is how MarketBetter uses Exa as a fallback data source.

When you are researching a specific company — maybe a website visitor showed up on your dashboard, or a champion just changed jobs — MarketBetter searches for contacts at that company. The primary data provider handles most lookups, but coverage is never 100%. Some companies, especially smaller or newer ones, simply are not in every database.

When the primary source comes back empty, MarketBetter automatically falls back to Exa. The system searches for people at that company with matching job titles across LinkedIn and the broader web. This waterfall approach means you get coverage where a single-provider setup would give you a dead end.

You do not have to configure this. It happens automatically. The system tells you which provider sourced each contact, so you always know where your data came from.

Finding Lookalike Companies

Beyond people search, Exa's similarity engine helps you find companies that look like your best customers.

Give it a company URL — say, your highest-value customer — and it returns companies with similar characteristics. Not just the same industry and size (any database can filter for that), but genuinely similar in terms of what they do, how they position themselves, and who they serve.

This is particularly useful for:

• Expanding into adjacent verticals — Find companies similar to your champion customer in a new industry
• Building account lists from a seed — Start with 5 ideal customers and discover 500 that look like them
• Competitive displacement — Find companies that use a competitor's product by looking for companies similar to the competitor's known customers

The results feed back into MarketBetter's audience workflow. You find lookalikes, enrich them with contact data, layer on intent signals, and route to reps — all without leaving the platform.

What This Means for Your Prospecting Workflow

The practical impact comes down to three things:

1. Speed

What used to take a full afternoon — defining filters, running searches across multiple tools, exporting CSVs, deduplicating, enriching through a separate tool, uploading to your CRM — now takes minutes. Describe, search, enrich, route. One workflow, one platform.

2. Precision

Natural language search captures intent that filters cannot. When you describe "companies that recently started selling into the EU and need to comply with GDPR," you get companies that actually match that profile — not just companies tagged "Europe" in someone's database from 2019.

3. Coverage

The waterfall architecture means you are never stuck with one provider's blind spots. If your primary data source does not have contacts at a company you care about, Exa fills the gap automatically. More coverage means fewer missed opportunities and less manual research for your SDRs.

Getting Started

If you are already on MarketBetter, the Exa integration is available in your workspace settings. Add your Exa API key (you can get one at exa.ai), and the provider appears as an option in your audience builder.

If you are evaluating MarketBetter, the Exa integration is one of several data providers available out of the box — alongside Fiber for company and contact data, and Lusha for verified business contacts. You choose the right provider for each search, or let the platform combine them automatically.

The bottom line: your ideal customers exist. Describing them should be the hard part. Finding them should not be.

---

MarketBetter integrates with Exa, Fiber, Lusha, and more to give your sales team the most complete view of your total addressable market. See it in action.