A list of names is not a list of contacts. Ariel uses waterfall contact data enrichment to look for missing email addresses and phone numbers across multiple sources. Ariel verifies newly found email addresses before charging credits. Invalid addresses are discarded, and the search continues through the waterfall. The Contact Details drawer shows whether a record is Enriched or Partial. Direct dials are checked against suppression and do-not-call controls before a representative can place a call. AI-written messages require human approval before sending. Ariel also checks suppression records before each send and each Power Dialer call.
For teams considering an AI SDR for contact data enrichment, the aim is to find usable contact details without repeatedly paying for fields that are already present.
Trusted by 10,000+ users across all sizes and industries
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
Every contact database has gaps. One provider is strong on mid-market SaaS and thin on manufacturing. Another has better coverage in Europe than in North America. Query any single source and you get back a list that is missing a third of the emails and half the direct dials, and the missing third does not distribute evenly across the accounts you want most.
The fix is not a bigger single database. It is running several sources in sequence, so each one only has to solve the part the one before it missed. That sequence is the waterfall, and it is the difference between an enrichment step that quietly gives up on a chunk of your list and one that keeps trying until it has exhausted its options.
When evaluating B2B contact data providers, check how they find missing fields and verify newly found emails. Ariel stores source information for individual fields, while the interface uses the neutral label Enrichment rather than a visible or configurable provider list.
A provider that covers mid-market SaaS well can be thin on manufacturing, and the gap does not distribute evenly across the accounts you want most.
Coverage that is solid in one geography can be patchy in another, so a single source quietly gives up on a chunk of your list.
Running several sources in order means each one only has to solve the part the one before it missed.
Here is a worked example, using the same batch shown on Ariel's own product tour. Treat the numbers as an illustration of how the mechanic behaves, not as a hit rate you should expect on your own list. Your result depends on your industry, your seniority mix and your geography.
| Pass | What runs | Result on this 25-contact batch |
|---|---|---|
| Pass 1 | Primary data provider searches for emails and direct dials | 17 of 25 found |
| Pass 2 | A second provider runs only on the 8 contacts pass 1 missed | 5 more found |
| Pass 3 | A third provider runs only on the 3 contacts still missing | 1 more found |
| Verification | Every email from every pass is checked in real time | 2 rejected as invalid or risky |
| Result | Contact records with verified email and/or other useful contact details | Example states 23 of 25 ready, including 21 verified emails and 14 direct dials |
Two things matter more than the final count. First, each pass only works the contacts the previous one missed, so you are not paying for the same lookup three times over. Second, verification happens before a credit counts as spent, not after, so a found address still has to prove it is real before it is charged for or marked usable.
Finding an email address and confirming it exists are different jobs, and treating them as the same one is how bounce rates climb.
Every address that comes out of the waterfall is checked in real time before it is marked usable. The result lands in one of four states: verified, invalid, risky or unverified. An invalid address is discarded and the ladder keeps looking. A risky address, one the system cannot confirm is safe to send to, is flagged rather than treated as ready. Only a verified address is marked usable, no matter how confident the source that found it seemed.
This matters because a sending domain does not care how the address was sourced. It cares whether the mail bounces. One provider's confident guess and another provider's verified hit look identical in a spreadsheet, and only one of them is safe to mail. It also matters for your credits: an address that fails verification is never charged for, so a bad hit costs you nothing but the wait.
A phone number is not automatically ready for a call step just because it was found. Direct dials pulled through the waterfall are checked against workspace suppression and do-not-call rules before they can be used, so a number already blocked in your workspace must not be dialled from a queue. Finding a phone number does not automatically make it eligible for a call. Before each Power Dialer call, Ariel checks the relevant suppression and do-not-call records and the prospect’s local calling window. Calls that fail these checks are blocked, with the reason shown to the representative.
The built-in calling window is 9 a.m. to 6 p.m. on weekdays in the prospect’s time zone. This is a human-operated Power Dialer workflow, not AI voice calling. Ariel checks eligibility; the sales representative conducts the conversation.
That same check runs a second time, on the server, the moment a call is about to be placed, alongside the person's local calling window and any recording rules for that market. A direct dial that clears data enrichment but fails that check at call time still will not ring, and the rep sees the exact reason rather than a call that silently never happens.
This is a human-operated Power Dialer workflow, not AI voice calling. The system checks eligibility; the sales representative conducts the conversation.
A phone number comes out of the waterfall pass.
Checked before the number can be used in a call queue.
The same check runs again, plus the person's local calling window.
Every waterfall pass and every verification call has a cost, and letting that run unmonitored is how a data budget disappears without anyone noticing. Ariel shows a running "Credits: {balance}" pill on every contact you enrich, and a spending brake automatically stops a bulk run once it reaches 80 percent of what is left in the balance, so a large list can not silently empty the account.
Credits spent chasing an address that turns out invalid or risky are not the ones that hurt, because those are never charged. Verification runs before a credit counts as spent, which is what keeps an enrichment run from turning into credits well spent and credits wasted in equal measure.
Enrichment finds and verifies contact details. It does not decide whether the list was worth building in the first place. A perfectly enriched list of the wrong people produces verified emails to prospects who were never going to reply.
It also will not turn a stale, three-year-old list into a live one on its own. A provider working from outdated employment records returns fewer matches on people who have since changed jobs, and no amount of waterfalling recovers information the sources do not have. There is no scheduled sweep that refreshes a stale database for you today. Refreshing a contact is a manual, per-contact Force re-enrich, so revisiting an old list is worth doing deliberately rather than assuming it happens on its own.
And it does not promise 100 percent coverage. Some contacts will not be found by any of the sources run, and some found addresses will not survive verification. A workflow that rejects unverified addresses rather than sending to them anyway is the point, even when that means a smaller list than you started with.
run in sequence before a contact is called unfindable
users running their pipeline on Salesflow
before a contact is eligible for another enrichment pass
No implementation project. The waterfall and verification run on the first list you enrich.
Contacts enter the waterfall in the order they are queued.
Each provider after the first only works what the one before it missed.
Invalid and risky addresses are rejected before anything enters a campaign.
Running several data providers in sequence for each contact, where each provider after the first only works the contacts the previous ones missed. It finds more usable contacts than querying a single source.
No. Every email is verified before it is marked usable, and only a verified address is ready to enter a campaign.
One the system cannot confirm is safe to send to. Every address lands in one of four states: verified, invalid, risky or unverified, and only a verified one is treated as ready.
Direct dials are checked against the do-not-call register and suppression rules when they are first found, and checked again against the same rules, plus the person's local calling window, the moment a call is about to be placed.
It depends on how many passes are needed to find and verify that contact. Only a newly found email or phone number is charged, LinkedIn URLs and job titles are always free, and a running credit balance is shown on every contact you enrich.
Multiple data providers, run in sequence. The specific mix depends on the region and the type of contact, and no single provider name is the whole story for any one workspace.
No. It finds and verifies details for the people already on the list. It does not judge whether those people match your buyer profile.
Book a 30 minute demo and watch Ariel find buyers, write the outreach and line up meetings, with you approving every move.
Book a Demo See Ariel