AI Leads has two additions. You can exclude job titles you never want to see, and you can run a two credit probe that tells you how many leads a search would return before you spend credits pulling them.
What changed
Building a lead search used to be a guess with a price attached. You described who you wanted, ran it, spent the credits, and found out afterwards whether the search was any good. If the criteria were too narrow you got a handful. Too broad and you got a list full of people who were never going to be relevant.
Either way the credits were gone, and the natural response was to widen the search on the next attempt, which usually made the list worse.
The probe changes the order. For two credits it tells you how many leads match your criteria. You see the size of the result before you commit to it, so you can adjust and check again cheaply until the number looks right.
Title exclusions handle the other half. Most searches by seniority or function pull in adjacent roles you do not want. A search for heads of marketing returns marketing interns, marketing assistants, and anyone whose title happens to contain the word. You can now name the titles to exclude and have them filtered out before you pay for them.
How to use it
Set up your search as usual: the roles, the company profile, the geography.
Add exclusions for titles you know you do not want. Be specific. Excluding "assistant" removes assistants. Excluding "marketing" would remove everything you were looking for.
Run the probe. It returns a count for two credits. Read the number against what you expected.
A count much lower than expected means the criteria are too tight. Usually it is geography or company size doing the narrowing rather than the role. A count much higher than expected means the role definition is catching more than you meant, and the fix is usually a more specific title or a tighter exclusion.
Adjust and probe again. Each probe is two credits, so several rounds of tuning cost far less than one badly aimed full pull. When the count looks right, run the full search.
Why it matters
A bad lead list is expensive twice. You pay credits for it, and then you pay again in domain reputation when you send to people who were never plausible prospects and some of them mark the message as spam.
The second cost is the larger one and the one that lasts. Credits are replaceable. A sending domain with a complaint problem takes weeks to recover, and it affects every campaign you run in the meantime, not only the one that caused it.
Being able to see the shape of a list before buying it moves the decision to before the money and before the sending. Two credits to find out that a search returns eleven thousand people, when you expected two hundred, is a good trade.
Using exclusions well
Exclusions match against the title text, so think about what the unwanted titles actually say rather than what they mean.
Seniority words are the most useful: assistant, intern, junior, coordinator, trainee. These reliably identify roles below the level you are targeting and rarely appear in the titles you want.
Be careful with words that appear in both. Excluding "manager" to remove junior managers also removes general managers and senior managers. When a word cuts both ways, exclude the longer specific phrase instead.
Build the exclusion list up over time. The titles that keep appearing in your lists and never convert are the ones worth adding, and after a few campaigns you will have a list that reflects your actual market rather than a guess.
What the probe costs and covers
Two credits per probe, regardless of how many leads match. The count it returns is the number that would come back from a full pull with the same criteria.
The probe does not return the leads themselves and does not reserve them. It tells you how many exist.
Reading a count that surprises you
A count in the single digits usually means two criteria are fighting each other. A specific role combined with a narrow geography and a company size band leaves very little, and the fix is to loosen one of the three rather than all of them.
A count in the tens of thousands usually means the role is doing no work. Broad function words match a great deal, and adding one level of specificity typically cuts the number by an order of magnitude.
A count that looks right is worth a second look at the criteria line before you commit. The number being plausible does not confirm the search is asking what you meant.
What this changes about how you buy leads
The old pattern was to buy large and filter afterwards, because you could not see what you were getting until you had it. That leaves you paying for the rows you throw away and tempted to send to them anyway.
Probing first makes small, specific lists cheap to arrive at. Several tightly aimed searches usually beat one broad one, and now they cost less to find.
Credits and what they buy
A probe is two credits whatever the result. A full pull charges per lead returned, so the probe is the cheap part of the process by a wide margin.
Nothing is charged for a probe that returns zero, and nothing is charged for leads that turn out to be duplicates of contacts you already hold.
Probes and pulls share one credit balance with the rest of AI Leads, visible in your billing screen with the rest of your usage.
Availability
Both are available now in AI Leads on all plans. Probes and full pulls draw on the same credit balance.