How AI scoring works
|MagicPost Leads scores each person who engages with your LinkedIn posts in two steps: a free pre-qualification against the keywords and filters of your ICP, then an AI scoring of the profiles that passed. The AI gives a fit level: Excellent, Good or Medium.
A profile that the MagicPost AI scores below Medium is set aside: it is not a lead and it costs no credit.
AI scoring is part of the MagicPost Leads module, included in the Growth, Creator, Agency, and Team plans. One credit is spent per profile scored with at least a Medium fit, per ICP. MagicPost presents the module on the LinkedIn leads feature page.
The fit levels
MagicPost shows three fit levels on a lead, from the closest match to the weakest: Excellent, Good and Medium. Only a profile with one of these three levels becomes a lead.
A profile that the MagicPost AI judges too far from your ICP is set aside. It does not appear in the Leads tab, it is not counted in your leads, and it spends no credit.
The Engagers tab of a post in MagicPost Leads lists everyone who liked or commented, set-aside profiles included, and shows no fit level. A set-aside profile carries the same Lead pill there as a lead and sits under the Already a lead filter, because MagicPost keeps a record of the profile. That pill does not make the profile a lead.
Step 1: pre-qualification, free
MagicPost first checks each profile against the keywords and the filters of the ICP. No credit is spent at this step, whatever the outcome. There are three outcomes:
The profile is ruled out. An exclusion keyword matches, or the profile fails a strict filter: the list of included countries, any exclusion on country, industry, company size or city, Followers, or LinkedIn profile. During a detection, the profile is not added to your leads and is not sent to the AI scoring. The person stays in the Engagers tab of the post, under the Unqualified filter, unless the person is already one of your leads.
The profile is a clear match. One Essential keyword matches, or two Secondary keywords do, and no filter objects. The profile goes to the AI scoring.
The profile is uncertain. It has too few keyword matches, or it is outside your included industries, company sizes or cities. MagicPost does not reject the profile on that basis. A lighter AI check, also free, decides whether the profile is worth a full scoring. A profile that the lighter check rejects is not added to your leads either.
MagicPost skips a filter when the profile has nothing in the matching field. A profile with no country is not rejected by a country filter. To tune keywords and filters, see how to edit an existing ICP.
Step 2: AI scoring
The MagicPost AI then reads each profile that passed pre-qualification and compares it with your ICP, that is, its description and its example profiles. The scoring produces the fit level and its explanation.
What the AI reads
The MagicPost AI scoring reads the LinkedIn profile of the person:
Name, headline and current role
Company, industry and company size
Country
About section
Company description
The MagicPost AI scoring does not read the comment the person left under your post. MagicPost saves the comment and shows it on the lead, but the fit level depends only on the profile.
Where to read the score of a lead
In MagicPost, click a lead to open its card, then open the ICP affinity tab. For each ICP, the tab shows:
The fit level.
A quote from the profile that carried the decision.
The reasoning, in a few short points.
The signals of the ICP that the profile matched or contradicted.
The pre-qualification result, Pre-qualified or Disqualified, with the keywords that matched.
In the ICP affinity tab of a MagicPost lead card, the pre-qualification line reads Pre-qualified for a clear keyword match and Disqualified in every other case. Disqualified on that line does not mean the lead was ruled out. A lead that went through the lighter AI check shows Disqualified there, although the AI scored it and it can be Excellent, Good or Medium.
What to do when you disagree with a score
A MagicPost fit level is a recommendation, and you have three ways to act on it:
Score the lead again. In the Leads tab, click the fit level of the lead, click Re-check affinity, then confirm with Re-check. The new score replaces the old one.
Add a profile to an ICP by hand. In the lead card, Add to a list adds the profile to another of your ICPs, at no cost.
Fix the ICP. Repeated misses usually come from the description, the keywords or the filters of the ICP.
What costs a credit
Collecting the people who engaged with a post: free.
Pre-qualification, including the lighter AI check: free.
A profile scored below Medium and set aside: free.
A profile scored Excellent, Good or Medium on an ICP: one credit.
A lead scored again: one credit if the new fit is at least Medium, nothing otherwise.
Adding a profile to an ICP list by hand, or removing it: free.
For the monthly quota of each MagicPost plan and the ways to buy more credits, see understanding and buying Leads credits.
FAQ
Is there a Low fit level in MagicPost?
Not in your leads. MagicPost shows three levels, Excellent, Good and Medium. A profile scored below Medium is not a lead: it is set aside, hidden from the Leads tab, left out of your lead counts, and free.
The same person matches two of my ICPs. Does that cost two credits?
Yes, if the person is a lead on both. MagicPost scores a profile against each ICP separately, and each score of at least Medium costs one credit. In the Detect leads dialog, you choose which ICPs a detection runs on.
Does the comment someone wrote change their score?
No. The MagicPost AI scoring reads the LinkedIn profile only. The comment is saved and shown on the lead, without any effect on the fit level.
What is the difference between Disqualified and Set aside?
They come from two different steps of MagicPost Leads, and both are free. Set aside is the outcome of the AI scoring: the AI read the profile and judged it too far from the ICP, so the profile is not a lead and is hidden from the Leads tab. Disqualified is a label of the pre-qualification line in the ICP affinity tab of a lead card. It appears whenever the profile was not a clear keyword match, including on a lead that the AI then scored.
Where do the profiles ruled out by pre-qualification go?
Nowhere in your leads. During a MagicPost detection, a profile ruled out by a keyword or a filter is not added to the Leads tab and costs nothing. The person only stays in the Engagers tab of the post, with everyone who liked or commented.
Does scoring stop when I have no credits left?
Yes. At zero credits, MagicPost pauses detection and AI scoring until you have credits again. See understanding and buying Leads credits.
Need help?
Reach out via the in-app chat, or book a 30-minute onboarding call: https://cal.com/magicpost/book-a-call