# How AI lead scoring works for LinkedIn outbound
> Learn how AI lead scoring compares LinkedIn profiles to your ideal customer profile and assigns a fit score, so you can focus outreach on better matches.
- HTML: https://getkinlead.com/explaining-ai-lead-scoring-logic
- Markdown: https://getkinlead.com/explaining-ai-lead-scoring-logic.md

**Short answer:** AI lead scoring compares a prospect's LinkedIn profile against your ideal customer criteria. It reads job titles, company descriptions and experience for meaning, then assigns a fit score from 0 to 100. Higher scores mean a closer match, so you can spend invites and messages on the people most likely to buy.

## What is AI lead scoring?

Most LinkedIn searches start with hard filters. You type a job title, pick a location, and get everyone who matches those exact words. The trouble is that people describe the same job in different ways. A founder at a 12 person agency might call herself CEO, Managing Director, or Principal. A demand generation lead might appear as Head of Growth, Marketing Lead, or something else.

AI lead scoring replaces the exact-word filter with a meaning check. You describe the customer you want, or point the system at your website. The scoring model reads each profile and asks how closely this person matches the work, company and problem you described.

You still get a list. The difference is that the list is ranked. A score of 90 means the profile looks like your buyer. A score of 40 means the system found a loose connection but not enough evidence. You decide where to draw the line.

**Exact filter**

- Search for Marketing Manager
- Misses Growth Lead and Head of Demand Gen
- Every result looks equal

**AI scoring**

- Describe the buyer in plain language
- Finds related titles and company types
- Each lead gets a 0 to 100 score

## How the scoring logic works

The model does not only scan for a list of keywords. It builds a picture of your ideal customer from the description you give it. That picture includes the kind of company, the person's role, the problems they likely handle and the language they use about their work.

Each profile is then compared with that picture. A headline that mentions the same problem your product solves raises the score. A company outside your target market lowers it. A job history that shows the buyer has handled the same work before also raises it. The output is a number from 0 to 100, not a yes or no.

Company fields matter as much as the title. A Head of Sales at a 20 person software company is a different buyer from a Head of Sales at a 2,000 person bank. The model reads the company description, employee range and industry to separate them.

1. **Read your target.** The system takes your website or a written description and pulls out the buyer's role, company type, size range and the problem you solve.
2. **Read the profile.** It reads the headline, about section, current company and past roles. It looks for meaning, so related titles can match even when the words differ.
3. **Compare the evidence.** The model checks whether the profile shows the same job, company context and problem. A close match on all three pushes the score up.
4. **Assign the score.** The lead gets a number from 0 to 100. The score is a ranking signal, not a promise that the person will reply.

## Setting your fit line

The fit line is the lowest score you keep. It is the one setting that changes how much of your list survives. If you keep every lead above 50, you will have a large list and a lot of weak matches. If you keep only leads above 85, you will have a smaller list and a cleaner reason to reach out.

LinkedIn does not publish exact limits for invites or messages. Every account has some ceiling, and every low-fit invite uses one of those actions. The fit line is how you protect that budget. You want enough leads to run your week and few enough that the notes you write still sound specific.

Start with a line that feels slightly strict. Run it for a week. If you run out of leads before the week ends, lower the line by a few points. If the leads look right on paper but you cannot write a relevant first line, the line is too low or your target description is too broad.

Your fit line should match your capacity. If you can only send a handful of invites in a day, keep the line high. If you have a lot of capacity and a broad market, lower it and test. The line is a dial, not a permanent rule.

Keep two groups if you are unsure: a strict group above 85 and a test group from 70 to 84. Send to the strict group first. If you run out, use the test group. Do not mix them in one campaign, because you will not know which group produced the replies.

| | Score band | What it usually means | What to do |
| --- | --- | --- | --- |
| 90 to 100 | Strong match on role, company and problem | Write to these first; use the profile detail in your note |
| 70 to 89 | Good match with one weak signal | Keep them; check the profile before you send |
| 50 to 69 | Loose match | Use for tests only; expect to write generic notes |
| Below 50 | Poor match | Do not keep |

> A high score is a reason to look at the profile, not a reason to skip reading it. If you cannot name the detail that made the person a lead, do not send the invite.

## Improving your scoring accuracy

If the model keeps scoring bad leads too highly, the target description is usually too broad. Words like B2B, enterprise or innovative mean different things to different people. The model has to guess, and the guesses show up as high scores for weak matches.

Give the model constraints. Name the company type, the size range, the role you want, the problem you solve and the kinds of companies you want to exclude. A short paragraph of plain language works better than a long list of adjectives.

```
Target profile example:

I want founders and heads of sales at B2B software companies in the US and Canada with 10 to 50 employees. They sell a product that needs outbound pipeline and they are likely handling lead quality themselves. Exclude recruiting agencies, staffing firms and companies that sell only to consumers. My product helps them find and contact LinkedIn leads without wasting invites on poor fits.
```

## Where scoring fits in your outbound workflow

Scoring is most useful before you write the note. The score tells you who to look at. The profile tells you what to say. If you skip the profile and write from the score alone, the note sounds like a template.

A simple workflow: score the list, read the top profiles, write one line about the work they do, then send. After a few replies, check whether the people who answered were near the top of the list or near the bottom. That tells you whether your fit line is in the right place.

Write the first line from the profile, not from the score. A score of 92 tells you the person is worth a look. It does not tell you what to say. The detail that made the model score them highly is the same detail you use to open the note.

Keep a simple record. When someone replies, note their score band. After ten replies, you will see whether the high scores are doing the work. If the replies come from the middle of the list, your description is catching the wrong signal.

- **you:** Saw you're building outbound at Northwind. Are you running LinkedIn and email together, or keeping email separate for now?
- **them:** LinkedIn mostly. Email gets ignored.
- **draft:** That matches what I see with small sales teams. If you want, I can send over how I score LinkedIn profiles before writing a note, so you are not spending invites on bad fits.

## Common mistakes that hurt your scores

Most scoring problems come from the input, not the model. A vague target description gives the model room to guess. A target you change every day makes it impossible to compare one week with the next. A lead you cannot write a specific first line to is not a good lead, even if the number is high.

- Keeping leads you cannot write to. If the profile gives you no detail for a first line, drop the lead.
- Using adjectives as requirements. Innovative and fast-growing do not tell the model who to find.
- Changing the target every day. Pick one description, run it for a week, then adjust.
- Treating the score as a promise. A high score means the profile matches. It does not mean the person will reply.

## Doing this with Kinlead

Kinlead finds people on LinkedIn who match the customer you describe or your website, then scores each lead from 0 to 100 against that profile. Leads below the fit line are not kept. Leads are stored in lead groups, and any campaign can point at any group, so you can keep a strict list and a test list apart. Lead search uses LinkedIn's standard search, so a Sales Navigator seat is not required. You can see the scoring setup in [lead groups and scoring](/features/lead-groups-and-scoring).

## Related

- [Book a demo](https://getkinlead.com/demo.md)
- [Lead scoring features](https://getkinlead.com/features/lead-groups-and-scoring.md)
- [Find leads](https://getkinlead.com/tools/find-leads.md)

## Common questions

**Can I change the fit line after leads are scored?**

Yes. Raise or lower it as you see which leads you can write to and which ones reply. The score stays the same; the line decides what you keep.

**Does AI scoring require Sales Navigator?**

No. Scoring works on the profile data LinkedIn shows in standard search. Sales Navigator adds filters, but it is not required for the scoring step.

**How does the AI handle profiles with little information?**

Sparse profiles usually score lower because the model has less evidence to compare. If the person is a strong fit in real life but has a thin profile, keep the lead and check manually.

**Should I use the same target description for every campaign?**

Not always. A founder campaign and an enterprise campaign may need different descriptions. Keep one base profile and add constraints for each campaign.

[Create a Kinlead account](https://getkinlead.com/signup) or [book a call](https://getkinlead.com/demo) to talk through pricing and setup.
