AI lead enrichment: from a bare list of names to leads worth calling
Add place details, then contacts, then an AI score with a reason. When to enrich a list once and when to let a flow keep enriching new leads as they arrive.
Part of the guide AI employee for sales prospecting
AI lead enrichment means taking a thin list of leads, often just a business name and a city, and adding what a salesperson needs before calling: where the business is, its phone, website and rating, the person to ask for, and finally a score with a sentence explaining it. The order matters. Facts first, people second, judgment last, because the AI can only score what the earlier steps found.
This guide is for anyone holding a spreadsheet from a trade fair, a purchased list, a CRM export or a form on their website. It walks through the three layers, shows how to tell a good enriched row from a bad one, works through a made-up example, and explains when to do it once and when to set up an AI employee that enriches every new lead as it comes in.
Layer one: the facts about the business
Most lists arrive with a name and something about where it is. The first job is to match each row to a real place. In xMap Flows, Map business scan can look up each incoming business by its name near its location and bring back its address, phone, website, rating, whether it is open, and its position. When the business is already in xMap's places database, Nearby places adds its brand, whether it is part of a chain, and its rating and number of reviews.
This layer also cleans the list. A row that matches no place is often a business that closed or moved. A row that matches a chain branch tells you the decision may sit at head office.
Layer two: the people
With a website on each row, the Apollo block can ask your own Apollo account for people at each company by job title (owner, operations manager, head of purchasing, whatever fits what you sell), and it can add the company's size, industry and the year it was founded. Size is the most useful extra field for scoring later: a team of four and a team of four hundred are different sales.
Expect gaps. Larger companies come back with several people; many small local businesses come back with nobody. Leave those rows in. They still have a phone number from layer one.
Layer three: the AI score and the reason
Now xMap LLM reads each enriched row and writes new fields. Tell it in plain words who your best customer is and what makes a poor fit, and ask for a fit score from 0 to 100 and a one-sentence reason. The reason is what makes the score usable: a rep who reads "independent, 4.8 rating from 300 ratings, opened this year, team of twelve" knows what to say on the call.
Add Check the work after it if the score must follow rules, for example that every reason names something real from the row, or that a lead without a phone can never score above 50.
Once, or every time a lead arrives?
| Enrich a list once | An enriching flow | |
|---|---|---|
| Best for | A list you already have: a fair, an export, a purchase | Leads that keep coming: a sheet, a form, a daily search |
| How you start it | Upload a CSV on My data and press Enrich a list | A trigger on new rows, or a schedule |
| What it adds | Place details, and an AI score with a reason if you ask | Anything a flow can do: details, contacts, scores, checks, routing |
| How many rows | Up to 100 rows that have a position | Each run's new leads, as many as the flow is built for |
| Where results go | A new table in My data and a CSV download | Email, Slack, a sheet, your CRM, the Team inbox |
The two ways to enrich leads in xMap Flows.
Enrich a list once shows its price before it starts, runs as a job with a progress bar, and saves the result next to the original as a new table. It is the quick answer for a list that will not grow. Anything that arrives every week deserves a flow, so nobody has to remember to upload it.
A made-up example
Everything here is invented to show the steps. An events company in Dubai collects 140 venue and caterer names at a hospitality fair, typed into a sheet in Google Sheets by three colleagues, some with a city and some with only a name.
- The sheet is brought into My data. Rows with an address or a city get a position; 22 rows with only a name are set aside for a person to check.
- Map business scan looks up each remaining business by name near its position and adds its phone, website, rating and whether it is open. Nine turn out to be closed.
- Apollo adds company size and asks for the general manager and the events or sales manager at each.
- xMap LLM scores each venue for fit (capacity for corporate events, rating, size of team) and writes a reason.
- Sort and limit puts the highest scores first, and the list goes to the sales team's sheet and as a short email to the manager.
The useful part was not the score alone. It was learning, before anyone called, that nine names were closed, that six were branches of one hotel group, and that the strongest leads were mid-sized venues no one at the fair had written notes about.
Mistakes that make enrichment useless
- Scoring before matching. An AI scoring a bare name invents a picture; score only after the facts are in.
- Trusting a match without its address. Keep the column that pins each row to a place.
- Dropping rows with no contact found. A phone number from the place details is still a way in.
- Scores with no reason. Nobody acts on a number they cannot explain to a customer.
- Enriching the same leads every week. Put Only new or Match with my table before the paid steps so only new rows are looked up.
The AI employee that keeps enriching
For a sheet that grows: New rows as the trigger, Map business scan to look up each place, Apollo for company size and contacts, xMap LLM for the score and reason, Check the work, then your CRM or Google Sheets and a short Email. What lands in the inbox each time is the new leads only, best first, each with its facts, a named person where one exists, and why it scored as it did. The templates Score new leads from a sheet in Google Sheets and Enrich a lead sheet and push it to HubSpot are built this way. More on building a pipeline of leads with AI employees is in the sales prospecting guide.
Questions
What is the difference between data enrichment and lead scoring?
Enrichment adds facts to each lead: address, phone, rating, size, contacts. Scoring judges the lead from those facts. Enrich first, then score.
Can the AI find details that are not in any source?
No, and it should not try. xMap LLM works from the fields the earlier blocks found and the reference files you give it. Check the work can refuse a reason that mentions something the row does not hold.
What happens to rows that match nothing?
In a one-off enrichment they stay in the new table without details, so you can see them. In a flow, a Filter can send them to a person to check instead of to the sales team.
Is my uploaded list shared with anyone?
No. Tables in My data belong to your organization, and the enriched copy is saved next to the original there.