How to automate lead lists from three buying signals
New openings, companies hiring and businesses like your best customers each point to a different buyer. How to automate lead lists from each, and keep them from overlapping.
Part of the guide AI employee for sales prospecting
To automate lead lists, pick the signal that says a business is likely to buy now, give that signal its own flow that finds new matches on a timetable, drop the ones you already sell to and the ones you have seen before, score what is left, and deliver a short ranked list where your reps work. Three signals cover most local B2B selling: a business that just opened, a company that is hiring, and a business that looks like your best customers.
A bought list is a snapshot of everyone; an automated list is a stream of the businesses that changed in a way that matters to you. This article is for sales leaders and founders who sell to local businesses and want new names every week without someone searching by hand. It covers the three signals, how to keep the lists apart, a worked example and the mistakes that fill a CRM with names nobody calls. The sales prospecting guide covers the whole job.
Signal one: just opened
A business in its first months is setting up suppliers, software, insurance, signage and services, and has not settled on any of them. That makes a new opening one of the strongest reasons to call. Nearby places can keep only places opened in the last 6 or 12 months among the kinds you sell to, in the cities you name. Places only appear once they have a listing, so this finds openings within weeks, not on the first day.
Suits: payment terminals, food and drink distribution, cleaning, insurance, local marketing, point of sale.
Signal two: hiring
A company advertising for warehouse staff, store managers or buyers is growing or opening somewhere. Public job postings say who, for what and where. Job postings finds them by keywords and place, and keeps one posting per company so a firm hiring ten people counts once. The role tells you who to sell to: a company hiring drivers needs vehicles, uniforms and fuel cards; one hiring office staff needs desks and software.
Suits: staffing agencies, payroll and benefits, equipment and fleet suppliers, training.
Signal three: like your best customers
Your best customers have something in common: a kind of business, a size, a kind of street. Businesses near them that look the same are worth a call even with no event. Start from a table of your best customers in My data, search around each one for similar businesses, and let an AI step score how alike each is. The template More businesses like my best customers does this every month.
Suits: anyone with at least twenty good customers to learn from.
Keeping the lists apart and clean
Run each signal as its own flow, because each has its own pace: openings and hiring change daily, lookalikes monthly. Then make sure no list wastes a rep's time:
- Not your customers. Match with my table, set to keep only items not found in your customer table, removes current accounts. With a CRM, CRM lookup does the same against the CRM itself.
- Not seen before. Only new drops anything this flow has sent before, so each business is offered once.
- Not twice across lists. A business that opened and is hiring can appear in two lists. Send all of them to the same sheet or CRM, where a duplicate shows up beside the original, or put the signals into one flow with Remove duplicates before the end.
- Short. An xMap LLM block scores each lead for your offer with one line on why, and Sort and limit keeps the best few.
A worked example
This example is made up. A company that leases delivery vans in Mexico City wants companies that will soon need more vans. Openings are a weak signal for it, because most new shops do not run vans. Hiring is a strong one: a company advertising for drivers is adding routes.
- A Schedule runs every weekday at 07:00, Mexico City time.
- Job postings looks for drivers and delivery staff posted in the last 24 hours in Mexico City, one posting per company.
- Only new drops companies already sent, and Match with my table drops companies on the current customer list.
- An xMap LLM block scores each company from 0 to 100 as a van leasing prospect, from the role and the kind of company, with one sentence on why.
- Sort and limit keeps the eight best, and Email sends them to the two reps with the posting linked for each.
Each morning the reps get a handful of companies, each with a reason to call that day: they advertised for drivers yesterday. To add the person to ask for, the Apollo block adds names and job titles at each company from the team's own Apollo account. The ready-made version for staffing is New companies hiring, daily.
Mistakes that fill a CRM with names nobody calls
- Every signal for every product. Pick the one or two signals that predict a sale for what you sell. A van leasing firm does not need every new cafe.
- Lists without a reason. A name alone gets ignored. Each lead should carry the signal that put it there, so the first line of the call writes itself.
- Too many a day. If reps can call ten, send ten. The rest of a long list goes stale before anyone reaches it.
- Not removing customers. Calling a current customer as a prospect is embarrassing. Match against the customer list in every flow.
- Pushing straight into the CRM. For the first weeks, send the list by email and read it. When the scores match your judgement, connect the CRM, or put an Approval before it.
Each flow runs as an AI employee with its own card on the Flows page and its own thread in the Team inbox, where it writes a daily note of what it found. For the openings signal, start from New places every morning, only new; to keep CRM accounts out, see Prospects not yet in my CRM.
Questions
Which signal finds the best leads?
The one tied to what you sell. New openings suit suppliers of things every new business sets up; hiring suits firms that sell to growing companies; lookalikes suit teams with many good customers to learn from.
How do I stop sending leads that are already customers?
Use Match with my table set to keep only items not in your customer table, or CRM lookup against your CRM, before the list is sent.
How many leads should an automated list send?
As many as your reps can act on that day or week. Sort and limit keeps the highest scores and drops the rest.
Can the list include who to contact?
Yes. The Apollo block adds names and job titles at each company from your own Apollo account.