Skip to main content
GullySystem
AI for Sales · Solutions

Draw up a list of likely buyers from your own records and public listings.

AI Lead Generation studies the customers you have already won, then searches directories, tender notices and your own website visits for companies that look like them. Each name arrives with a reason. A salesperson reviews the list and accepts or rejects every row, and nobody outside your business is contacted until that decision is made.

What it is

Prospecting is often one salesman scrolling a directory with no rule for who deserves a call, so good names get skipped and the same dozen firms get rung twice. The list is rebuilt every quarter.

Here the rule is written down once, taken from your own closed orders, and the software proposes names that fit it with the reason beside each. Your team keeps the final say. They know which of those firms owe you money, which once threw a salesman out, and which are cousins of a competitor.

Where it goes wrong today

The problems this is meant to ease.

Prospecting by whoever has time

One salesman scrolls a directory with no rule for who deserves a call. Good names get skipped, and the same dozen firms get rung twice.

No written picture of the ideal buyer

Each executive carries a private idea of a good customer. The owner cannot say which trades or towns bring the orders that matter.

Enquiries scattered across channels

A form mail, a chat message and a missed call reach three different people. Nobody holds the whole list, so some buyers wait a week for a reply.

Money spent without a source record

A directory subscription and a trade fair stand are paid for every year. Of the names those bring in, no one can count how many reached a quote.

Follow-up that depends on memory

A promised call after the first contact is noted on paper or not at all. The lead cools while the salesman moves on to the next visit.

Inside the module

What AI lead generation does.

A profile built from your wins

Your own orders set the rule. Recent orders are read for trade, size, town and items bought, and the common pattern is shown in plain words before the software uses it.

Names from permitted sources

Directory listings, tender notices, trade fair exhibitor lists and registers you already have access to are scanned for firms that match. Basic details such as town and trade are added to each row.

Signs of buying interest

A new branch, a tender notice or a visit to your price page is attached to the row. The reason is written as a short fact that a salesperson can check.

Enquiry capture on your channels

Website forms, a chat box and the WhatsApp number on your van all feed one list. Every enquiry keeps the channel it came through.

Ranked, with a draft opener

Rows are ordered by how closely they resemble past buyers. Beside each one sits a draft first message in your wording. The salesperson edits it or throws it away.

Accept, reject or park

The queue clears row by row. A rejection asks for a short reason, and those reasons are read back when the next batch of names is drawn up.

Existing contacts held back

Firms already in the CRM, on a do-not-contact list or owing an old invoice are removed before the list reaches anyone.

CRM, reminders and attribution

Each accepted row turns into a CRM lead. It carries a date for the next call and a source tag saying it was researched. Later reports compare these leads with portal enquiries, referrals and walk-ins.

Where it gets used

What this looks like in a working week.

Machinery manufacturer

The profile is drawn from past lathe and press orders. Tender notices and exhibitor lists are scanned for workshops that look similar, and the sales head accepts the ones in reach of a service visit.

Building material supplier

New contractor registrations and site boards in the permitted listings are matched against past bulk buyers. Each name carries the reason, such as a second site opened in the same district.

Software company

Pricing page visits tied to a company are listed with the pages read. A marketing coordinator decides which visitors deserve a personal email and which are students or rivals.

Interior design studio

Enquiries from the website form, an Instagram link and a showroom visit land in one list. The studio compares which channel produced the briefs that became signed projects.

Industrial tools distributor

Trades that bought drill bits and abrasives over the past two years form the base list. Reps accept a dozen names each Monday and note why they dropped the rest.

Who touches it

The people who open this every day.

Sales executive

Opens a short queue each morning. Accepts the firms worth a call and drops the rest, writing a one-line reason for each rejection so the next batch improves.

Sales head

Reads the profile it drew. Corrects it when the software describes the wrong customer, for example a one-off bulk buyer from last year who will never order again.

Marketing coordinator

Sees which visitors were matched. Decides which of those companies deserve a personal email and which should be left alone because it was a student, a rival or a supplier checking prices.

Owner

Reads one monthly count. It shows how many researched leads were accepted and how many of those became quotes or orders, set beside the same counts for walk-in leads.

How it is set up

From first conversation to daily use.

  1. Discovery with the sales head

    We sit with the sales head and two executives to learn how names are found today. We list the directories, tender notices and forms you use, and the ones you are allowed to use.

  2. Getting closed orders and enquiries ready

    Closed orders are checked for trade, town, size and items bought. Gaps are filled by the team, and duplicates in the CRM are merged before any profile is drawn.

  3. Trial on one trade or town

    The profile is applied to a single trade or a single town. It is a narrow test. It produces a short list from your own orders, with the reason beside each name.

  4. Checking the list with the salesmen

    The executives accept and reject names in front of the sales head. Every rejection reason is read aloud, and the profile is corrected where it described the wrong customer.

  5. Going live on the channels you choose

    Forms, the chat box and the accepted-names queue are switched on for the whole team. Contact limits and the do-not-contact list are confirmed before the first message is drafted.

  6. Reviewing after a few cycles

    After several batches, the sales head compares accepted names with quotes and orders by source. Weak sources are dropped, and the profile is redrawn with the new closed orders.

Measuring it

How you would know it is working.

Record where you stand before it goes live, so the change is measured against your own figures.

  • Qualified leads are counted as accepted names and enquiries that a salesperson marks as worth a quote, set against the monthly count before go-live.
  • Cost per lead is the spend on each source divided by the leads it produced, using last year’s directory and fair bills as the baseline.
  • Lead-to-opportunity conversion is the share of accepted leads that reach a quote, measured against the same share for portal enquiries in the CRM.
  • Accepted-to-rejected ratio shows how well the profile fits, and it starts from the first batch your team reviews.
  • Time to first contact is read from the lead’s creation and the first logged call, against the gap in your current records.
Used responsibly

Where a person stays in charge.

A person accepts every name

No message is sent, no call is made and no profile is stored for an outsider until a salesperson accepts the row. Rejected names are not contacted.

Permitted sources only

Only listings and notices your business may lawfully use are scanned, and each row shows where it came from. Countries treat business contact and visitor data differently. Your own lawyer should confirm what is allowed. Treat the tool as a finder, not a legal adviser.

Thin data gives thin lists

With few closed orders, or buyers who come by referral, the profile will be vague. The page says so early, and we advise a hand-written profile in that case.

Do-not-contact lists are kept

Customers who asked not to be rung, and firms in dispute, stay out of every list until a person clears them.

Enrichment can be stale

Public details such as a firm’s size or town may be out of date. The row shows the source date so the salesperson can check before calling.

How it fits

It is one part of a system, not an island.

A module earns its place by what it passes to the next one. These are the connections we set up most often.

  • Lead management in your CRM
  • Opportunity management
  • Your website forms and analytics
  • Trade directories and tender notices
  • WhatsApp or email for later contact
Common questions

Questions owners ask about AI lead generation.

How many closed orders does it need before the list is useful?

It needs a few hundred closed orders with trade, town and items recorded, because fewer than that give it nothing to learn from. Below that, write the profile by hand.

Does it contact anyone by itself?

No contact is made. It proposes names and drafts an opener, a person accepts the name, and nothing is sent, rung or posted until someone on your team decides to reach out.

What if our best customers came through relatives and referrals?

Then the pattern it finds will be thin, because a referral leaves no trace in a directory or on a website, and we would tell you that early. This suits firms whose buyers are findable.

Is it lawful to use public listings and visitor data this way?

Contacting businesses and storing visitor details are governed by local law, which varies by country. Your own lawyer or adviser should settle the sources and the consent wording. This page gives no legal advice.

What does it connect to?

It reads and writes the lead records in your CRM. Website forms, a chat box and WhatsApp need your own accounts, and some directories offer feeds only on certain plans.

Who corrects it when it suggests the wrong kind of firm?

The sales head edits the profile in plain words, and every rejection reason from the team is read back. The software never changes the profile without a person seeing the change.

When is it not worth the effort?

When one person sells to a small circle of known buyers, a diary does the job. The return begins once several people prospect and the source of each order is unknown.

How much work is it to start?

Cleaning closed orders and agreeing the profile with your sales head take the most time. Our quote follows your sources and channels, and scope is fixed once we have seen your records.

Talk to us

Tell us how you handle AI lead generation today.

A spreadsheet, a register or another system: say which, and we will tell you plainly what is worth changing.

  • No obligation
  • A reply within one business day
  • Your details stay private

Start with your requirements

Seven short steps that tell us what you need. An engineer reads them before getting in touch.

Share your requirements