Almost half the calls for every meeting booked.

The most expensive part of real estate sales isn’t the lead. It’s the salesperson’s time spent calling the wrong ones. Brigade Group’s salespeople were calling 40–50 leads to book 10 meetings. LatentIQ now runs Brigade’s lead pipeline end to end: we generate leads across every platform, including Instagram and LinkedIn, and our AI lead filter checks every one, scores it on spending capacity, preferred area, profile and authenticity, and removes junk before anyone picks up the phone. Qualified leads reach salespeople on WhatsApp. Three months into the pilot, with about 3,000 leads a month generated and filtered by LatentIQ, 4 out of every 10 leads called now book a meeting, so the team needs almost half the calls for every meeting booked.

• ClientBrigade Group
• IndustryReal estate
• What we doGenerate leads for Brigade across every platform, then filter them with AI so only qualified leads reach sales
• What it checksSpending capacity, preferred area, the lead’s profile (occupation and current situation), whether the lead is real or fake, and an overall lead score
• Lead sourcesAll platforms, including property portals, ads, Instagram and LinkedIn
• VolumeAbout 3,000 leads a month, generated by LatentIQ
• Delivery to salesQualified leads sent to salespeople on WhatsApp
• Pilot so far3 months, about 9,000 leads
• BeforeAbout 40–50 leads called to book 10 meetings
• After4 out of every 10 leads called book a meeting
• ImpactAlmost half the calls for every meeting booked
• Team2 people, who scoped, built and now run the whole pipeline
• StatusPilot

Brigade Group is a Bengaluru-based real estate developer. Like every large developer, it depends on a steady flow of leads for its sales team to call. In this pilot, LatentIQ generates those leads for Brigade, around 3,000 a month, from every platform: property portals, advertising, Instagram, LinkedIn and more.

In real estate, the sales call is where everything starts. A call leads to a meeting, a meeting to a site visit, and a site visit to a booking. So the number that matters most to a sales team isn’t how many leads arrive. It’s how many calls it takes to book a meeting.

Most leads aren’t ready to buy. Some are wrong numbers or duplicates. Some are people browsing with no budget or timeline. Some are enquiries that don’t match the project at all. These are common types of junk lead across real estate.

But until someone calls, every lead looks the same in a list. So salespeople called them all.

What it looked like before:

• Salespeople called about 40–50 leads to book 10 meetings, roughly 4–5 calls for every meeting.
• Most of the working day went on calls that led nowhere.
• Serious buyers waited in the same queue as junk leads, and in real estate, a buyer who waits often moves on to another project.

What it cost: the sales team’s time, which is the most expensive and least scalable resource in real estate sales, was being spent mostly on leads that were never going to buy.

The usual way to deal with bad leads is after the fact: a salesperson calls, discovers the lead is junk, and marks it. By then, the time is already spent.

The most expensive part of real estate sales is a salesperson’s time. So junk leads should be filtered out before anyone picks up the phone, not after.

That turned the problem around. Instead of asking salespeople to sort leads by calling them, we let AI sort the leads first, so salespeople only call the ones worth calling.

And because we generate the leads as well as filter them, we own the whole pipeline: what we learn from the filter, about which leads qualify and which don’t, comes from the same pipeline that brings the leads in.

Three principles shaped the filter:

1. Check every lead. No lead reaches sales without being assessed first.
2. Pass on only qualified leads. The sales list should contain leads worth a call, not everything that arrived.
3. Never throw a lead away silently. Every rejected lead is logged, so the team can review the filter’s decisions and nothing valuable disappears.

1. 01

   4.1 Generates leads across every platform

   LatentIQ generates leads for Brigade across every platform, including property portals, ads, Instagram and LinkedIn: about 3,000 leads a month.
   Who uses it: Brigade’s sales team receives the output, with no lead-generation work on their side.
   Why it matters: one partner is responsible for both the volume and the quality of leads, so nothing is lost between a lead agency and a filter.

4.2 Checks every lead on five signals. As each lead arrives, from any platform, the AI assesses it on five signals:

• Spending capacityWhether the lead can realistically afford the property. A buyer outside the budget won’t book, however interested.
• Preferred areaThe location the lead is likely to prefer. A lead looking elsewhere is a poor match for the project.
• ProfileWhat the lead does and their current situation. Shows whether this is a serious, ready buyer.
• Real or fakeWhether the lead is genuine. Fake and junk entries never reach a salesperson.
• Lead scoreOne overall score combining the signals. Decides which leads go to sales, and which first.

Based on these, the AI decides whether the lead is qualified and worth calling, or junk.
Who uses it: runs automatically on every incoming lead.
Why it matters: sorting happens in the background, not on the salesperson’s phone.

1. 01

   4.3 Filters out junk before it reaches sales

   Leads the AI identifies as junk are kept out of the sales team’s calling list.
   Who uses it: the sales team, whose list now holds fewer, better leads.
   Why it matters: salespeople stop spending their day on calls that were never going to convert.

2. 02

   4.4 Sends qualified leads to sales on WhatsApp

   Only leads that pass the filter go to the sales team, delivered straight to salespeople on WhatsApp, with no new software for them to learn.
   Who uses it: salespeople, on the phone they already use.
   Why it matters: serious buyers reach a salesperson sooner, instead of waiting behind junk leads.

3. 03

   4.5 Logs every rejected lead for review

   Every lead the filter rejects is logged, so the team can review it. On review, about 1 in 10 rejected leads turns out to be worth a call. Those leads are recovered and sent to sales instead of being lost.
   Who uses it: sales managers and the LatentIQ team running the pilot.
   Why it matters: no AI filter is perfect. Reviewing rejections means a good lead wrongly filtered out is caught, and the filter’s mistakes become lessons that improve it.

1. LatentIQ generates a lead on one of the platforms: a property portal, an ad, Instagram or LinkedIn.
2. The AI checks it for spending capacity, preferred area, profile and authenticity, and gives it a score.
3. If it’s junk, it is kept out of the calling list and logged for review.
4. If it’s qualified, it is sent to a salesperson on WhatsApp.
5. The salesperson calls a lead that is far more likely to book a meeting.
6. Rejected leads are reviewed. About 1 in 10 is recovered and sent to sales, and the filter’s decisions stay accountable.

1. 01

   Salesperson

   Before: calls 40–50 leads to book 10 meetings; most calls go nowhere.
   After: 4 of every 10 leads called book a meeting; more time with real buyers.

2. 02

   Sales manager

   Before: sees a busy team but few meetings per call.
   After: sees more meetings from the same calling effort, and can review every rejected lead.

3. 03

   Serious buyer

   Before: waits in the queue behind junk leads.
   After: reaches a salesperson sooner.

The pilot has run for 3 months, with about 3,000 leads a month, around 9,000 leads so far.

• Leads called to book 10 meetingsBefore: About 40–50 · With the AI lead filter: About 25
• Leads called that book a meetingBefore: About 2 in 10 · With the AI lead filter: 4 in 10
• Calls per meeting bookedBefore: About 4–5 · With the AI lead filter: About 2.5

The “about 25” and “2.5” figures are calculated from the pilot result of 4 meetings for every 10 leads called.

What this means in practice:

• Almost half the calls for every meeting. The same 10 meetings now take about 25 calls instead of 40–50.
• Close to double the meeting rate per call. From roughly 1 in 4–5 leads to 4 in 10.
• The same team can book more meetings. The calls no longer spent on junk leads become time for qualified buyers.
• Few good leads lost. Reviewing rejected leads recovers about 1 in 10 of them, so the filter saves time without quietly throwing away buyers.

A team of 2 people scoped the pipeline, built the filter and has run lead generation and filtering for the 3 months of the pilot, about 3,000 leads a month. Leads go to the sales team on WhatsApp, so salespeople needed no new tools or training. Keeping the team small and close to the work means the filter’s decisions are reviewed by the same people who built it, so the rejection log turns directly into improvements.

1. It fixed the most expensive step. It protected salespeople’s time, not just the lead list.
2. One partner owned the whole pipeline. Generating and filtering leads together means volume and quality are managed as one job, not handed between two vendors.
3. It acted before the call. Filtering before the phone is picked up saves the time; filtering after only records the waste.
4. It stayed accountable. Every rejected lead is logged and reviewed; about 1 in 10 is recovered, so the team can trust the filter and correct it.
5. It fitted how salespeople already work. Leads arrive on WhatsApp, so there was nothing new to adopt.
6. It started as a pilot. Proving the result on real leads first means any wider rollout is based on evidence.

Most developers measure lead volume. The better measure is calls per meeting booked, because that is where sales time goes. A filter that removes junk before the call doesn’t just clean a list. It gives a sales team back most of its day.

Don’t make your salespeople find the buyers. Let them talk to them.