A buyer browsing listings late at night will often send inquiries to several teams at once, and the first team to call back usually wins the client. During business hours, most agents can keep up with that pace. After hours and on weekends, many of those inquiries sit until the next workday or never get a reply, according to Prakhar Gupta, founder of Mergro, a company that builds voice AI agents for real estate teams. For Gupta, who works with teams in the U.S., the U.K., Canada, and the UAE, that gap decides which teams should automate their first call and which should keep it in human hands.
Gupta came to real estate from software. An engineer and graduate of the Indian Institute of Technology, he previously co-founded LazyApply, a job application automation tool used by more than 50,000 people, and served as its CTO. He started Mergro after a frustrating property search in London, where he and a friend inquired about a number of listings and heard nothing back for more than a week. The company began working with U.K. teams in January 2026 and is now concentrating on the U.S., particularly New York and Texas.
The teams Gupta now speaks with describe the same breakdown he experienced as a buyer, seen from the other side. Leads arrive in the CRM, and agents work through them with a qualifying script covering financing, budget, and preferred areas. The script is rarely the issue. Capacity is. “They are not able to handle the volume,” Gupta says. “They are not able to call back every lead.”
That delay matters because buyers rarely inquire with just one team. In Gupta’s experience, those who are actively looking fill out multiple forms, and the first team to reach them usually keeps them. “The first ring is always better than a better scripted call,” he says. He points to published research on lead response time, including work from Harvard Business School, and says what he sees with teams is consistent with it.
One recent team of roughly 15 to 20 agents, drawing mostly on Zillow and Facebook Marketplace, reported a typical response time of six to eight hours. After it routed a portion of its leads to automated callbacks, Gupta says bookings roughly doubled, and the booking rate rose from about 4% to 9%.
Many teams remain skeptical about handing that first call to AI in a business built on relationships, and Gupta does not dismiss the concern. “Human touch is important, that is for sure,” he says. “Nothing works without human touch.”
The shortfall is widest outside business hours. On weekdays, Gupta says, most teams keep up with inquiries until around 6 or 7 p.m. The trouble starts with the lead that comes in at 11 p.m. on a Sunday. By his estimate, 40% to 50% of leads that arrive after hours or on weekends are never answered.
Those buyers are often the least willing to wait. “Buyers are very aggressive,” Gupta says. Many submit several inquiries at the same time and expect someone to show them a property the next day, so a callback that arrives the following afternoon may already be too late.
When buyers do pick up, most already know what they can spend. “The budget is very fixed in their mind always,” Gupta says. He estimates that roughly 60% of the buyers reached are asking about one or two specific listings and are already within range for them. The other 40% or so are searching more broadly, looking for a home in a city such as Austin, with a more flexible budget and often a need to arrange financing.
Where the lead comes from also appears to shape how serious the buyer is. Gupta says he has not yet broken out the numbers, but inquiries from a team’s own website tend to convert better than portal leads, because those buyers have already researched the team and seen the listing before reaching out. “Zillow is a general thing,” he says.
Across the four countries where Gupta works, buyers ask largely the same questions, in part because each team qualifies leads with its own fixed script. What changes is how they prefer to be reached. In the UAE, follow-up over WhatsApp tends to work best. In the U.K., a text sent after a missed call usually draws a reply. Canada, he says, is roughly an even split between calls and texts.
In the U.S., the phone call still leads. “Mostly phone calls,” Gupta says, though he adds that a text is worth sending because some buyers reply that way instead. A typical cadence he describes starts with a second call 15 minutes after a missed one, then continues over the next two to three days, with texts and emails in between. Any reply is logged in the team’s CRM so agents can see which leads need action.
Not every team benefits. Early on, Gupta tested automated callbacks with very small teams handling around 10 leads a day, and the approach did not hold up. “The manual reach out is much easier than our AI taking the call,” he says. With so few leads, each one carries more weight, and agents can reasonably handle them by hand.
He is direct about where the product does not fit. Without a CRM and enough lead volume, there is little to automate. “This is where Mergro is not useful at all,” he says. He now considers roughly 500 leads a month a reasonable starting point. The pressure is greatest, in his experience, among teams of up to about 50 agents that receive more inquiries than their agents can cover and keep buying more.
Old leads offer only a partial fix. Many teams carry three or four months of inquiries that were never called, and Gupta has tested what happens when those contacts are finally reached. “Most of the buyers have already purchased, moved to another vendor,” he says. By his estimate, only one or two of every 100 turn out to be active buyers who have not yet bought.
For team leaders, Gupta argues, the more useful question is not how many leads to buy but how well existing ones are worked. The most common response he hears to a conversion problem is to buy more leads. Hiring inside sales agents to keep pace carries its own cost, since they are typically paid a salary plus a percentage of each sale. A team running seven of them, he says, could save considerably by tightening follow-up and cutting back to two.
“Instead of buying more leads, you need to work on your follow-up process more,” Gupta says. “You need to convert those leads you already have.” He allows that buying more sometimes works, but says it has not been the answer for the teams he works with.
About the Expert: Prakhar Gupta is the founder of Mergro, a company that builds voice AI agents to call back and qualify leads for real estate teams in the U.S., the U.K., Canada, and the UAE.
This article is based on information provided by the expert source cited above. It is intended for general informational purposes only and does not constitute legal, financial, or real estate advice. Readers should conduct their own research and consult qualified professionals before making any real estate or financial decisions.
Explore similar articles from Our Team of Experts.
The 1031 exchange is one of the most powerful tax-deferral tools in commercial real estate, and one of the most likely to collapse before closing. Between 50 and 60 percent of 1031 exchanges fail, according to Scott Abeel, who spent six years as a commercial retail broker in Texas before launching Exchango, a platform connecting ...
Across Central Texas, a construction defect hidden inside backyard pools is ending home sales before they close. The problem is alkali-silica reaction, or ASR, commonly called concrete cancer. According to Bill Cafferata, a real estate agent with CaffGroup at Keller Williams Realty in Dripping Springs, most buyers and many agents have never heard of it. ...
Investors are supposed to dominate the market for distressed and estate properties. They have the cash, the contractor networks, and the tolerance for risk that families typically lack. In Oakland’s East Bay market, that assumption does not hold, according to Delores Johnson, an independent broker who has worked Oakland real estate for 24 years. Probate ...
The economics of building new homes in Southern California have shifted in ways that aren’t visible from a listing page. Development costs for turning raw land into finished lots have more than doubled over the past two decades in parts of the Inland Empire – from roughly $55,000–$60,000 per lot to $110,000–$115,000 – while home ...
A $400,000 home with $5,000 a year in property taxes looks like a bargain to a buyer arriving from New York or New Jersey – and like a barrier to someone who already lives and works in Pennsylvania’s Lehigh Valley. That gap in perspective is repricing the region. Out-of-state buyers with higher incomes and different ...
Buyers in the short-term rental cabin market around Broken Bow and Hochatown, Oklahoma, have less leverage than they did a year ago. In 2025, active listings climbed past 325. With that much choice, buyers could negotiate roughly 6% to 10% off the asking price, with the high end reserved for extreme cases, by the rough ...
