Lofty, the Phoenix-based maker of CRM and agentic AI software for real estate agents, on Nov. 12 put out the October edition of a monthly scorecard of the work its AI does for agents. The Lofty AI Index says the company’s AI tools completed 1.13 million tasks in October 2026 and saved agents 131,248 hours of lead follow-ups, social posts, homeowner check-ins and CRM updates.
The company announced the index in a GlobeNewswire release that bills it as the industry’s first AI index. The numbers come from Lofty’s own platform records and Lofty’s own estimates of how long each task would take a person, so they describe one vendor’s customers, not the real estate industry as a whole. Even so, they offer a look at what agents actually hand to AI software, as opposed to what they tell surveys they plan to do.
How the Lofty AI Index counts hours saved
The method has three steps, according to the index page. Lofty counts each task its AI completed during the month from platform records, such as messages sent, posts published and calls connected. It then assigns each task an estimate of how long an agent would need to do it by hand, with a low, middle and high value. Multiplying counts by minutes gives the hours saved, and the headline figure uses the middle estimate.
That choice matters. On Lofty’s low estimate, October’s savings were 69,920 hours; on its high estimate, 206,368. The page does not list the minute values assigned to individual tasks, and it does not describe any outside review of the figures.
Lofty says it tries not to inflate the count. Lead follow-ups and homeowner emails are counted once per lead or homeowner per month, not once per message; social posts count once per channel; and test and Lofty staff accounts are excluded. The period covered is Oct. 1 to Oct. 31, 2026.
Where the hours went
Lofty’s Sales Agent, which the company describes as a virtual inside sales assistant, accounted for the largest share: 56,018 hours. The index says 387,824 leads got an AI follow-up by text or email, 152,487 morning briefs were delivered to agents, and 118,010 lead records were updated. The AI also handled 7,944 phone conversations and booked 1,698 appointments, and 20,922 leads replied to it.
The Social Agent, which writes and publishes posts, saved 31,474 hours, according to Lofty. It published 115,585 posts across agents’ channels in October, 52,561 of them showcasing listings, and created 1,468 AI listing videos.
The Homeowner Agent, which keeps in touch with past clients, saved 28,885 hours. It sent 258,009 emails to 153,239 homeowners, filled in 38,755 property records with value, equity and mortgage data, and handed 695 homeowners back to their agents as possible sellers. Cowork, the plain-language assistant Lofty expanded earlier this year as part of its agentic AI platform, saved 14,871 hours; its most common jobs were looking up or filtering leads (26,757 times) and writing and sending emails and texts (20,744).
The four figures add up to the 131,248-hour headline.
Small teams lean in hardest
The more revealing numbers are in the breakdown by account size. Lofty says 74% of agents on independent small teams, which it defines as two to 10 agents outside a large brokerage, had its AI do work for them in October, the highest share of any group. The figure was 70% for solo agents, 68% for mid-size teams and brokerages, 60% for small teams inside large brokerages and 55% for large brokerages of 100 or more seats.
What each group delegates differs, too. Solo agents rely most on the Sales Agent, which produced 57% of their hours saved, and 77% of solo agents using AI received a daily morning brief. Large brokerages lean most on social content: the Social Agent produced 35% of their hours saved, and Lofty says 29% came from it even after leaving out reposts of office listings, more than in any other group.
On our reading, that split fits how the businesses are built. A solo agent has no assistant to chase leads, while a large brokerage’s problem is keeping hundreds of agents visible online. Lofty’s data only shows what its own customers did, but it points to the same conclusion for vendors: the AI jobs that stick are the dull, repetitive ones.
A baseline, then a trend
Lofty says it will publish the index monthly. Its site describes September 2026 as the starting line, with 111,246 hours saved that month. By our math, October’s total is about 18% higher. Some of that could reflect more customers using the tools rather than more work per agent, and the index page does not separate the two.
The index lands as brokerages and software vendors race to put AI agents in front of real estate professionals, from Realtor.com’s RealAssist to Real’s token-priced Leo Pro agents. Most of those launches have come with promises about time saved. Lofty has chosen to publish a number every month, and its methodology, with all its vendor-defined assumptions, is now on the record for customers and competitors to test. More coverage of the beat is in our Technology & AI section.
