When Compass recently debuted a new AI assistant to 83,000 real estate agents, the company’s senior leadership used every sales tactic in their arsenal to encourage adoption.
The real-estate brokerage held a demo day to announce the launch of the new tool, called AI Assistant, which was watched live by more than 10,000 Compass agents. CEO Robert Reffkin, President of Growth Rory Golod, Chief Technology Officer Shay Artzi, and President Neda Navab were among the top executives who presented the new tool, alongside agents who had used it during the pilot and feedback phase.
Since then, senior leaders have visited the company’s local offices to promote AI Assistant. Corporate staff located at these regional offices were also trained on how to teach agents how to use the tool, which was designed to handle repetitive administrative tasks and nudge agents to pursue new sales opportunities.
“It takes a lot of work to get agents to actually adopt these tools,” says Golod.
Compass reports that within weeks of when it debuted its AI Assistant in July, about 15,000 agents generated more than 97,000 conversations.
Nine out of every ten agents across the entire industry report that they use AI today or are planning to do so, according to the trade association National Association of Realtors (NAR), with more than two-thirds reporting that they save at least one hour of work each week using AI. And yet, the NAR reports that the median age of a real estate agent is 57, presenting a challenge for adoption. Older Americans use AI chatbots less for work than younger employees, according to Pew Research.
Brokerages also have a very mixed track record of embracing new technologies, so there is some historical skepticism that tech-focused firms like Compass have to confront. They were slow to embrace the internet and mobile applications, leaving a big market opening for online real estate marketplaces like Zillow and Redfin to respond to homebuyers moving to digital channels at a faster clip than agents.
Another risk that looms is from the democratization of the services that real estate agents provide through the AI systems themselves. There are early indications of some home buyers and sellers experimenting with chatbots for do-it-yourself transactions to save money on fees.
One physical therapist assistant reported saving thousands using an AI platform for a home contract, while a man in Florida says he used ChatGPT to close on the sale of his home in five days for $100,000 more than what real estate agents estimated the property would fetch. Earlier this year, a technology reporter at the New York Times reported that he used AI almost entirely to sell a home in upstate New York.
Yet another issue involving AI in real estate is emerging regulation. In New York City, there’s a proposal that would require landlords, brokers, and online listing platforms to disclose when they’ve used AI to generate or digitally alter rental listing photos or videos.
Golod says he believes AI and technology are not able to replace the experience of an agent going into a seller’s living room, seeing the property and having a face-to-face conversation with them, and outlining a plan to sell it at the highest price possible with the least amount of hassle.
“To actually sell a home properly and help a buyer buy a home properly, it requires incredible judgment, high EQ, and relationships,” says Golod. “Those are just not things that AI has the capability of driving right now, and maybe not for a very, very long time, if ever.”
Over the past few years, building an AI platform internally at Compass has evolved as AI models have advanced. An earlier version of the AI tool, called Compass AI, mostly focused on giving agents ways to use the technology for content creation, including the generation of marketing materials and social media posts.
But AI Assistant was conceptualized to act more like an actual human assistant and take real actions, including proactively producing a list of prospective clients that may be ready to sell a home and drafting a personalized note to reach out and reconnect. The tool is also voice-enabled, allowing agents to dictate notes while driving between appointments. AI can also compile a list of potential homes for a homebuyer, called Compass Collections, and share this data with agents and their customers.
“We moved in a similar path to the industry, moving from chat to agentic,” says Artzi. “We saw the value is really in agentic, and so it took us a while, because you want the AI to be aware of all the platforms of tools.”
AI Assistant connects 100 different real estate tools within a single platform, but Compass says it was designed in a manner that means agents just need to know how to prompt the AI tool to take an action, rather than understand any Compass-specific branding or software jargon.
“We’re starting to see this amazing shift where more and more agents, who in the past used parts of the platform but not all of it, are now really starting to take full advantage,” says Golod. “It’s going to drive a lot more usage.”
John Kell
NEWS PACKETS
Anthropic reportedly will delay its IPO to November. The Wall Street Journal reports that the Claude maker will launch its IPO a little later in the fall than originally anticipated, less than two weeks after CEO Dario Amodei wrote an essay that called for an industry slowdown in AI development to ensure the technology’s safety. WSJ reports that some advisors pushed for the new timeline to give Anthropic time to disclose its third-quarter financials, hoping that these results would show the company remains in a strong position against archrival OpenAI. To that end, Reuters reports that Anthropic is mulling releasing a new AI model ahead of the IPO in response to OpenAI’s GPT-6 Astra, which has gained traction with enterprise customers that have been central to Anthropic’s success in AI thus far.
Calls for AI slowdown result in a lawsuit that alleges the agreement is illegal. On Friday, a lawsuit filed in the U.S. District Court for the Northern District of California alleged that Anthropic, OpenAI, Google, and SpaceXAI made an “illegal” deal that violated antitrust laws when they agreed to work together on slowdown efforts for the development of AI. The lawsuit alleges that doing so would reduce the value that customers get for the services they pay for in the form of AI subscriptions. Fortune reports that the plaintiffs in the lawsuit say they aren’t against these AI giants asking Congress, the White House, or any other agency to develop regulations for AI, nor are they against these firms seeking an antitrust exemption.
Trump announces a new “AI Force,” but with few details. Over the weekend, President Donald Trump pledged that he would form an “AI Force” and hire an AI czar as he continues to push back against calls within the industry for more guardrails. There is very little known about the composition of an AI Force or what its mandate would be, though Trump continues to stress the importance of AI to the American economy and his desire to remain competitive with China. Axios reports that executives are already privately angling to get tapped as Trump’s AI czar, and has published a short list of names to watch. Politico, meanwhile, reports that technology industry insiders weren’t consulted before the administration announced its new approach and believe that, at least for now, the plan lacks focus.
Disney appoints first-ever CTO. Disney has appointed Karandeep Anand to the newly created role of CTO, effective October 2, in which he will report directly to CEO Josh D’Amaro and oversee enterprise technology, infrastructure, data and AI platforms, and engineering. Anand joins Disney from generative AI chatbot platform Character.AI, where he served as CEO, and will bring some of the startup’s technical team members with him to join the entertainment and media conglomerate. The two companies have scuffled in the past, with Deadline noting that Disney sent Character.AI a cease-and-desist letter in September 2025, accusing the startup of unauthorized use of Darth Vader, Spider-Man, and other Disney characters.
Meta may have a hit on its hands with Muse. Over the past few years, Meta CEO Mark Zuckerberg has faced criticism that the Facebook and Instagram company’s AI strategy has lacked a clear vision compared to peers like Alphabet. But, early signs show that the company’s new AI agent, Muse, has gained traction with consumers, as it ranks as the top free app on Apple’s iOS App Store and Google Play stores in the U.S. Shares of chipmakers including Advanced Micro Devices and Intel soared higher on Monday, as investors expect more demand for agents to also increase the need for chips. Still, Amazon did rain on Muse’s parade a bit, as it blocked the agent from its retail site because it has a policy that prevents autonomous tools from shopping on Amazon.com.
ADOPTION CURVE
AI is bolstering the top line for insurers, but many still lack enterprise-wide integration and training. Insurers have boasted that they are seeing meaningful revenue growth from their AI and data initiatives, with 81% of 263 senior insurance executives surveyed by consulting giant Accenture reporting that these technologies have improved underwriting cycle time and accuracy as well as growth in written premiums. Yet just about one in four insurance executives say that their company has achieved enterprise-wide AI integration and only 14% have launched targeted AI upskilling for all of their workers.
Accenture says that the AI use cases that insurers have prioritized have been focused on claims and underwriting, reflecting the outsized importance of those parts of the business to the industry’s total revenue generation. But in the process, they may be delaying the more difficult work required to support data, infrastructure, and organizational changes necessary to ensure insurance companies can re-engineer their businesses for AI.
Ravi Malhotra, a senior managing director at Accenture and co-author of the report, tells Fortune that leaders need to clearly explain to their workforce how AI is changing workflows, and specifically, share the metrics showing how these new tools can generate value.
“The challenge is not solved by technology alone,” says Malhotra. He adds that industry leaders must connect their AI vision to strategy more deliberately and “translate that into leadership alignment and expectations of outcomes.”
Courtesy of Accenture
JOBS RADAR
Hiring:
-
Eskaton is seeking a chief data and technology officer, based in Carmichael, California. Posted salary range: $250K-$355K/year.
-
Guardian is seeking a head of Park Avenue wealth management technology, based in New York. Posted salary range: $178.7K-$293.6K/year.
-
University of Wisconsin-Green Bay is seeking a chief data and technology strategist, based in Green Bay, Wisconsin. Posted salary: minimum of $170K/year.
-
Transamerica is seeking a head of life and annuities technology, based in Philadelphia. Posted salary range: $300K-$350K/year. Hired:
-
Arch Insurance appointed Gaurav Sharma as CIO, effective immediately, where he will lead the insurance company’s technology modernization agenda. Prior to joining Arch, Sharma was head of technology at health insurance provider AXA Health. He also previously held senior technology roles at RSA Insurance UK, Zurich Insurance and Aspen Re.
-
Chubb announced the appointment of Sean Ringsted as chief scientist, where he will lead the insurance company’s global AI, data, and analytics strategy and initiatives. Ringsted has worked at Chubb for 22 years, initially joining in 2004 as chief actuary. He also previously served as chief digital officer and the chief risk officer.
-
BD announced the appointment of Arthur (Art) Hu as chief information and digital officer, effective September 30. Hu joins the medical technology company from personal-computer manufacturer Lenovo Group, where he most recently served as CIO. He also previously spent over a decade at the consultancy McKinsey.
-
American Express Global Business Travel promoted Michael Esquibel to the role of CTO. He initially joined the corporate travel management company in 2018 as a vice president of technology strategy and operations. Prior to joining Amex GBT, Esquibel held various technology leadership roles for eight years at payments network Western Union.
-
Quantum named Mike Heuer to the role of CIO. Heuer joins the data storage company from Legence, an engineering and maintenance services company, where he served as CIO. He also previously served as head of global IT at Blackhawk Network and held tech leadership roles at Peet's Coffee, Blue Shield of California, and Applied Materials.
-
K-Love appointed Debashree "Dee" Sethmajhi to serve as CTO, where she will lead the technology strategy for the Christian radio network. Previously, Sethmajhi spent nearly a decade at software giant Oracle, where she served as a senior technical program manager.
-
A Place for Mom named Jason Conley as CTO, where he will lead the technology strategy, engineering, and data organization for the senior care referral service provider. Most recently, Conley served as CTO at specialized medical clinics group VIP Medical Group. He also previously served as CTO and VP of product, engineering, and technology at Envera Health.
-
EverCommerce announced several executive appointments, including naming Chris Rogers to serve as the software-as-a-service company’s CTO. Previously, Rogers served as CTO of online security brokerage Instinet. Earlier in his career, he founded Sidewalk Software.
Sign upfor free.