In the National Association of Realtors’ 2026 Technology Report, only about 12% of agents say they have no plans to use AI, and one brokerage survey found the share of firms with no AI plans has fallen to under 2%. But adoption and impact are different things. Most of that use is shallow. AI writes a listing description or a social post, and the deal still moves at its old pace. The change that matters now is the move from AI that writes to AI that acts.
This piece lays out where the market stands, and where it is going.
Almost universally. NAR’s 2026 report puts daily AI use at 23% of agents, weekly use at 25%, and occasional use at 31%, with only 12% having no plans to adopt it. “Do you use AI?” has stopped being the useful question. Nearly everyone does.
The depth is where it gets interesting. The single most common use is writing listing descriptions, cited by 75% of AI users, followed by social media posts (56%) and emails (52%). Those are marketing tasks. Fewer than a quarter of agents reach for AI every day.
Because it is pointed at the margins of the job. Writing a description faster saves minutes, not deals. The work that actually moves a transaction — qualifying a buyer, booking the viewing, completing the disclosure paperwork, keeping compliance in order — still happens by hand. The AI drafts text beside the workflow instead of doing the workflow. That is the honest state of the market in 2026: adopted eagerly, used narrowly.
Agentic AI completes tasks on your behalf instead of handing you content to use. An AI tool drafts a listing description. An AI agent creates the listing, books the viewing, and fills the related forms inside the system. The difference is execution.
This is the frontier, and it is early. One 2026 estimate puts agentic-AI adoption in real estate at roughly 14%, far below the near-universal use of generative tools, and Morgan Stanley expects agentic AI to reach mainstream use between 2026 and 2027. The gap between using AI and letting it act is where the next advantage sits.
It takes the task off your plate instead of handing you a draft. An AI agent built for real estate can create a listing and complete its paperwork, schedule a viewing, and answer plain-language questions about a property from the system’s own data. Work the agent used to do becomes work the agent reviews.
This is the thinking behind our own agent, AIDA, built to act inside the brokerage system rather than suggest text next to it. It has generated descriptions for more than 40,000 listings and now handles tasks such as creating listings and booking viewings. As Linear’s CEO Miro Eriksson puts it: “The real value of AI is created when it is connected to data and integrated into the agent’s daily tools.” We raise it here as one example of the shift. The broader point holds whatever software you use: value is moving from generation to execution. (More on how that works on our AI page.)
Three things are coming into focus for 2026 and 2027. AI stops being a writing assistant and becomes an operating layer that does tasks rather than produce text. The data underneath starts to matter more than the model on top, because an agent is only as useful as the brokerage data it can reach. And compliance joins the AI conversation, because the routine anti-money-laundering checks tightening across the EU through 2027 are exactly the repetitive work an agent should handle.
The practical move for an agency is to stop measuring AI by how much text it produces and start measuring it by how much work it removes. Look for AI that acts, that is grounded in real transaction data, and that treats compliance as part of the job. That is what separates the agencies using AI from the few getting a real result from it.
In NAR’s 2026 Technology Report, 23% of agents use AI daily, 25% weekly, and 31% occasionally, while only about 12% have no plans to adopt it. That is roughly half using it weekly or daily.
The most common uses are writing listing descriptions (75% of AI users), social media posts (56%), and emails (52%) — mostly marketing tasks rather than deal-moving workflows.
Agentic AI completes tasks on the agent’s behalf, such as creating listings, booking viewings, and completing paperwork inside the system, rather than just drafting text. Real-estate adoption is still early, estimated around 14%.
Because most use is concentrated in low-leverage marketing tasks. Only about a quarter of agents use AI daily, and agentic AI that handles real workflows is still around 14% adoption.
AI that acts rather than only writes, that is grounded in real brokerage data, and that handles compliance — the areas where AI turns time saved into results.
Nearly every agency now has AI in some form, so simply using it is no longer an advantage. The edge over the next two years belongs to the ones who let it do the work instead of just writing about it. If you want to see what an AI agent that acts looks like in practice, you can try Linear free.