November 9, 2026 · 7 min read
NAR's 2026 Tech Report: How Agents Actually Use AI, Task by Task
NAR's 2026 REALTORS Technology Report shows agents use AI mostly for writing. Here is each number, set against 2025, and what it means for your office training plan.
According to NAR's 2026 REALTORS Technology Report, agents use AI mostly for writing. Listing descriptions lead at 75% of AI users, then social media posts at 56% and emails and follow-up at 52%. Nearly half of agents use AI daily (23%) or weekly (25%), and just 12% say they have no plans to use it.
That is a useful picture for a broker, because it tells you where your agents are already working with AI, with or without a policy. Our page on AI for real estate marketing covers what we think AI is good for. This post goes through the NAR numbers one at a time and turns each into a decision.
This year the story is routine, not adoption. The question for an office is no longer "should agents use AI." It is "which tasks, with what checks, and who trains them."
How often agents use AI, 2025 against 2026
NAR's 2025 Technology Survey, as its press release on the survey put it, found that 20% of REALTORS used AI daily, 22% weekly, 27% a few times a month, and 32% had not yet used AI in their business.
NAR's press release on the 2026 report says nearly half of agents now use AI daily (23%) or weekly (25%), and just 12% say they are not using it and have no plans to.
Read those side by side and two things stand out:
Daily use moved from 20% to 23%, and weekly use from 22% to 25%. Each is up three points. That is steady growth, not a jump.
The group that was not using AI shrank from 32% to 12%. Be careful: the 2025 line says "had not yet used AI" and the 2026 line says "not using it and have no plans to." The wording differs, so treat it as direction, not a measured drop.
The practical reading: plan on about half your agents already having a habit. That changes the job from persuasion to standards. One note on sample size: NAR's news coverage says the 2026 report is based on responses from 1,200 agents, so quote the figures as what those agents reported.
The task ladder: what agents use AI for
NAR's 2026 REALTORS Technology Report page states that the top AI use case among REALTORS is writing listing descriptions, cited by 75% of AI users. NAR's coverage ranks social media posts next at 56% and emails and follow-up at 52%.
A caution: the report page words the second and third numbers as shares of REALTORS, while the ranking is framed around AI users. Do not repeat them as "56% of all agents." Say "social posts ranked second at 56%." The order is the finding.
The three top tasks share a shape. Each is a first draft that a person edits, and each puts text in front of a client or the public.
That is where AI fits best, and where it fails in a predictable way. A description drafted from a few bullets says what every listing says. A post drafted with no brief sounds like every other agent's. A follow-up with no context reads like a form letter. The tool is fast. The sameness is the cost.
What each number means for a training plan
Here is how we turn each of the top three uses into a training rule.
Listing descriptions (75%). Teach a facts-first method. The agent pastes in the verified facts, the AI drafts, and a person who has seen the home finishes it. Add a fair housing check before anything goes live. We wrote up that routine in AI listing descriptions without the fair housing risk, and the listing description prompt is free.
Social media posts (56%). Teach a voice guide, not a clever prompt. If ten agents prompt ten ways, your feed has ten voices. Our post on the voice guide that fixes AI copy shows how one page of rules keeps the output consistent.
Emails and follow-up (52%). Teach a context rule. Every draft starts with what the person asked for, which page they came from, and what has already been said. A follow-up that knows the context gets a reply. A generic one gets deleted.
The work agents leave alone
NAR's coverage also lists the least common uses, cited by 10% or fewer respondents. They include summarizing disclosures and contracts and providing pricing or risk insights. For a broker that gap is a protection. Those tasks touch contracts and client data and need a licensed person in the loop, so spend the first training quarter on the three writing tasks.
NAR also reports that fifty-four percent of real estate pros expressed interest in using generative AI for content creation over the next 12 months, so expect more drafts.
Is it paying off?
NAR's news coverage of the 2026 report says 55% of agents who use AI report a positive impact on their business. In 2025, NAR's survey found that 17% reported a significantly positive impact, 33% a moderately positive impact, and 46% said AI had no noticeable impact.
Do not stack those into a trend line, because the question bases differ. What you can say is that most AI users in 2026 report some payoff, where a year ago nearly half noticed none.
Why do agents say they want it? NAR's 2026 release found that 81% said saving time is their primary goal in embracing new technology, up from 66% a year ago. The same release found that 63% of agents named the learning curve as their biggest challenge, and 59% pointed to cost.
Put those together and you get the training brief. Agents want time back and say learning is the hard part. So training should be short, tied to a task they do every week, and done with them.
Which tools agents use
Among agents who use AI, NAR's coverage reports that 91% use ChatGPT, 42% use Gemini, 27% use Copilot, 25% use Claude and 14% use Apple Intelligence. The shares add up to well over 100%, so agents use more than one tool.
You cannot assume everyone is on the same tool, so your standards have to work in any of them. A voice guide, a facts-first rule and a fair housing check are tool-neutral by design.
A 45-minute audit you can run this week
You can turn this report into a plan without hiring anyone.
Send agents a three-question form: how often do you use AI, for which tasks, and with which tool?
Sort the answers against NAR's ladder. Do your agents match the 75, 56 and 52 pattern, or are they further along?
Pull five recent listing descriptions, five social posts and five follow-up emails. Mark any that read like a template, contain a claim nobody verified, or describe people instead of the home.
Pick the weakest of the three tasks. Write one rule for it: the facts, the voice or the context.
Hold one 30-minute session on that rule only. Show a bad draft and a fixed one.
Re-pull five samples in 30 days and compare.
In practice: if descriptions are the weak spot, the rule might be "no draft goes live until the agent adds two details only someone who walked the house would know." A small rule, and the difference between a draft and a description.
How Urban Marketing Edge handles this
We use AI to move faster on drafts and variations, and we keep strategy and taste human. Every post goes through our post approval system, so nothing publishes unseen. Monthly content themes and brand rules are the context every draft starts from, the same discipline we ask agents to adopt.
For the training side, our Lunch and Learn catalog includes The AI Fast Track, a run of 90-minute hands-on sessions that takes an office from curious to capable. If you want one person in the office to own this, our 12 Week Program trains an in-house marketer, with AI workflows in the curriculum. Our AI guide is a good place to start.
Questions brokers ask
What are real estate agents using AI for in 2026?
According to NAR's 2026 REALTORS Technology Report, the top use is writing listing descriptions, cited by 75% of AI users. Social media posts come next at 56%, then emails and follow-up at 52%. Writing leads, and the higher-stakes work with contracts and pricing is used far less.
How many agents use AI every day?
NAR's 2026 release says 23% of agents use AI daily and 25% use it weekly, which NAR describes as nearly half. Just 12% say they are not using it and have no plans to.
Does the 2026 report show that AI works for agents?
It shows that most users say so. NAR's coverage reports that 55% of agents who use AI say it has had a positive impact on their business. That is agent opinion, not a measured result, so track your own office's numbers.
Which AI tool do most agents use?
Among agents who use AI, NAR's coverage reports ChatGPT at 91%, Gemini at 42%, Copilot at 27%, Claude at 25% and Apple Intelligence at 14%. Agents use more than one tool, so write standards that work in any of them.
What should a brokerage train first?
Start with the three writing tasks at the top of NAR's ladder. Teach one check for each: verified facts, a written voice guide and the context of the lead. Short sessions tied to weekly work beat a long course.
Book a 30-minute strategy call
If your agents are already using AI and you have not told them how, a short call can close that gap. We will look at where your listing copy, social posts and follow-up are today, pick the one task that needs a rule first, and map the simplest training and posting plan around it. No pitch deck, no pressure.
Book a 30-minute strategy call and bring five recent drafts. Or compare how our plans handle content and approvals before we talk.
