AI Tools for Real Estate Agents: What to Use and What to Delegate

The question has moved. In the National Association of REALTORS® 2026 Technology Report, published in September 2026 from a survey of 1,200 agents, nearly half of agents now use AI daily or weekly. Twelve percent say they do not use it and have no plans to; a year earlier, 32% had not yet tried it at all.

So the useful question is no longer whether to use AI. It is which parts of your week it actually takes off you, and which parts only look like they move.

What Are Agents Actually Using AI For?

The same report asked AI users which tasks they use it on. The answers lean heavily one way:

• Writing listing descriptions — 75%
• Social media posts — 56%
• Emails and follow-ups — 52%
• Market summaries — 30%
• Marketing content in a personal tone — 30%
• Document review and summarizing — 27%
• Client education, FAQs and guides — 26%
• Pricing insights and comps support — 23%

The tools are concentrated too. Among agents who use AI, 91% use ChatGPT, 42% Gemini, 27% Copilot and 25% Claude.

Read that list and a pattern shows up immediately. The top of the list is writing tasks. Agents have adopted AI fastest exactly where the job is to produce words.

The More Interesting Half of the Same Data

Now look at the bottom of the list.

Lead generation and prioritization sits at 11%. At 10% or below: recommending the next steps in a transaction timeline, sending clients automatic status updates, summarizing disclosures and contracts, and auto-filling forms.

Those four have something in common, and it is not that nobody has thought of them. Each one needs to reach into a system — your MLS, your CRM, your transaction platform, your forms — and each one ends with something having happened rather than something having been written.

That is the real split, and the survey found it without looking for it. AI has been adopted fastest on the work it can finish by itself. Adoption stalls on the work that has to land somewhere, be checked by someone, and be followed up until it is done.

Which is, of course, the work that was eating your week.

The Same Pattern Outside Real Estate

It would be easy to read that as a real estate quirk. It isn’t.

Title and escrow. A Qualia survey of 378 title and escrow professionals, published through ALTA in July 2024, found 90% had already adopted generative AI in some form — and about 94% had no company-wide AI policy. Near-total adoption, almost no governance. The tasks named are the familiar ones: pulling data out of contracts, drafting client responses, summarizing reports, writing marketing copy. The closing file still has to be built, chased and kept current by somebody. More on title and escrow support.

Insurance agencies. The 2026 Agency Universe Study, from Future One — a collaboration of the Big “I” and leading independent agency companies — found 46% of agencies using AI, up from 15% in 2024. The top three uses: marketing content generation at 49%, coverage form analysis at 43%, contract reviews at 35%. Content and analysis again. And the most-cited industry challenge in that study wasn’t AI at all. It was finding and screening job candidates, at 45%, just ahead of keeping up with AI at 43%. More on insurance agency support.

Home services and the trades. Housecall Pro surveyed more than 400 contractors for its 2025 AI Adoption in Home Services report. More than 70% had tried AI; about 40% were actively using it. That gap is worth sitting with. Trying costs an afternoon. Using it week after week means somebody owns the follow-ups, the data entry and the emails it produces, and that is where the other thirty points went. The contractors who stuck with it reported saving over four hours a week on admin — so the value is real, and somebody had to be there to collect it. More on home services support.

Then the broadest figure of the lot. In the Census Bureau’s Business Trends and Outlook Survey, covering December 2025 to May 2026, 37% of firms with at least 250 employees reported using AI, against under 20% of firms with four or fewer. Over that period, use rose among firms with at least 20 employees and did not change significantly among the firms below that.

That is the whole problem in one line. The tools cost a two-person brokerage the same as a national firm. What the large firm has is somebody whose job it is to run them.

Where Does AI Earn Its Place?

None of this is an argument for using less of it. Four in five agents in that survey say they pick tools on whether they save time, and on these tasks AI does:

• The first draft of listing copy. Faster than a blank MLS field, and easy to shape.
• Social posts and newsletters. Volume and variations, which is what a content calendar needs.
• Market summaries for a farming area or a listing presentation.
• Long documents. A summary to orient you before you read the thing properly.
• Objection practice. Only 13% use AI for role-play, an underrated use — it costs nothing and nobody is watching.
• Client education. The buyer FAQ you have explained out loud four hundred times.

Where It Stops, According to NAR

NAR’s own broker guidance is direct about the limit: AI-generated content “should not go straight from prompt to publication.” If inaccurate information reaches a listing or a client communication, the agent and the brokerage remain responsible for it.

The title survey above suggests how rarely that review is written down anywhere: near-universal use, almost no company-wide AI policy.

The practical consequence is the thing that gets skipped: somebody has to read what AI produces before it goes out, against the source. That review is the agent’s judgment and the agent’s responsibility. It stays with you, and no assistant takes it over.

There is a second kind of failure, and it is quieter than a bad draft. A workflow you set up months ago and stopped thinking about — the lead routing from your portal, the drip that follows up after a showing, the reminder that fires before an option period ends — breaks without announcing it. Automation does not fail loudly. It stops, and because it was built to run without you, nobody is watching the place where it stopped. You find out when a lead says they filled out the form three weeks ago.

The harder version is a workflow still running that no longer does what it was built to do. Nothing errors, the campaign still sends, and the thing it was supposed to accomplish for the client at the other end quietly stopped happening. That needs somebody close enough to the week to notice a lead count that has not moved, and to know that is wrong.

What an assistant can take over is everything standing around it.

What Should You Hand to a Person?

The list is unglamorous, which is the point:

• Listing input into the MLS, media coordination, booking the photographer and chasing the files
• Disclosure packets assembled, sent out for signature, and chased until the last one is signed
• CRM upkeep: new leads entered the day they arrive, the database actually current, the same client merged out of three records
• Follow-up on last month’s leads, and the ones that went quiet in the spring
• Showing coordination, and the three reschedules after the inspector moves
• Transaction timelines: who owes what, and the reminder that goes out before the deadline rather than after it
• Marketing executed on a schedule instead of in the gaps

A Keller Williams Memorial client described the version of this that matters most, two weeks in:

“If I don’t make a phone call to somebody during the day, he makes sure it’s in front of me first thing the next morning that I still need to make that phone call.”
— Ken Foster, Keller Williams Memorial

That is not a prompt. Nothing was asked for. An assistant who has been inside the business for two weeks noticed what did not happen and put it back in front of the person who needed to do it.

The Listings You Win Are the Hours You Lose

There is a particular trap in this business, and it is not a technology problem.

The listings arriving this month are the return on the prospecting, the follow-up and the market presence you built two months ago. Servicing them takes exactly the hours that used to go into prospecting, follow-up and market presence. Win more and you have less time to win more. The pipeline fills and then the top of it quietly empties.

“I have the most listings I’ve ever had. And so now he is doing the follow-ups for me so that we can keep the pipeline going.”
— Melanne Carpenter, Commercial agent, Keller Williams Commercial

Two jobs, split between two people, is what makes that sentence possible. A better drafting tool does not do it, because the constraint was never the drafting.

The Part That Isn’t About Tools

And there is a reason the obvious fix has waited. A hire is a salary, plus payroll taxes, benefits and equipment, carried whether the quarter is good or not. If the work is clearly there and the money for another salary isn’t yet, the admin stays where it is, and the things that would grow the business — the follow-up, the database, the marketing, the process nobody has written down — wait behind the things that keep today running. You should not have to just endure that, and AI does not resolve it. It makes the writing faster and leaves the rest with you.

What resolves it is capacity. A virtual assistant or your first employee goes through that comparison in detail.

An Assistant Who Already Uses AI

The choice was never AI or a person. It is a person who uses AI.

Each VRA assistant completes the Essential Toolkit before their first client interview, and Claude and ChatGPT are part of it, alongside Canva, a CRM, Zapier and Asana — which is to say the most-used and the fourth-most-used AI tool among agents who use any. Assistants who need deeper AI skills for a client’s business can go further, through a dedicated AI track.

So the tools you were going to open yourself are already in the hands of the person doing the work, and what reaches you is the finished result, handled to your standards, rather than a draft you still have to act on.

That sits on top of the rest. Assistants come through screening, a mandatory four-week instructor-led program, and a final evaluation before they touch client work. 1,800 applications produced 16 graduating assistants, or 0.89%.

A VRA assistant administers the back office of a real estate business under your direction — and your license, your advice and your judgment stay exactly where they are. More on how that works on our real estate virtual assistant page, and on the wider question in should you just use AI instead of hiring a virtual assistant?

What Does It Cost?

A VRA assistant is $1,500 a month full-time, or $800 a month part-time for 20 hours a week. One flat rate, with no setup fee, no overage charges and no upcharge for a bilingual assistant. Increased coverage is available at any time.

Part-time exists for the situation above: the work is there before the business is ready to carry another salary.

AI Tools for Real Estate Agents: Common Questions

Who notices when an automated workflow stops working?

Somebody has to, and it is not the tool. Lead routing, drip campaigns and deadline reminders fail without announcing it, and because they were built to run unattended, nobody is watching the point where they stopped. The agent usually finds out from the lead who says they filled out the form three weeks ago. An assistant who works your pipeline every day is the person positioned to catch it, because they see the inquiries that were not answered and the numbers that have not moved.

What are the best AI tools for real estate agents?

The 2026 NAR Technology Report found that among agents who use AI, 91% use ChatGPT, 42% Gemini, 27% Copilot and 25% Claude. Which one matters less than what you point it at: the same report shows agents getting the most out of AI on writing tasks — listing descriptions, social posts, market summaries and client education — and the least on anything that has to happen inside the MLS, the CRM or the transaction file.

Can AI write my listing descriptions?

It can write the draft, and 75% of agents who use AI already have it do exactly that. NAR’s broker guidance is that AI content should not go from prompt to publication without review, and the agent and brokerage stay responsible for what is published. The draft is quick; the read-through is yours.

Will AI replace real estate virtual assistants?

The NAR data points the other way. The tasks where AI adoption is lowest — 10% or less — are recommending transaction timeline steps, client status updates, summarizing disclosures and contracts, and filling forms. Those are the tasks that need access to your systems and someone accountable for the outcome, which is what an assistant is for.

What should I delegate to a virtual assistant instead of using AI?

Anything that has to be sent, entered, tracked, chased, confirmed or followed up: MLS input and media coordination, disclosure packets chased to signature, CRM upkeep, follow-up on older leads, showing coordination and reschedules, transaction timelines, and marketing executed on a schedule.

Do VRA assistants use AI?

Yes. Claude and ChatGPT are part of the Essential Toolkit each assistant completes before their first client interview, and there is further AI training for assistants whose clients need it.

How much does a real estate virtual assistant cost?

$1,500 a month for full-time at 40 hours a week, or $800 a month part-time for 20 hours. One flat rate, with no setup fee and no overage charges, and increased coverage is available.

Most of this carries across to an investor’s portfolio, where the admin repeats on every property rather than every client.