Skip to content
SellingRoostHaven Team

California Seller Disclosures: Expect Buyers to Use AI

Buyers now do their due diligence with AI, asking it about every page of your disclosures. Here's what that AI reads and what it tends to catch.

Buying or selling, you, your agent and AI work it out in one chat. Try RoostHaven

If you're selling a house here, at some point your agent hands you a stack of California seller disclosures, and you spend a weekend trying to remember when the water heater was replaced. You sign, hand it all back, and figure buyers will skim it like everyone does.

Know what's in every offer as soon as it's submitted.

Ask Roost in the chat with your agent. It reads the terms of any offer on your home you can see.

MC
AO
Selling 412 Alder Way
3 members
@roost what's in the $2,210,000 offer?
Roost AI

A 3% deposit and 20% down with a loan, closing in 21 days. The appraisal contingency is waived, and it expires Thursday at noon.

MC
Maya Chen · Listing agent
It's the strongest so far. I'll walk you both through all three tonight.
Sample conversation
Create a free account

They might not anymore. A buyer can drop your whole package into an AI tool and ask whether the roof has ever leaked, whether the foundation's been looked at, and what the pest inspector found under the house. The answers come back in seconds, often with a page number.

AI has changed how buyers do their due diligence. So let's look at what it reads, what it tends to catch, and what it still can't do.

What AI reads in California seller disclosures

Every package starts with two required forms: the Transfer Disclosure Statement (TDS) and the Seller Property Questionnaire (SPQ). Together they're the seller's own account of the house, from the roof and plumbing to past repairs, and the SPQ goes into more detail than the TDS. Most packages also include a Natural Hazard Disclosure, which covers things like flood and fire hazard zones.

In the Bay Area there's usually more. Many packages include a home inspection and a pest inspection done before listing. Some add roof, sewer lateral or foundation reports, and condos and townhouses come with HOA documents.

It adds up to a lot of pages, and every one of them is now something a buyer can ask about.

How AI is changing due diligence

Due diligence on a Bay Area house has always been short on time. Disclosures go up, offers are due a few days later, and the package is too long to read closely in between. So buyers skimmed, leaned on their agent to point out what mattered, and counted on their inspector for the rest.

AI takes the time pressure out of the reading. A buyer can ask the whole package pointed questions the night it goes up and read the pages it points to. Comparing two houses? Same question, both packages, and a minute later they know which one has the older roof. A lot of the digging that used to wait for the inspection contingency can now happen before the offer is written.

Some buyers go further. In July, an agent told Inman that some buyers now upload inspection reports to a chatbot and negotiate off its repair estimates. Those numbers don't reflect local pricing, and they can blow a small fix up into a demand that puts the deal at risk.

AI also doesn't get tired on page 40. It reads the pest report as carefully as the TDS, and it's good at the thing people are worst at: noticing when two documents don't agree.

What AI tends to catch first

Documents that disagree are the big one. The TDS says no known leaks, and the inspection mentions old water staining under the kitchen sink. There may be a perfectly good explanation, but an AI tool will put those two lines side by side.

Work mentioned in passing comes next. A pest report notes a "converted garage," or a contractor invoice mentions a new bathroom, and the AI connects it to what the forms say about additions.

Then there are reports that get mentioned but aren't there. "See roof report," and there's no roof report. "Sewer lateral inspected 2023," and no sewer report. To a buyer, a missing document reads like a missing answer.

An "unknown" on a major system gets noticed too. Sometimes it's the honest answer, especially on an inherited house or one that was rented out for years. AI will still point it out, and buyers tend to follow up.

Finally, older systems: roof age, an original sewer lateral, an older electrical panel. None of these kill a sale. Buyers just want to know how old they are and whether anything's been done.

What AI still can't do

It reads documents, not houses. It can tell a buyer what the inspector wrote about the crawlspace, but it has never been in the crawlspace. The inspection is still the inspection.

It doesn't know local prices. That's the problem with chatbot repair estimates: a number that sounds precise can be far off from what a Bay Area contractor would actually charge.

It can misread a page. Good tools show where each answer came from, so anyone relying on one can check the source. Tools that don't show their work deserve less trust.

And it doesn't decide what any of it means. Whether a finding matters, and what it does to the price, is still worked out between buyers, sellers and their agents.

A 1962 ranch in Mountain View

Say you're selling a 1962 ranch, the kind you'll see among single-family homes in Mountain View. The garage became a family room in 2009 and the sewer lateral is original. You've been there twenty years, and you honestly can't remember when the roof was done.

Your agent uploads the TDS, the SPQ, the home and pest inspections and the Natural Hazard Disclosure, then runs an AI review before the listing goes live. Three things come back, each with the document, page and exact wording:

  • The pest report mentions a "converted garage with finished interior," and the AI flags it against what the forms say about additions.
  • The inspection recommends a sewer lateral video inspection, and there's no sewer report in the package.
  • The roof's age is listed as unknown, while the inspection estimates 15 to 20 years.

None of it is fatal. But a buyer asking questions would have found all three in the first ten minutes. This time you and your agent see them first, with the exact pages, and the two of you decide what to do about each one before anyone else reads the package.

The same kind of review works on older homes in Los Altos, Redwood City or among the San Jose houses under $1.5M.

How RoostHaven helps

A quick word about our own product. RoostHaven is a workspace where buyers, sellers and their agents, at any brokerage, work on a listing together, with Roost AI built in. This fall it won Inman's 2026 Best of Proptech award for Conversational AI and Virtual Assistants.

Your agent uploads the disclosures and reports to the listing and sets who can see each one: everyone, client only, listing parties only, or specific people. When a report gets added, everyone with access sees the current file, so there's no re-sending.

On Pro, the AI Report reviews the disclosure package and cites the document, page and exact quote for every finding. A second pass drops anything it can't find on that page. Your agent can run it before you list, and buyers with access can run it too, so everyone is looking at the same findings. Here's what's in Pro.

When buyers ask Roost AI about your listing, it answers from your documents and points to where it found each answer, so their answers come from the actual package rather than a guess. Once offers come in, your agent can set a deadline with a reminder, keep every offer and counter in one place, and have Roost AI compare them. Your agent still sends the counters.

There's more on selling with AI in how to sell your home with AI tools. Getting ready to list? Tell us about your home.

FAQ

How is AI changing due diligence when buying a house? Buyers can ask AI about every page of a disclosure package and read the pages it points to, so a lot of the close reading happens on the first night instead of the last week.

Can buyers use AI on my disclosures? Yes. Anyone who has your documents can ask an AI tool about them.

Will buyers negotiate off AI repair estimates? Some already do, an agent told Inman in July. Chatbot estimates don't reflect local pricing, so they tend to start a conversation between the agents rather than settle one.

What does an AI review of a disclosure package look for? Mostly consistency. It reads every document and points out where they disagree, where a report is mentioned but missing, and where an answer is left as unknown, with the page for each.

Does an AI review replace a home inspection? No. AI reads what's in the documents. An inspector looks at the house.

Sources

This post is general information, not legal advice. What to disclose and how to prepare a listing are decisions for you and your agent, and a real estate attorney can answer legal questions.

Keep exploring RoostHaven.