Failure to Disclose: The Claim Behind 70% of Agent Lawsuits

If there is one mistake that consistently lands real estate agents in court, it is failing to tell buyers what they already know about a property. According to CRES Insurance, a leading E&O provider, failure to disclose known property defects drives roughly 70% of real estate E&O lawsuits arising from sales transactions — and one veteran defense attorney believes that figure is "probably low." The National Association of Realtors reaches the same conclusion, calling failure to disclose the top claim against agents.

The Case That Changed Everything

The modern duty to disclose traces back to a single California lawsuit. In Easton v. Strassberger (1984), a buyer purchased a home for $170,000, only to watch landslides gut its value to around $20,000, with repair estimates topping $213,000. The jury awarded $197,000, and the California Court of Appeals established something new: brokers have an affirmative duty to conduct a reasonably competent and diligent inspection and disclose facts that materially affect a property's value.

That ruling effectively ended caveat emptor — "buyer beware" — in residential real estate. Today, more than 90% of U.S. transactions involve home inspectors, and over 30 states license them, a direct legacy of the disclosure standard Easton created.

Water Intrusion: The Defect That Guarantees a Lawsuit

Not all defects carry equal legal risk. Attorney Mark Carlson, who has defended real estate professionals since 1993, puts it bluntly in the CRES disclosure analysis: "There's nothing like water intrusion to guarantee a lawsuit." Legal analysis of failure-to-disclose claims confirms that water infiltration in basements is the most commonly alleged defect in these suits.

The money at stake is real:

• A court ordered one agent to pay $170,000 after finding "reckless disregard for the truth" for concealing prior water damage.
• The Easton jury award of $197,000 exceeded the home's original sale price.
• In Wisconsin, theft-by-fraud claims can bring triple damages plus attorney's fees — and homeowners' insurance typically won't cover defense costs.
• E&O policies cover negligent non-disclosure only; agents who intentionally conceal defects face uncovered personal liability.

That last point deserves emphasis. An E&O policy is a safety net for honest mistakes, not a shield for deliberate silence. An agent who hides a known defect isn't just risking a lawsuit — they're risking one their insurer won't defend.

The Disclosure Duty Is Expanding to AI

The same legal logic — you are responsible for what your tools communicate — now extends to artificial intelligence. The California DRE's advisory on AI in real estate states that if an AI tool generates misleading communications, responsibility rests with the licensee and broker, not the technology provider. California's new AI image disclosure law took effect January 1, 2026, and NAR warns that "speed without guardrails creates risk" for brokerages adopting AI.

This is why disclosure-by-design matters when choosing AI tools. CallMyLeads, for example, builds honest AI disclosure into every interaction — callers always know they're talking to AI, and every caller can reach a human. In a legal environment where the agent bears all the liability for what AI says, that transparency isn't a courtesy. It's compliance.

Why Good Agents Still Get Sued Over Disclosure

Even the most careful agent can land in a disclosure lawsuit through no dishonest act of their own. The uncomfortable truth: a survey cited by NAR found that 60% of sellers admit they didn't disclose a known problem with their property to buyers.

That statistic creates a trap. Agents get sued not because they lied, but because a seller did — and the buyer's attorney argues the agent should have caught it. Wisconsin legal analysis notes that agents face conflict-of-interest scrutiny because they "may be motivated to finalize the sale as effortlessly as possible."

The stakes are severe. In Wisconsin, sellers (and their agents) can face triple damages plus attorney's fees under theft-by-fraud claims — and homeowners' insurance typically won't cover defense costs. And E&O policies cover only negligent non-disclosure; fraud and intentional misrepresentation are excluded, leaving agents with uncovered personal liability.

So how do good agents protect themselves? Defense attorneys who've spent decades in this arena offer concrete guidance:

• Never fill out a seller's disclosure form — that's the seller's job, and completing it for them transfers their liability onto you.
• Probe anything that "looks wrong." Fresh paint on one ceiling patch or a musty basement should trigger questions, not assumptions.
• Document everything — every conversation, every seller refusal, every disclosure you made independently.
• Walk away from bad listings. As one defense attorney puts it, sometimes it's about the listings you don't take.

The same disclosure discipline now extends to AI. California's Department of Real Estate has advised licensees that if an AI tool generates inaccurate information or improper consumer communications, responsibility rests with the licensee and broker — not the technology provider. That's why services like CallMyLeads build disclosure in from the start: callers always know they're talking to AI, and every caller can reach a human. NAR's own guidance is blunt: speed without guardrails creates risk.

The pattern is the same whether the disclosure involves a leaky basement or a chatbot: the agent owns the outcome. Diligence, documentation, and honest tools are the only reliable defense.

The New Disclosure Risk: AI You're Using Without Guardrails

Disclosure risk no longer stops at water stains and cracked foundations. Regulators are now treating undisclosed or unsupervised AI the same way they treat an unlicensed assistant doing licensed work — and the liability lands squarely on the agent.

California's Department of Real Estate, which oversees roughly 434,000 licensees, has issued an official advisory on AI in real estate with a blunt message: using AI changes nothing about your duties. Only licensed persons may perform licensed activities, brokers must supervise everything under their license, and fiduciary duties remain intact.

The most consequential line for any agent using AI tools: if an AI tool generates inaccurate information, misleading advertising, or improper consumer communications, responsibility rests with the licensee and their broker — not the technology provider. The DRE compares unsupervised AI use to asking an unlicensed assistant to do licensed activity, a violation of California real estate law.

The disclosure mandates are getting specific, too:

• Effective January 1, 2026, Bus. & Prof. Code § 10140.8 requires clear disclosure when listing images have been digitally altered in ways that change a property's appearance — including AI enhancement — with the original image made available to consumers.
• First-point-of-contact rules under § 10140.6 apply to AI-generated solicitation content, meaning your name, license ID, and responsible broker must appear even when AI writes the outreach.
• Failing to review AI output can trigger discipline for breaching fiduciary duties, negligence, misrepresentation, and missed disclosures.
• AI provides no shield from liability under fair housing and lending laws, including FEHA and the Holden Act.

NAR frames the same problem at the brokerage level. Its guidance on AI use policies identifies four core risk categories: accuracy (hallucinated square footage or features), fair housing (biased training data producing discriminatory language or steering), privacy (client data fed into unapproved tools), and licensing. NAR's warning is direct: "Speed without guardrails creates risk." AI cannot make agency decisions, provide legal advice, or independently determine property value without human oversight.

This matters most where AI adoption is already happening — lead response. Agents increasingly use AI to answer new inquiries, which is precisely where first-contact disclosure and supervision rules bite. An AI that pretends to be human, or that improvises answers nobody approved, creates exactly the exposure the DRE describes.

The compliant model looks different: AI that always identifies itself as AI, works only from approved scripts and client-set response rules, and hands off to a human whenever the caller wants one. That's the framework behind honest AI disclosure as a feature, not a courtesy — every caller knows what they're talking to, every conversation runs within boundaries the licensee controls, and client data stays in the client's own CRM rather than leaking into unsecured tools.

The parallel to traditional disclosure law is hard to miss. Just as the E&O claims data shows agents pay for what they failed to reveal about a property, regulators now expect the same transparency about the technology doing the talking. The agents who treat AI as a supervised, disclosed support tool — as the DRE puts it, "a support tool rather than an independent decision-maker" — are the ones who keep the speed without inheriting the lawsuit.

A Practical Compliance Playbook for AI-Powered Lead Response

Fast lead response and legal compliance are not opposites — but only if you build the safeguards in from day one. NAR's warning that "speed without guardrails creates risk" applies directly to AI-powered lead response, because agents are already using AI to reply to leads, and regulators are watching.

Here's how to capture speed-to-lead without creating the kind of liability that already drives an estimated 70% of E&O lawsuits in disclosure cases.

1. Always disclose that callers are talking to AI. California's DRE — which oversees roughly 434,000 licensees — makes clear that responsibility for AI-generated communications rests with the licensee and broker, not the technology provider. Hiding the AI doesn't hide the liability; it compounds it. Honest disclosure up front is your first line of defense.

2. Keep humans in the loop with clear escalation paths. The DRE instructs licensees to treat AI as "a support tool rather than an independent decision-maker," and NAR stresses that AI cannot make agency decisions or provide legal advice without human oversight. Every caller should be able to reach a human, switch to text, or book online — no dead ends.

3. Use only approved scripts and client-set response rules. Unsupervised AI output is what regulators compare to "asking an unlicensed assistant to do licensed activity." Your AI should answer from approved FAQs, qualify leads against criteria you define, and route conversations to your team at the moments you choose — never improvise on pricing, property details, or legal questions.

4. Protect client data end to end. The DRE explicitly warns against feeding confidential client information into unsecured AI tools. Your leads, your data, and your calendar should stay yours — flowing into your own CRM, not a vendor's black box.

5. Collect explicit consent and honor opt-outs automatically. Booking flows should gather clear consent, and any opt-out must take effect immediately — not after a manual review. US carrier rules (A2P 10DLC) and telemarketing quiet-hours laws make this non-negotiable for business texting.

In practice, a compliant AI lead-response setup looks like this:

• AI introduces itself as AI on every call and message
• Human escalation available at any point in the conversation
• Responses limited to approved scripts and client-defined rules
• Explicit consent captured in the booking flow
• Opt-outs honored instantly and automatically

This is exactly how CallMyLeads builds its done-for-you lead response: callers always know they're talking to AI, every plan runs on client-set response rules with routing to your team, consent is collected in the booking flow, and opt-outs are automatic. The result is speed with guardrails — every new lead answered in seconds, 24/7/365, without the compliance exposure regulators are now enforcing.

Because when AI goes wrong, the law is unambiguous about who pays: you do. Choose tools that were built assuming that from the start.