> The Appraiser Coach  > AMC Guide > OREP E&O ★★★★★ “They are knowledgeable, professional, and understand urgency.” – Joe Thweatt

AO-41: TAF’s New Tech Guidance and What It Means

 by Isaac Peck, Publisher

Every appraiser reading this is using technology tools in their assignments. MLS platforms, adjustment software, sketch tools like CubiCasa, data aggregation services, and increasingly, AI-powered platforms like Spark that do more of the analytical lifting. The question is not whether appraisers will use technology. The question is: what happens when a tool gets it wrong, and who is responsible? In January 2026, the Appraisal Standards Board (ASB) at The Appraisal Foundation released an exposure draft of Advisory Opinion 41, Use of Technology in an Appraisal or Appraisal Review Assignment. AO-41 consolidates and replaces two existing Advisory Opinions: AO-18, which addressed AVMs in 1998, and AO-37, which addressed computer-assisted valuation tools in 2018. After two rounds of exposure and public comment, the ASB officially adopted AO-41 on April 23, 2026. AO-41 is now the interpretive framework that state boards, regulators, GSEs, and opposing counsel will use when evaluating how an appraiser used technology in any assignment that comes under scrutiny. Why AO-41 Exists The ASB was direct about why AO-41 was necessary. In the introduction to the exposure draft, the Board wrote that “it is insufficient for appraisers to maintain only the skills and knowledge they possessed when they entered the profession” and that appraisers “must continuously improve their skills to remain proficient; this is especially true when new technological tools are introduced into the appraisal workflow.” The existing guidance was built for a different era. AO-18 was published when AVMs were novel. AO-37 expanded the conversation to regression software and computer-assisted tools, but it predated generative AI, AI-powered adjustment software, and tools like CubiCasa. Neither anticipated the world appraisers are working in today. The profession needed a single framework covering everything from spreadsheets to machine learning. AO-41 is that framework; retiring AO-18 and AO-37 and replacing them with guidance addressing the full spectrum: AVMs, AI, regression software, sketch tools, data platforms, and generative AI. What AO-41 Actually Says AO-41 organizes its guidance around three elements: technology, the appraiser, and assignment results. The foundational statement is this: “A tool cannot comply with USPAP. The appraiser decides whether to use a tool and whether reliance on its output is appropriate.” That sentence, from lines 34 and 35, is the one around which the entire Advisory Opinion turns. If you take nothing else away from AO-41, take that. On the appraiser’s role, the draft states that “it is the appraiser’s judgment and not the tool that determines whether, and to what extent, reliance on a tool’s output is appropriate.” On assignment results, the language is equally direct: “Assignment results are the appraiser’s opinions or conclusions developed in an assignment,” and “the appraiser remains responsible for ensuring that the tool is used ethically and competently, in compliance with USPAP.” One of the most practically important clarifications concerns what appraisers are not required to know. In Illustration Question 2 of the draft, the ASB states that “in most instances, an appraiser is not required to replicate or fully understand the technical algorithms underlying a statistical tool.” But you do need “sufficient knowledge and experience to use the tool competently, which means applying judgment in every case to interpret its output and determine whether reliance on it is appropriate.” On disclosure, AO-41 makes an important clarification: USPAP does not require you to disclose tools simply because you used them. The obligation to disclose arises only when omission would make the report misleading. If you use a spreadsheet to run calculations, you do not need to name the software. If you rely substantially on a proprietary AI system to develop an adjustment, disclosure may be necessary so intended users understand how you arrived at your conclusions. On record keeping, AO-41 ties documentation to reliance. If a tool’s output was relied upon, your workfile must include the output, data used, any prompts or instructions, and enough to show how that output contributed to your conclusions. On confidentiality, AO-41 has some of its sharpest language: “Knowing that a system may inappropriately disclose or transmit confidential information and choosing to use it anyway demonstrates a lack of due care and may constitute gross negligence in violation of the ETHICS RULE.” Appraisers entering client data into generative AI platforms without understanding how those platforms handle data should take notice.

(story continues below)

(story continues)

 The Real Question: What Does “Competent” Mean? If there is one issue in AO-41 that drew the most comment, it is this: what level of understanding does an appraiser actually need before relying on a technology tool? AO-41 contains a genuine tension. On one hand, it states that appraisers “must also be competent to recognize when the design or training of advanced tools, such as generative AI, may reflect assumptions, limitations, or embedded biases introduced by their developers.” But later it says appraisers are “not required to replicate or fully understand the technical algorithms underlying a statistical tool.” So which is it? Can you be expected to recognize embedded bias in a tool whose algorithms you are not required to understand? The public comment letters made clear that a lot of people were asking that same question. The Appraisal Institute picked up on this. In its comment letter, signed by Amy McClellan, Chair of the Professional Standards & Guidance Committee, the Institute wrote that “this ambiguity creates uncertainty for appraisers working to comply with USPAP” and warned that it “increases the risk of inconsistent state enforcement.” James (Jim) Park, President of the Collateral Risk Network and a former Executive Director of the Appraisal Subcommittee, was more direct. AO-41 “overstates the degree of control and understanding individual appraisers can reasonably be expected to have over advanced technological systems,” Park wrote. Many modern valuation systems are “intentionally non-transparent, continuously learning, and protected by intellectual property restrictions,” and framing competency around a level of system understanding that is “functionally unattainable risks exposing appraisers to compliance and enforcement risk without providing them with meaningful control or insight.” Lee Kennedy, a certified appraiser and Managing Director of AVMetrics, supported AO-41’s overall direction but saw a different problem. “The most significant implementation risk I see is not misuse of technology, but unrealistic interpretations of competency when applied to opaque tools,” Kennedy wrote. He compared the situation to when multiple regression analysis was first introduced into appraisal education. The problem was not regression itself. It was that appraisers were encouraged to use it “without sufficient conceptual grounding in when the relationships made sense and when they did not.” Not everyone sees AO-41 as a step forward. Chris Daniel, an SRA-designated appraiser, called the draft “deeply flawed” and argued it is “too long, too general, and ultimately adds confusion rather than clarity.” David Samnick called it “regulatory noise, not guidance,” warning that the lack of an objective standard for sufficient understanding “invites after-the-fact enforcement based on outcomes, not violations.” NACVA raised a separate problem. The draft describes an Ethics Rule violation as an appraiser who “intentionally presents the tool’s output as their own assignment results.” NACVA pointed out that ethical violations under USPAP do not require intent, and recommended clarifying that presenting tool output as assignment results “whether intentionally or through lack of due care” may constitute a violation. Mark Schiffman, Executive Director of the Real Estate Valuation Advocacy Association (REVAA), an organization that represents the largest AMCs in the country, wrote in his public comment that AO-41 is “a common-sense approach.” He clarified, stating that “These tools don’t follow USPAP, appraisers do.” Meanwhile, on AppraisersForum.com, a thread on AO-41 has generated over 60 replies. The sharpest frustration centers on a perceived double standard: the GSEs offer liability protection to appraisers on hybrid and property data collector products, yet USPAP holds appraisers fully responsible for every tool output they rely on. As one poster put it: “It has to be all or nothing.” If a tool is reliable enough for a GSE-endorsed product, why does USPAP treat the appraiser as the sole backstop when that same tool is used in a traditional assignment? (story continues below) (story continues)   What This Means for Your Practice For all the debate, AO-41 is not creating new liability out of thin air. It is putting into writing a standard of due diligence that thoughtful appraisers already follow. An appraiser who verifies a CubiCasa measurement against county records before relying on it is already doing what AO-41 describes. So is one who cross-checks adjustment software output against their own paired sales analysis, or who documents in the workfile why a particular adjustment was accepted or rejected. Now that it’s adopted, that practice is an explicit benchmark. When a state board investigates a complaint, when a lender pushes back, or when a case ends up in litigation, AO-41 is part of the evidentiary framework. The question will not be whether the tool failed, it will be whether you exercised the professional judgment the Advisory Opinion requires. The practical takeaway is straightforward: document your reasoning. If you use a technology tool in an assignment, your workfile should show that you evaluated its output, considered its limitations, and made an independent judgment about reliance. From an E&O perspective, demonstrated judgment is always the first line of defense. AO-41 simply creates clearer expectations around what that documentation looks like. What to Watch AO-41 was adopted on April 23, 2026 after two rounds of exposure and public comment. What’s left to watch is how state boards, regulators, and the GSEs apply it in practice. The biggest unresolved question is what “competent” really means when the tool is opaque, continuously learning, or its internal processes are obscured by proprietary restrictions. That tension did not vanish with adoption, and neither did the related concern about how state enforcement actually plays out across jurisdictions. Both will surface in board complaints, enforcement actions, and litigation. Appraisers should also watch for FAQs or follow-on interpretive guidance from the ASB, which often arrives in the months after a new Advisory Opinion is issued. AO-41 and the comment record are available at AppraisalFoundation.org. The bottom line is the same one appraisers have always lived with. You sign the certification and you own the work. AO-41 extends that principle to an era where the tools are doing more of the analytical lifting. For appraisers who already use technology thoughtfully and document their reasoning, this should not require a major change. For appraisers who have been treating software outputs as gospel without a second look, this is a good time to reconsider. About the Author Isaac Peck is the Publisher of Working RE magazine and the President of OREP Insurance, a leading provider of E&O insurance for real estate professionals. OREP serves over 10,000 appraisers with comprehensive E&O coverage, competitive rates, and 14 hours of CE at no charge for OREP Members (CE not approved in IL or AK). Visit OREP.org to learn more. Reach Isaac at isaac@orep.org or (888) 347-5273. CA License #4116465.