MindBridge’s Response to the PCAOB’s Strategic Priorities on AI Audit 

The future of audit will be shaped not only by new technologies but by the standards and oversight frameworks that govern how those technologies are applied. 

MindBridge recently submitted formal comments to the Public Company Accounting Oversight Board (PCAOB) in response to its Request for Public Comment on Strategic Priorities (PCAOB No. 2026-001). The submission reflects our perspective as a company working at the intersection of AI, financial analysis, and audit methodology. 

Our comments focus on how the PCAOB can help the profession responsibly adopt AI-enabled audit procedures while continuing to advance audit quality, professional skepticism, and investor protection. They also highlight areas where additional guidance could help firms confidently implement technology-assisted audit approaches as AI becomes increasingly embedded in audit workflows. 

We believe thoughtful collaboration between regulators, audit firms, and technology providers is essential to ensuring that innovation strengthens—not compromises—the integrity of the audit profession. 

Below is the full text of MindBridge’s submission to the PCAOB. 

Re: PCAOB No. 2026-001 — Request for Public Comment on Strategic Priorities 

Dear Chairman Logothetis and Members of the Board, 

MindBridge Inc. welcomes the opportunity to comment on the PCAOB’s request for public input on its strategic priorities. As a company whose platform applies machine learning and AI to the analysis of financial transactions in support of the audit process, we sit at a practical intersection that is directly relevant to several of the questions you have raised. We offer these comments in the spirit of constructive partnership with the Board as it charts its next chapter. 

We want to begin by acknowledging what Chairman Logothetis described as a return to basics. We share that instinct. But we would add one qualification: the basics themselves have changed. The foundational obligations of audit quality, investor protection, and professional skepticism remain constant. What has changed is the environment in which those obligations are discharged. 

Technology is not a supplement to good auditing. It is increasingly the means by which good auditing is achieved. A strategic plan that engages seriously with that shift will be better positioned to deliver durable long-term improvements to audit quality. 

We offer the following perspectives, organized around the PCAOB’s seven thematic questions, along with a brief offer of further engagement at the close. 

1. Strategic Priorities in Registration, Inspections, and Enforcement 

The PCAOB’s core mission has not changed, but the operating environment for that mission has. Audit procedures that were once performed through sampling and manual review can now be conducted across entire populations of transactions using AI-driven platforms. The relevant question for strategic planning is not whether technology changes the mission, but how the PCAOB’s oversight model can keep pace with an audit profession that is already being reshaped by it. 

We encourage the Board to include in its strategic priorities an explicit commitment to understanding and evaluating AI-augmented audit procedures as a recognized form of substantive testing. Auditing frameworks of the past decades were built primarily in a world of sampling. They should be updated to accommodate a world in which coverage can be comprehensive and continuous. Specifically, the 2024 amendments to AS 1105 and AS 2301 created enforceable standards for technology-assisted analysis (TAA) in tests of details and multi-purpose procedures. 

Regulatory guidance or illustrative examples clarifying what constitutes compliant documentation of these matters contemplated by the Board would give firms a basis for confident implementation rather than conservative avoidance. 

The effect of inspection unpredictability on technology adoption is not a theoretical concern, it is a finding the Board documented in its own economic analysis accompanying the 2024 amendments. The adopting release cited multiple surveys finding that auditors were cautious about implementing TAA due to a lack of explicit regulation, and that some firms did not adopt advanced analytic tools partly due to uncertainty about how regulatory authorities would perceive the quality of evidence produced. One study cited in the release stated directly that uncertainty about regulators’ response and acceptance of emerging technologies significantly hinder adoption.

The Board adopted the 2024 amendments in part to reduce that regulatory uncertainty. Yet the amendments left unresolved precisely the questions that generate inspection risk in full-population transaction risk-scoring environments such as how auditors should define investigation criteria when technology assigns risk scores to every item in a population, and what the evidentiary status of below-threshold scores is for purposes of AS 2301. Firms and technology providers operating in this environment cannot design methodologies that are both technically correct and inspection-ready, because no further guidance addresses these questions. The PCAOB’s 2026-2030 strategic period is the right window to complete what the 2024 amendments started. 

2. Changes to the Inspections Program 

Chairman Logothetis noted his intention to shift the inspections focus toward firms’ systems of quality management, with engagement-level reviews serving as validation of those systems rather than the primary emphasis. We strongly support that direction. 

We would go one step further: inspections should actively evaluate how auditors exercised professional skepticism, including whether and how they used available technology to sharpen that skepticism. An inspection model that treats a well-documented, AI-assisted full-population test identically to a manually selected sample of several items in a world of big data is not evaluating audit quality – it is evaluating process familiarity. The two are not the same. 

In the absence of authoritative guidance on TAA procedures, inspectors are applying their own interpretive judgment to engagements that use these tools. The result is inconsistency: firms performing substantively identical procedures reach different inspection outcomes depending on which team reviews the file. This leaves firms in the position of defending their own interpretation of how technology contributed to audit quality, rather than demonstrating compliance with a known standard. 

A predictable consequence we see is over-auditing, wherein firms running AI-assisted procedures in parallel with the manual procedures they have always performed, not because the duplication adds evidentiary value, but because they cannot be certain the technology-assisted work alone will survive inspection. This redundancy consumes audit resources without improving audit quality, and its costs are ultimately borne by the investing public that the PCAOB’s standards are designed to protect. That is not a framework that encourages investment in better tools or rewards the firms that have made that investment thoughtfully. 

3. Inspection Information and Reporting 

We encourage the Board to consider publishing aggregated, anonymized data on the types of TAA procedures inspectors observed and the deficiencies identified in their application. This would serve two purposes: it would give firms and technology providers visibility into where the profession is struggling to implement the 2024 amendments correctly, and it would give the Board’s own research function a basis for evaluating whether the amendments are achieving their intended effect. 

Transparency about inspection findings related to technology-assisted procedures is particularly valuable at this stage of adoption, because firms are making methodology decisions with limited information about what compliant practice looks like in practice. Inspection reporting that illuminates the gap between current practice and the standard’s intent would be more useful to audit quality than generalized deficiency counts. 

4. Standard-Setting Priorities 

We encourage the Board to prioritize two specific guidance or standard-setting projects in the 2026-2030 period, both of which flow directly from gaps left open by the 2024 amendments. 

First, guidance on the definition of investigation criteria for transaction-scored populations. AS 2301.49 requires auditors to investigate items identified for further investigation, but presupposes a clear binary between flagged and unflagged items. That binary does not exist when a platform assigns a continuous and dynamic risk score to every item in a population. The investigation threshold is created entirely by the auditor’s judgment about where to set the score cutoff. 

Guidance clarifying that thresholds must be set before reviewing output, linked to assessed risk and the relevant assertions, and documented as part of the procedure design would resolve the most consequential source of inspection uncertainty in this space. 

Second, guidance on the evidentiary status of below-threshold risk scores for purposes of the AS 2301.50 remaining-population assessment. The Board’s own adopting release for the 2024 amendments acknowledged the boundary of what principles-based standard setting can resolve here. In addressing what happens when a TAA run over an entire population does not identify any items for investigation, the release stated that it was not possible to specify scenarios that would provide sufficient appropriate audit evidence, given that facts and circumstances vary. That is an accurate description of the limits of principles-based rulemaking, and a precise description of the gap that remains. The release confirmed that the Board could not tell auditors what a clean result from a full-population analysis means for the sufficiency conclusion, and then declined to address it. 

That unanswered question is where inspection risk concentrates in practice. An auditor who runs a continuous transaction scoring platform over four million journal entries, investigates the top risky scores, and finds no misstatements has done something analytically meaningful. The current standards give the auditor no sufficient framework for documenting why the result supports the sufficiency conclusion, and no assurance that an inspector applying independent judgment to the same file will reach the same view. Guidance addressing this scenario is not prescriptive; it is the minimum necessary to make principles-based standards workable in a full-population context. 

Beyond these two near-term priorities, we encourage the Board to undertake a broader review of where its standards presuppose human performance and supervision, and to assess where clarification or revision is warranted to accommodate AI systems increasingly executing audit procedures. This review should address at minimum the supervision requirements in AS 1201 and their application when AI systems rather than human assistants are executing procedures, the documentation requirements in AS 1215 and what constitutes an adequate record of AI-generated work product, and quality control considerations under QC 1000 for firms deploying agentic AI tools. This review should be conducted in coordination with the IAASB, IESBA, and other international standard setters. 

5. Alignment with International Auditing Standards 

Global harmonization matters to us directly, as MindBridge serves audit firms in multiple jurisdictions. Auditors operating across borders should not face a patchwork of inconsistent requirements for how technology-assisted procedures are documented, evaluated, or reported. 

We encourage the PCAOB to engage actively with the IAASB and national standard-setters on the question of AI in audit, with the goal of developing shared frameworks for evaluating technology-enabled procedures. This is an area where the PCAOB can exercise genuine leadership rather than simply converging with existing international standards, because neither the PCAOB nor its international counterparts have fully worked through these questions yet. The opportunity to co-develop rather than converge is available now, and we encourage the Board to take it. 

6. Deploying Technology, Including AI, in PCAOB Operations 

We welcome Chairman Logothetis’s emphasis on leveraging technology within the PCAOB itself, and we want to be candid about what that will require. Technology adoption does not fail because of bad tools. It fails because of insufficient investment in the people and mindset shifts needed to use those tools well. The PCAOB has built significant expertise and institutional knowledge over more than two decades. That expertise is an asset. But realizing the benefit of technology-enabled oversight will require treating technology partners as genuine collaborators in that work, not simply as compliance subjects. 

Inspectors have developed deep expertise evaluating whether human-performed procedures were sufficient and well-executed. The emerging challenge is different: assessing whether the judgment applied to technology-generated insights was appropriate. That requires a willingness to engage with how these tools work, what they surface, and what good follow-up looks like in practice. Training alone will not produce that capability. It requires deliberate investment in how inspectors think about evidence, coverage, and professional skepticism in a technology-enabled audit environment. 

Practically, this means: 

• Investing in training that builds inspectors’ ability to evaluate AI-assisted audit procedures on their merits, including understanding what good and poor use of these tools looks like in practice.  

• Treating relationships with technology and audit analytics providers as a source of intelligence about where the profession is heading and needing support.  

• Building internal data and analytics capabilities that allow the PCAOB to work from population-level data rather than samples when conducting its own oversight activities.  

We recognize that the PCAOB is an oversight body, not a technology organization. The market is already determining what assurance looks like in the age of AI, how quickly it must be delivered, and what form it takes. The PCAOB’s oversight role will need to evolve at the same pace, or risk regulating a version of the audit profession that the capital markets have already moved past. Investing in people who can navigate that line is not optional. 

7. Enhancing Transparency with Stakeholders 

The decision to solicit public input before developing the strategic plan, rather than after, is a meaningful signal. We encourage the Board to carry that spirit through the full process, including providing stakeholders with interim updates on how the feedback received is shaping the plan’s development, not only the final product. 

For those of us working at the intersection of technology and audit, transparency about the PCAOB’s evolving views on AI-assisted audit procedures would be particularly valuable. Audit firms make long-term technology and human investment decisions based in part on their expectations about what the inspection and standard-setting environment will look like. Clearer signals from the Board reduce uncertainty and allow these investments to flow toward practices and methodology that genuinely improve audit quality. 

Closing 

MindBridge AI is committed to the advancement of high-quality, technology-enabled audit. Our platform processes billions of transactions annually across audit engagements in multiple countries, and we have accumulated meaningful insight into how AI-assisted procedures can strengthen professional skepticism when they are deployed and embedded in audits thoughtfully. 

We do not offer these comments as advocates for a particular outcome. We offer them as a partner that wants the audit profession to navigate this transition well. A profession that adopts technology without the standards, oversight frameworks, and inspector competencies to evaluate it serves no one’s interests, least of all investors. The PCAOB has a pivotal role to play in ensuring that does not happen. 

We would welcome the opportunity to engage further with PCAOB staff on any of the themes raised in this letter, whether through the formal comment process, the practitioner engagement channels Chairman Logothetis described, or through the PCAOB’s active Data and Technology research project, which invites stakeholders to share analyses and data to help inform its work. MindBridge’s deployment experience across a broad cross-section of audit engagements, is precisely the kind of input that research project is designed to draw on. We are ready to contribute. 

Respectfully submitted, 

Wenzel Reyes, CPA, CA 
Head of AI Governance 
MindBridge Inc. 
wenzel.reyes@mindbridge.ai