Important takeaways.
- The SEC's Division of Examinations published its fiscal year 2027 examination priorities on 23 October 2026. They are a staff statement with no legal force and create no new obligation. Private credit advisers are still squarely in them.
- The 2027 priorities call AI super intelligence (SI) where the 2026 priorities said artificial intelligence. The Division says it will test whether firms using it still comply with existing, technology-neutral obligations.
- The 2026 sentence on reviewing firms' representations about their AI capabilities is gone. The test that representations are fair and accurate, and that operations match disclosures to investors, is still listed for automated tools.
- At a private credit adviser, the back-office AI use examiners will ask about is mostly loan monitoring. Spreading, covenant tests and monitoring notes feed the loan marks, so AI there is both SI use and a valuation input.
- Rule 206(4)-7 requires an annual review of the compliance programme, and the priorities say exams will typically analyse it. An annual review that does not mention AI, at a firm whose analysts use it weekly, is the first gap to close.
- The evidence an examiner needs is a record that reconstructs each AI-produced figure: the page it came from, the run that produced it and the person who approved it.
The SEC's Division of Examinations published its fiscal year 2027 examination priorities on 23 October 2026. Private credit and AI both appear, and the AI now has a new name. For an adviser to private credit funds, the priorities add no new rule. They do tell you what an examiner will ask to see.
This guide answers for the compliance lead or COO at an adviser registered with the SEC that manages private credit funds. That includes Australian and European credit managers registered there. Our reading is that the exam will test whether AI sits inside the compliance programme you already run. The evidence that it does is a record of what the AI did and who checked it.
The priorities create no obligation, but they name your asset class and your tools.
The document opens with a disclaimer. It is a staff statement with no legal force, and it creates no new obligations. What it does is set where examination resources go for the year.
Three passages land on a private credit adviser. The Division will prioritise advisers that recommend illiquid products, and it names loans provided by private funds or business development companies. For advisers to private funds, it will check whether controls address both fund-level and investment-level risks. Across all advisers, it will review whether they have assessed the reasonableness of their own valuation processes and methodologies.
The leadership message then describes registrants using generative tools for research, drafting, coding, portfolio management and compliance monitoring. The Division says it is intensifying efforts to understand how these tools influence outcomes. It also asks whether firms continue to comply with existing, technology-neutral obligations. That phrase is the whole approach. No AI rule is coming through this document. The existing rules apply to whatever the AI touches.
What changed from the 2026 priorities.
The fiscal year 2026 priorities said "artificial intelligence (AI)". The 2027 document says "super intelligence (SI)" throughout. The change is in the wording, not the scope. The examples given are the same generative tools for research, drafting and coding.
Two other changes matter more, and one priority stays.
- The AI-washing sentence is gone. The 2026 document said the Division would review, for accuracy, registrant representations about their AI capabilities. The 2027 document drops that sentence. It keeps the test, for automated tools, that representations are fair and accurate and that operations and controls are consistent with disclosures made to investors.
- The cyber priority now names SI-driven attacks. Examiners will look at how firms prepare for SI-generated phishing and for synthetic or deepfake voice and video used for social engineering.
- Monitoring SI stays a priority. The Division will prioritise examinations of policies and procedures for monitoring SI. It names fraud prevention and detection, back-office operations, AML and Reg BI compliance, and trading.
Do not read the dropped sentence as permission. The fair and accurate test is still listed, and an investor-facing claim about AI is still a representation.
In a private lender, the back office is the monitoring cycle.
"Back-office operations" reads as settlement and reconciliation at a listed manager. At a private credit adviser it is mostly loan monitoring. Borrower reporting is collected, financials are spread, covenants are tested and a monitoring note is written. Those outputs feed the loan marks.
That puts an AI step in the monitoring cycle in scope twice. Once as SI use the adviser is expected to monitor. Again as an input to a valuation whose methodology the adviser must be able to defend. A spread figure an assistant pulled from a borrower's PDF is a valuation input. So is a covenant headroom figure it calculated.
The compliance rule ties it together. Rule 206(4)-7 requires written policies and procedures, and a review of their adequacy and effectiveness at least annually. The priorities say examinations will typically analyse that annual review. If your last annual review does not mention AI and your analysts use it every week, that gap is the first thing to close.
What to have ready before an examiner asks.
- A list of where AI touches the monitoring cycle and the marks. Give each use a named owner. Include AI features inside the loan system, the document platform and the fund administrator's portal.
- A written review step. Who checks an AI-produced figure before it reaches a valuation, a committee or an investor report, and how that check is evidenced.
- A record that reconstructs the output. Which page each figure came from, which run produced it and who approved it. Without that, nobody can show that operations match what the firm told investors.
- DDQ and marketing answers about AI, checked against what actually runs. Our view is that the fair and accurate test applies to every AI claim made to an investor, not only to automated advice.
- Vendor terms. The cyber section assesses risks from increased use of third-party service providers and shared technologies. Know what your AI vendors retain; zero data retention covers less than most firms assume.
- A payments control that a deepfake cannot pass. A lender's exposure to SI-generated social engineering is the drawdown request and the changed bank details. A call-back to a number already on file is the control.
Our position is that the monitoring cycle is where an examiner should expect to find AI first in a private credit adviser. It is where AI output becomes a number an investor relies on. A firm that can produce the record behind each figure answers the valuation question and the SI monitoring question with the same document. The auditability of that record matters more than which model produced it. Singapore reached the same place through the MAS AI guidelines, and the Australian equivalent is FSC Standard No. 30.
Primary sources: the Division of Examinations fiscal year 2027 priorities, the SEC press release of 23 October 2026, the fiscal year 2026 priorities for comparison, and Rule 206(4)-7. This guide is general information, not legal advice. Exam scope for a given adviser depends on its business and on staff judgement.
Questions this guide answers.
What do the SEC's 2027 exam priorities say about AI?
The Division of Examinations calls it super intelligence (SI) and will prioritise examinations of firms' policies and procedures for monitoring it. It names fraud prevention and detection, back-office operations, AML and Reg BI compliance, and trading. It also names SI-generated phishing and deepfake voice and video as cyber threats firms should prepare for. The priorities were published on 23 October 2026 and create no new rule.
Do the SEC 2027 exam priorities apply to private credit fund managers?
Yes, to advisers registered with the SEC, including non-US credit managers registered there. The priorities name loans provided by private funds or business development companies among the illiquid products the Division will prioritise. They also cover valuation processes and whether private fund advisers' controls address fund-level and investment-level risks.
What changed on AI between the SEC's 2026 and 2027 exam priorities?
Three things. The 2027 priorities say super intelligence (SI) where 2026 said artificial intelligence. They drop the 2026 sentence on reviewing the accuracy of firms' representations about their AI capabilities, while keeping the fair and accurate test for automated tools. And they add SI-driven cyberattacks, such as SI-generated phishing and deepfakes, to the cybersecurity priority.
How should a private credit adviser prepare for an SEC exam on AI use?
List where AI touches the monitoring cycle and the loan marks, with a named owner for each use. Write down who reviews an AI-produced figure before it reaches a valuation or an investor, and keep a record of each check. Match DDQ and marketing claims about AI to what actually runs. Cover AI in the annual compliance review under Rule 206(4)-7. This is general information, not legal advice.
This guide is general information, not advice. Factual claims that rely on public sources link to those sources in the text. Practical guidance also draws on patterns Levercon observes across Australian credit funds. No client is named and no figure is attributed to one. Written by Levercon, reviewed before publication and revised in place as the facts change.
Trusted AI-powered portfolio monitoring built for private lenders.
Spend less time on routine monitoring and more time on credit analysis, credit judgment, and origination.