In our Ask Kaia question-trends report covering April 1 to June 30, 2026, lending compliance turned out to be the single subject bankers asked about more than any other, with 1,309 tagged questions representing 27.5% of all substantive questions over the quarter.
That concentration shows where compliance staff most often reach the edge of what a policy manual or a static reference page can answer, and where they want a second set of eyes on a fact pattern.
This article walks through why lending leads the data, how the theme grew from April to June, which subtopics generate the most questions, and why fact-specific lending scenarios are so hard for static guidance to handle.
The Growth Trend: From 19.8% to 28%
On a primary-theme basis, lending questions rose from 19.8% of the total in April to 28.0% in June, according to the Ask Kaia report. Over three months, roughly one in five primary-theme questions became closer to one in four.
Within that rise, appraisal-bias and valuation-fairness language grew noticeably in June. That shift lines up with sustained regulatory and industry attention to fair valuation in residential lending, and it suggests staff were increasingly bringing appraisal-review questions to an AI assistant.
A trend line points to where compliance teams are feeling pressure in near-real time, and it gives risk leaders an early read on which lending subtopics may need updated procedures, refreshed training, or a clearer escalation path.
What Bankers Actually Ask About Lending Compliance
The lending questions blended origination, servicing, advertising, adverse action, appraisal review, credit purpose, and state-law wrinkles, and most of them turned on specific facts. A few patterns recurred across the quarter:
Consumer versus business credit
Staff repeatedly asked whether a given loan fact pattern counted as consumer or business credit, because that single determination changes which disclosures and protections apply.
Adverse action handling
Many mortgage compliance questions centered on adverse action notice requirements: how to word denial reasons, how to treat incomplete applications, and how to handle files with multiple applicants.
Mortgage disclosures and servicing
Bankers asked whether mortgage disclosures, servicing notices, rate-change or ARM notices, PMI-cancellation timing, and escrow actions were required in each situation.
Appraisal and valuation review
A recurring request was help reviewing appraisal and valuation language for bias, an appraisal bias review task that calls for judgment about wording, comparables, and neutral description rather than a yes-or-no rule check.
Mortgage advertising review
Staff also asked how to review advertising for NMLS references, Equal Housing language, APR trigger terms, and product-specific disclaimers before a campaign went out.
Lending Subtopic Hotspots
The table below breaks down the six lending compliance hotspots identified in the Ask Kaia report, with the tagged-question count, its share of all substantive questions, and why each one lends itself to AI-assisted triage.
| Lending subtopic | Tagged questions | Share of substantive questions | Why it favors AI triage |
|---|---|---|---|
| RESPA, servicing, escrow, flood, PMI, and ARM timing | 254 | 5.3% | Answers hinge on loan lifecycle stage and timing rules that vary by scenario |
| Mortgage advertising and lending disclosures | 194 | 4.1% | Requires checking specific language against multiple disclosure triggers |
| TILA, Reg Z, HPML, HOEPA, APR, and credit cards | 191 | 4.0% | Thresholds and calculations depend on product terms and fact inputs |
| HMDA reporting and LAR decisions | 108 | 2.3% | Data-field determinations turn on transaction details and coverage tests |
| Adverse action and denial notices | 95 | 2.0% | Correct reason and timing depend on the specific decision and applicant facts |
| Appraisal bias and valuation fairness | 93 | 2.0% | Needs judgment about wording and comparables, not a single rule lookup |
Why Static Guidance Struggles with Lending Questions
Static guidance struggles with lending questions because the answer so often depends on the specifics. Product design, borrower facts, collateral type, disclosure timing, and the institution's own policy can each change the outcome.
The regulatory landscape adds to the complexity, as a single lending question can implicate several regimes at once: TILA and Regulation Z, ECOA and Regulation B, HMDA and Regulation C, and RESPA and Regulation X.
Understanding HMDA reporting requirements, for instance, means working through coverage tests and data-field rules that only resolve once you know the transaction details. Keeping those regimes current is its own discipline, which is why lending sits close to the work of managing regulatory change.
How Controlled AI Triages Lending Scenarios
A controlled AI workflow addresses lending questions by guiding the user through the facts a determination requires. In the operating model described in the Ask Kaia report, lending scenario triage walks users through the inputs needed for Reg Z, Reg B, RESPA, HMDA, flood, PMI, adverse action, or servicing questions, assembling the fact pattern.
The same operating model routes complex cases, or any case with adverse customer impact, to a compliance officer or counsel rather than resolving them automatically. Human review and an audit trail run throughout.
Ask Kaia is one implementation of this pattern, applying guided triage and mandatory escalation for higher-risk lending scenarios. The value is speed and consistency on the intake, with the accountability for the answer staying where it belongs.
Frequently Asked Questions
Why is lending compliance the top AI use case for banks?
Lending compliance is the top AI use case because it drew the most questions of any specific subject in the Ask Kaia report, 1,309 tagged questions or 27.5% of substantive questions over the quarter. Lending was the largest specific content category on a primary-theme basis. The volume reflects how often staff face fact-specific lending scenarios that a static reference page cannot resolve.
What are adverse action notice requirements?
Adverse action notice requirements govern how a lender communicates a credit denial or other adverse decision to an applicant, including the specific reason for the decision and its timing. Under ECOA and Regulation B, the notice must state accurate, specific reasons. Handling incomplete applications and files with multiple applicants adds complexity.
What are HMDA reporting requirements?
HMDA reporting requirements, set under the Home Mortgage Disclosure Act and Regulation C, require covered institutions to collect and report data about mortgage applications and originations in a loan application register. Determining coverage and the correct data-field entries depends on transaction details, which is why HMDA and LAR decisions generated 108 tagged questions in the Ask Kaia report.
How do banks use AI for lending compliance?
Banks use AI for lending compliance mainly to triage scenarios: the assistant guides staff through the facts a rule requires, such as the inputs for Reg Z, Reg B, RESPA, HMDA, flood, PMI, or adverse action, then organizes the analysis. Controlled workflows route complex or adverse-customer-impact cases to a compliance officer or counsel, with human review and an audit trail throughout.
To see how this connects to the broader shift in compliance work, the next step is to look at applied compliance judgment and how AI scenario triage supports it without replacing the human decision.
Learn how Ask Kaia can assist your organization’s compliance team in gaining clarity on regulatory changes.
Request Demo- Policy Drafting
- Compliance Automation
- Audit Trails
- Regulatory Intelligence