Your AI use is on theexam surface now.
Would it survive one? Mark the six domains an SEC or state examiner probes when an adviser uses AI — marketing, books & records, supervision, vendor oversight, Form ADV disclosure, communications capture — and get a defensibility verdict per firm, from your own evidence.
Not legal advice. This is a readiness self-assessment that grades a firm's evidence from your own marks — not a compliance audit, a mock exam, a certification, or an opinion of counsel. It renders no legal ruling, confirms no obligation, and scores no person. Confirm which rules apply to your firm, and every date, with your CCO or a qualified securities attorney.
Examiners stopped treating AI as a novelty. The controls it needs are the ones you already owe.
AI is in your marketing
AI-drafted posts, emails, and testimonials fall under the Marketing Rule like any other advertisement — and a lighter review path 'because AI wrote it' is a finding waiting to happen.
AI outputs are records — or should be
AI notetaker summaries and AI-drafted client comms function as records. Whether they're retained is the single most live books-and-records question an adviser faces right now.
A high score can hide a fatal hole
A firm can look strong on paper and still be indefensible if one load-bearing control is simply absent. An overall percentage won't tell you which; a gate will.
Mark six domains per firm. The fatal gate does the rest.
Mark each exam-surface domain 0/1/2 on evidence. The verdict falls out.
Fix first: Books & records for AI outputs — the weakest exam-surface domain across the portfolio.
The Northgate branch scores 73 on points yet reads NOT DEFENSIBLE — its books-and-records domain (dotted, a fatal domain) is at zero, and a strong file elsewhere can't cure a missing load-bearing control. Set that domain to 1 and it clears. The verdict comes from your own evidence marks; it grades a firm's exam evidence, renders no legal ruling, and scores no person. Not legal advice.
The same verdict from the command line.
The workbook and a zero-dependency Python engine produce identical results. Point the engine at a firms file and it prints the exam-surface read — verbatim below, from the shipped sample.
==============================================================================
ADVISORY-FIRM AI EXAM DEFENSIBILITY AUDIT
Exam-surface read - as of 2026-07-05
==============================================================================
PORTFOLIO VERDICT ..... DEFICIENCY LIKELY
Firms graded .......... 5
2 exam-ready - 1 open findings - 2 not defensible
Fix first ............. Books & records for AI outputs
------------------------------------------------------------------------------
FIRM / BRANCH MAR REC SUP VEN DIS COM SCORE VERDICT
------------------------------------------------------------------------------
Main RIA (HQ) 2 2 2 2 2 1 94 EXAM-READY
Northgate Branch 2 0 2 2 1 2 73 NOT DEFENSIBLE
Lakeside Advisors 1 1 1 1 1 1 50 OPEN FINDINGS
Summit Wealth (new) 1 0 0 1 0 1 24 NOT DEFENSIBLE
Harbor Point LLC 2 2 2 1 2 2 92 EXAM-READY
------------------------------------------------------------------------------
! Northgate Branch scores 73 on points yet reads NOT DEFENSIBLE:
a fatal exam domain is at zero - Books & records for AI outputs.
A strong file elsewhere cannot cure a missing load-bearing control.
KEY: MAR=AI in marketing & advertising | REC=Books & records for AI outputs | SUP=Supervision & written policies | VEN=Third-party / vendor AI oversight | DIS=Client disclosure & Form ADV | COM=AI communications capture
The verdict comes from your own evidence marks - no benchmark. It
grades a firm's exam evidence, renders no legal ruling, and scores
no person. Not legal advice - confirm every rule and date with counsel.
==============================================================================Six weighted domains, three of them fatal.
AI in marketing
Marketing Rule review, testimonial disclosure, substantiated claims.
Books & records for AI
AI summaries and AI-drafted comms retained where they function as records.
Supervision & policies
AI named in the written compliance program, with an owner.
Vendor AI oversight
Due diligence, contracts, and data-flow records for AI vendors.
Disclosure & Form ADV
Material AI use disclosed; no overstated capability claims.
AI comms capture
AI-assisted client communications captured and reviewable.
Each domain is marked 0/1/2 on evidence; the weights sum to 100 for a 0–100 score, banded EXAM-READY (75+) / OPEN FINDINGS (50–74) / NOT DEFENSIBLE (under 50). The gate is dispositive: a zero on any fatal domain forces NOT DEFENSIBLE regardless of the score — because each is a finding that draws a deficiency letter on its own.
A way to find your fatal exam gaps before an examiner does. Not an opinion of counsel.
What it is
- A readiness self-assessment for an adviser using AI
- Six exam-surface domains mapped to the Advisers Act frame
- A per-firm defensibility verdict with a fatal-domain gate
- A portfolio rollup across branches, with the domain to fix first
What it isn't
- Not a compliance audit, mock exam, or certification
- Not an opinion of counsel and not a legal ruling
- Not tied to a statutory date — it grades process, not a deadline
- Not a score of any person — it grades a firm's evidence file
Not legal advice. This is a readiness self-assessment that grades a firm's evidence from your own marks — not a compliance audit, a mock exam, a certification, or an opinion of counsel. It renders no legal ruling, confirms no obligation, and scores no person. Confirm which rules apply to your firm, and every date, with your CCO or a qualified securities attorney.
Anyone who owns an adviser's exam readiness.
RIA principals & CCOs
Find the fatal gap in your AI controls before an examiner asks — and know which branch is the weakest link.
Compliance consultants
Run a consistent, evidence-based read across every client firm, and hand each one a prioritized remediation list.
Multi-branch & roll-up firms
Grade every office on the same six domains; the portfolio verdict follows the weakest branch, so you know where to look.
Firms adopting AI notetakers
The books-and-records domain is where notetaker adopters quietly fail. See whether your summaries are actually retained.
Triage the broad gap; deepen the domains this audit flags.
Pass a 90-Day AI Audit? Readiness Triage
The generic, cross-industry front door — seven governance domains, which broad gap to close first. This audit is its adviser-specific counterpart.
ViewAI Output Audit-Trail & Record-Keeping Kit
If books-and-records is your fatal gap, this grades whether each retained AI entry is actually complete — who, what tool, when, reviewed by whom.
ViewAI Vendor & Sub-Processor Data-Flow Register
Builds the vendor-oversight artifact this audit's fourth domain asks for — what each AI vendor touches, and whether it's governed.
ViewBefore you buy.
Find the fatal gap first.
Before the examiner does.
One purchase, lifetime access, 12 months of updates. $99, once.
Not legal advice. This is a readiness self-assessment that grades a firm's evidence from your own marks — not a compliance audit, a mock exam, a certification, or an opinion of counsel. It renders no legal ruling, confirms no obligation, and scores no person. Confirm which rules apply to your firm, and every date, with your CCO or a qualified securities attorney.
Sold by RedHub AI LLC · Secured by Stripe · redhub.ai