L2 ended with the structured-output discipline that lets every artifact land in Wealthbox, RightCapital, Orion Eclipse, and Smarsh without paraplanner re-keying. L3 begins one layer up. The L2 advisor used AI to compress a single workflow at a time — meeting prep, follow-up, IPS draft, rollover memo. The L3 advisor designs the entire practice as a portfolio of workflows and chooses, deliberately and economically, which ones AI eats first. The L3 chapter you are starting is the thinking layer that precedes every multi-tool, multi-step workflow in chapters 2 through 10: Roth conversion seasons run across fifty households, RMD calendars run across the entire book over 73, estate-gap audits run across the trust population, equity-comp exercise modeling runs across the executive niche. None of those compound until the practice itself has been mapped. This lesson installs the workflow audit — a complete inventory of every recurring motion the practice runs, scored on time spent and AI leverage, with a 90-day intervention list that becomes the operating plan for the rest of L3.

Why the Audit Comes Before the Tools

The dominant 2024-2026 failure mode in advisor AI adoption is not vendor selection — it is the absence of a map. The advisor sees a Jump demo at a custodian conference, deploys it in two weeks, captures forty-five Zocks transcripts, and discovers in month three that the Smarsh archive integration is misconfigured, the Wealthbox follow-up template is the firm's pre-AI default, the Reg BI rationale paragraph still requires twelve minutes of manual writing per memo, and the paraplanner is still re-keying the action items into the household record. The AI tool is doing exactly what it was sold to do. The practice never mapped the workflow the tool was supposed to live inside. The Schwab 2026 RIA Benchmarking Study's adoption-doubling headline obscures a quieter finding: most firms with deployed AI tools are realizing a fraction of the productivity collapse the tools are technically capable of delivering, because the deployments are unmapped point solutions rather than designed workflow interventions.

The audit forces the inverse discipline. Before any new AI tool is selected, before any prompt is locked into the library (L2 Ch8 L1), before any workflow is automated end-to-end (L3 Ch2 through Ch7), the practice produces a complete inventory of every recurring motion it runs — what the motion is, how often it runs, how many person-hours it consumes per year, which household segment it serves, what regulatory artifacts it produces, and what current tooling supports it. The inventory is the practice's operational asset list. Every L3 workflow lesson assumes this inventory exists. Every L4 strategic decision (Ch1 strategy, Ch2 vendor selection, Ch5 ROI dashboard, Ch7 ADV disclosure, Ch8 M&A diligence) depends on the inventory being current and accurate.

The audit is also the artifact a 2026 SEC examiner will ask to see during a Compliance Rule 206(4)-7 review when the question is "how does your firm decide which client-facing communications are AI-touched and which are not?" The audit is the artifact an M&A buyer's diligence team will ask to see when the question is "what is the operational maturity of the AI integration here, and where are the gaps you have not yet closed?" The audit is the artifact the AI Governance Committee (L4 Ch6 L1) will reference quarterly when triaging new use cases. A practice that has never produced its workflow audit can buy tools and deploy them; it cannot run an AI-integrated practice in the L3 sense.

The Fifteen Canonical Recurring Motions of a 200-Household Practice

Every 200-household RIA, wirehouse advisor pod, or family-office team runs some subset of fifteen canonical recurring motions. The labels and the cadence vary; the underlying work does not. The audit's first pass simply lists each motion the practice runs, marks the ones it does not, and notes the household segment served.

Annual Reviews and Quarterly Meetings

The recurring meeting motion that defines the practice. For a 200-household book with a 60/40 split between annual-review and quarterly-meeting cadence, the practice runs roughly 320 review meetings per year. Each meeting produces a prep pack (L2 Ch3 L1), an in-meeting note capture (L2 Ch3 L2), a follow-up package (L2 Ch3 L3), CRM activity updates, planning-software annotations, and Smarsh archive entries. Pre-AI baseline: 3.5-4.5 advisor hours per review end-to-end. Post-AI achievable: 1.5-2.0 hours. Annual time spent across the book at baseline: roughly 1,200 advisor-hours. Annual recoverable hours: 500-700 if the workflow is integrated, not just AI-touched.

Prospect-to-Client Onboarding

The 30-day window after a "yes" — Form CRS and ADV Part 2A delivery, IPS draft (L2 Ch5 L1), engagement letter, ACATs initiation, beneficiary form collection, custodian NIGO recovery (L2 Ch7 L1), Reg BI documentation for any rollover (L2 Ch7 L2), CRM record build-out. For a practice adding 25-40 new households per year, onboarding consumes roughly 350-500 person-hours annually pre-AI. The Hendersons-style household onboarding (married couple, multi-account, multi-beneficiary) easily takes 12-18 paraplanner hours per household; AI-integrated workflows compress that to 4-6 hours while improving the Reg BI documentation completeness rate.

Client Offboarding

Quiet, infrequent, and overrun by NIGO friction. Termination paperwork, ACATs out, final fee billing, CRM archival, regulator-grade record retention under FINRA Rule 4511 and SEC Rule 204-2, books-and-records pull for the departing client, optional offboarding survey for the L4 Ch5 ROI dashboard. Roughly 4-6 hours per departing household at baseline; AI-integrated workflows compress to 1.5-2.5 hours and produce cleaner Rule 4511 retention packages.

RMD Season

October through December every year, for every client over age 73 (with the SECURE 2.0 step-up to 75 in 2033) and every inherited IRA holder under the 10-year rule. Calculation under the Uniform Lifetime Table or Single Life Table, custodian cross-reference, deadline confirmation, destination decision (cash to checking, in-kind transfer, QCD direction under IRC §408(d)(8) for the 70.5+ charitable household), client communication, execution, 1099-R verification. The L3 Ch3 L1 RMD calendar workflow is built directly on the audit's RMD-population count.

Year-End Tax-Loss Harvesting

November and December across every taxable account. Loss-identification at the lot level (Schwab / Fidelity / Pershing / BNY Mellon cost-basis feeds), wash-sale rule compliance (substantially identical security check across all accounts including spouse's IRA — the often-missed leg), replacement-security selection, IPS-aligned tax-loss-harvest execution, $3,000 ordinary-income offset planning, carryforward documentation. The pre-AI workflow is highly manual; AI-integrated workflows surface candidate lots with wash-sale-safe replacements and rationale memos at scale.

Beneficiary Reviews

Custodial brokerage TOD/POD, IRA primary and contingent, Roth IRA primary and contingent, 401(k), 529, HSA, life insurance, annuity. The L3 Ch5 L2 estate-gap memo is built on this motion. Most practices run beneficiary reviews ad-hoc rather than systematically; the audit forces a cadence (typically annual on the review meeting + triggered on life events) and a documentation standard.

IPS Updates

L2 Ch5 L2 covered life-event triggers (marriage, divorce, inheritance, business sale, retirement, death of spouse, birth of grandchild, special-needs diagnosis). The audit converts the lesson's logic into a population-level workflow: how many households have IPS editions older than 24 months, how many have an open life-event trigger that has not produced an IPS revision, how many have IPS-to-trade reconciliation breaches (L2 Ch5 L3) that have not been resolved. Most practices discover during the audit that 40-60% of their IPS population is stale.

Quarterly Commentary and Client Education

Market commentary plus household-personalized concept memos (Roth conversion, NUA, QCD, 72(t), mega-backdoor) generated from L2 Ch6 L1's combined template. The audit captures cadence, distribution list, Marketing Rule 206(4)-1(d) compliance posture, and the principal-review queue volume under FINRA Rule 2210.

Roth Conversion Season

Q4 of every year. L3 Ch2's three-lesson sequence — 50-household screen, per-client sizing memo, Reg BI-compliant recommendation + execution + archiving — is the AI-integrated form of this motion. Pre-AI, most practices run conversion conversations for the dozen most-obvious candidates and miss the next thirty households where the leverage is comparable but the household wasn't surfaced by ad-hoc memory. The audit's Roth-eligible population count drives the L3 Ch2 L1 workflow design.

Rollover Reg BI Memos

Triggered by every 401(k), 403(b), or 457 rollover. L2 Ch7 L2 covered the per-memo workflow. The audit captures volume (typical 200-household practice generates 15-40 rollover memos per year), the average per-memo time pre-AI (45-90 minutes), the four-alternatives-documented completion rate (the FINRA 2026 AWC vulnerability), and the principal-review SLA.

Estate Document Review

L2 Ch4 L3 covered the per-document Wealth.com / FP Alpha intake; L3 Ch5 L1 builds it into a population-level estate intake workflow. The audit captures the count of households with documents on file, the count with documents older than five years, the count missing standard documents (will, trust, POA, healthcare directive), and the count with documented gaps that have not been remediated.

Performance Reporting

Monthly or quarterly performance package generation, distribution, and Marketing Rule 206(4)-1 compliance review. The audit captures the principal-review queue volume, the disclosure-language version control, and the hypothetical-performance-rule exposure (none, ideally; the rule's strict requirements make hypothetical performance a high-risk category to handle without explicit policy).

Quarterly Billing

AUM-fee calculation, household billing memo generation, payment processing, fee-rebate or fee-credit memo generation when applicable, ADV Part 2A fee-disclosure reconciliation. The audit captures the household-segment fee schedule complexity (flat-fee, tiered AUM, capped, hybrid retainer + AUM) and the audit-trail completeness for the L4 Ch7 fee-disclosure strategy.

CCO / OSJ Periodic Reviews

FINRA Rule 3110 principal review of communications (Rule 2210), AI-drafted client content (L4 Ch3 L2), supervisory log entries, Smarsh archive sampling, ADV Part 2A annual update workflow (L5 Ch7 L6), Form U4 monitoring. The audit captures review volume, sampling cadence, and the exception-handling SLA.

Prospect Discovery and Conversion

L2 Ch2's three lessons covered pre-meeting research (Catchlight + LinkedIn + public filings), live note capture, and same-day follow-up. The audit captures prospect volume, conversion rate, average time-to-close, and the marketing communication compliance posture for any prospect-quote reuse.

The Scoring Rubric — Time Spent vs. AI Leverage

Every motion the audit lists gets two scores. The first is time spent — total practice person-hours per year, summed across senior advisor, associate, paraplanner, and CSA. The second is AI leverage — the practical percentage of the motion that can be compressed by AI under current tooling, given the constraints of Reg BI, Marketing Rule, FINRA Rules 2210/3110/4511, Reg S-P, and the firm's WSPs.

Time-Spent Scoring

The senior advisor and the practice manager produce an estimate per motion using either the practice's existing time-tracking data (if it exists), the L4 Ch5 ROI dashboard baseline (if a prior measurement has been done), or a calibrated estimate from sampling two weeks of practice activity. The estimate is per-role and per-cycle: how many senior advisor hours per review, per onboarding, per rollover memo, per quarterly billing run; same for associate, paraplanner, CSA. Multiply by annual cycle count to get annual person-hours.

Most practices, doing this for the first time, discover (a) the practice runs 50-70% more recurring motions than the senior advisor mentally tracked, (b) the senior advisor's personal time allocation diverges materially from the Kitces canonical split (~20% client meetings / 36% prep/planning/servicing / 15% prospecting / 20%+ admin), usually toward more admin and less meeting prep, (c) the paraplanner's actual time allocation is dominated by 3-4 specific motions (onboarding, RMD season, quarterly billing, IPS-to-trade reconciliation) that are also the highest-leverage AI candidates.

AI-Leverage Scoring

Each motion gets a 1-to-5 score on AI compressibility under current tooling. Score 1: motion is fundamentally human judgment (the actual meeting itself, the difficult client conversation, the trustee selection memo's discretionary call) — AI augments prep and follow-up but does not compress the core motion. Score 2: motion has some AI-compressible components (the pre-meeting prep, the post-meeting follow-up) but the core motion remains human time. Score 3: motion is half human, half AI-compressible (most planning workflows, where the calculation is AI but the recommendation is human). Score 4: motion is dominantly AI-compressible (extraction-heavy workflows — Holistiplan-led tax-memo workflow, FP Alpha-led estate intake, RMD calendar generation, beneficiary review across the population). Score 5: motion is almost entirely AI-compressible (pattern detection, NIGO pre-check at custodian submission, off-channel detection in Smarsh, lead-scoring at Catchlight or SmartAsset).

The score is current-tooling-constrained — agentic AI's 2027-2028 trajectory (L5 Ch6 L1) will shift several motions up the scale, but the audit scores what the practice can deploy this quarter, not what the industry might deploy in two years.

The Combined Leverage Score

Multiply time-spent (annual person-hours) by AI-leverage (1-5 scaled to 0.2-1.0 as a fraction). The product is the annual person-hours the practice can recover by AI-integrating that motion. Rank-order the motions. The top three are the practice's 90-day intervention candidates. The next five are the 6-month roadmap. The bottom seven are the 12-18-month roadmap or "augment but don't restructure" candidates.

A representative result for a typical 200-household RIA: (1) annual reviews and quarterly meetings — 1,200 hours x 0.6 = 720 recoverable; (2) onboarding — 425 hours x 0.7 = 298 recoverable; (3) RMD season — 280 hours x 0.8 = 224 recoverable; (4) Roth conversion season — 180 hours x 0.7 = 126 recoverable; (5) estate document review — 220 hours x 0.7 = 154 recoverable; (6) beneficiary reviews — 95 hours x 0.6 = 57 recoverable; (7) IPS updates — 165 hours x 0.5 = 83 recoverable; (8) rollover Reg BI memos — 140 hours x 0.6 = 84 recoverable; (9) year-end TLH — 75 hours x 0.6 = 45 recoverable; (10) quarterly commentary and client education — 85 hours x 0.5 = 43 recoverable. Total recoverable: roughly 1,800 person-hours per year, or about 0.9 advisor-equivalent FTE. That number is the practice's L4 Ch5 ROI baseline before any AI investment lands.

Producing the 90-Day Intervention List

The audit's deliverable is not a spreadsheet of scores. The deliverable is the 90-day intervention list: the three motions the practice will AI-integrate first, the named tools (from the L1 Ch3 AdvisorTech map and the L2 chapters' tool selections) that will support each, the prompt library entries (L2 Ch8 L1) and few-shot exemplars required, the schema definitions (L2 Ch8 L2) for downstream integration, the verification protocols (L1 Ch2 L3 + L3 Ch1 L3) embedded in each workflow, the Reg BI / Marketing Rule / Rule 4511 retention design (L1 Ch4 + L3 Ch1 L2), and the success metrics (L4 Ch5 L2 ROI dashboard) that will track whether the integration is working.

The 90-day discipline is intentional. Most practices that fail at AI integration fail because the rollout was either too small (single tool, single workflow, no compounding) or too large (firmwide platform replacement attempted without earned-credibility on the first three workflows). 90 days is the unit of time in which three workflows can be designed, deployed, debugged, and measured — long enough to produce learning, short enough to maintain organizational urgency, and matched to the typical quarterly cycle of CCO supervisory review and practice-owner financial review.

The first 90-day list typically includes (a) the L3 Ch2 Roth conversion season workflow if Q4 is approaching, (b) the L3 Ch3 RMD calendar if December is approaching, (c) the L2 Ch3 annual review prep pack workflow as the universal high-volume motion, or (d) the L2 Ch7 onboarding workflow if the practice is in active acquisition mode. The actual selection follows the leverage scores plus the practice's annual calendar plus the named-tool readiness (does the firm already have Holistiplan? FP Alpha? RightCapital? Wealthbox vs Redtail vs Salesforce FSC?).

The Audit as Living Document and Governance Anchor

The audit is not a one-time exercise. The practice updates it quarterly — adding newly-discovered motions, retiring obsolete ones, re-scoring as tooling evolves (FP Alpha's Estate Insights 2.0 launch shifted several estate motions from leverage-3 to leverage-4), capturing actual recovered hours from the L4 Ch5 ROI dashboard, and feeding the AI Governance Committee's (L4 Ch6 L1) quarterly review.

The audit's regulatory function is documentation. The 2026 SEC examiner asking the Compliance Rule 206(4)-7 question — how does the firm decide which client-facing communications are AI-touched — gets shown the audit, the scoring rubric, the 90-day intervention list, the workflow design diagrams (L3 Ch1 L2 next), and the verification checkpoints (L3 Ch1 L3 after that). The exam answer becomes "we ran a workflow audit on X date, scored each motion on time-spent and AI-leverage, prioritized the top three for AI integration, deployed under the WSPs in section [N], and instrument the ROI dashboard quarterly. The current AI-touched motions are [list], each with the verification protocol documented at [reference]." That answer is the difference between a practice that adopted AI and a practice that runs an AI-integrated operation.

The audit's M&A function is valuation defense. The L4 Ch8 framework treats documented AI workflow integration as a multiple driver — the top-quartile / premium-tier RIA EBITDA multiples (8x-10x per Mercer Capital and ECHELON Q3-Q4 2025 data, with the highest-rated transactions reaching ~11.6x at the premium top, vs. the broader-market median in the 6x-8x range) widen further for firms with documented AI maturity. The audit is the artifact that demonstrates the maturity to the buyer's diligence team. A firm with a current, scored, intervention-tracked audit is materially easier to underwrite than a firm with equivalent AUM and scattered tool use.

The Handoff to the Rest of L3

L3 Ch1 L2 (next) takes the audit's prioritized motions and draws the human-AI handoff diagrams for each — where AI drafts, where AI extracts, where AI proposes, where the advisor signs under Reg BI. L3 Ch1 L3 then embeds the three-tier verification pattern (source-system / regulatory / client-fit) as discrete checkpoints inside every multi-step workflow, with Rule 4511 retention baked into the checkpoint chain. Chapters 2 through 7 then build out the specific workflows the audit identified as 90-day candidates: Roth conversion (Ch2), RMD / inherited IRA / backdoor (Ch3), Social Security and IRMAA (Ch4), estate and advanced vehicles (Ch5), business owner and equity comp (Ch6), non-standard family and education (Ch7). Chapters 8, 9, and 10 then layer the ESG/charitable strategy, advanced prompt techniques, and compliance archiving plumbing across the workflow portfolio. Every workflow in L3 starts from a motion the audit identified.

The Robert Chen executive-comp scenario, the Wilsons family-of-four onboarding, the Hendersons Q4 review with its $96,000 Roth conversion window and the unfunded revocable trust — these reappear throughout L3 as the worked examples. The audit is what surfaced them as priorities in the first place.

Key Takeaways

• The audit comes before the tools. Most 2024-2026 advisor AI deployments under-realize their productivity potential because the deployment was an unmapped point solution. The practice never produced the workflow inventory the tool was supposed to live inside.
• Fifteen canonical recurring motions cover most 200-household practices: annual reviews, quarterly meetings, onboarding, offboarding, RMD season, year-end TLH, beneficiary reviews, IPS updates, quarterly commentary, Roth conversion season, rollover Reg BI memos, estate document review, performance reporting, quarterly billing, CCO reviews, and prospect discovery.
• Score each motion on time-spent (annual person-hours summed across roles) and AI-leverage (1-5 current-tooling-constrained). The product is annual recoverable person-hours. Rank-order to produce the intervention list.
• The 90-day intervention list is the deliverable — three motions the practice AI-integrates first, with named tools, prompt-library entries, schema definitions, verification protocols, Reg BI / Marketing Rule / Rule 4511 retention design, and success metrics. 90 days is the unit in which three workflows can be designed, deployed, debugged, and measured.
• Most 200-household RIAs surface 1,500-2,000 recoverable person-hours per year (roughly 0.9 advisor-equivalent FTE) when the audit is run for the first time — annual reviews, onboarding, RMD season, Roth conversion season, and estate document review typically dominate the top of the list.
• The audit is a living document and a governance anchor. Updated quarterly, fed to the AI Governance Committee (L4 Ch6 L1), referenced in SEC Compliance Rule 206(4)-7 exam responses, and used as the M&A buyer-diligence artifact that defends the top-quartile / premium-tier EBITDA multiple (8x-10x per Mercer Capital and ECHELON Q3-Q4 2025, ~11.6x at premium top).
• The audit hands off to L3 Ch1 L2 and L3. The next lesson draws human-AI handoff diagrams for each prioritized motion; the lesson after embeds the three-tier verification pattern as discrete checkpoints. Chapters 2-7 then build the specific workflows the audit identified.