A practice owner who can't answer "what did we get for our $48,000 in AI spend last year?" with a specific number in five business categories is going to lose the budget conversation, the partner conversation, or — when the day eventually comes — the buyer-diligence conversation. This lesson installs the five metrics that matter, names the leading indicator that predicts operational AI maturity, shows how to instrument them inside the CRM and time-tracking stack a typical 5-to-50-advisor practice already runs, and turns the result into a one-page monthly ROI dashboard that is board-ready, buyer-ready, and FINRA-Rule-3110-supervisory-ready. The dashboard is the operational follow-on to the 90-day adoption curve and the pre-diligence artifact for L4 Ch8.

Why Five Metrics, Not Fifteen

The advisor AI vendor stack — Jump, Zocks, FinMate AI, Sybill, Zeplyn, Holistiplan, FP Alpha, Wealth.com, fpPathfinder, RightCapital, eMoney, MoneyGuidePro, Orion Eclipse, 55ip, Wealthbox, Redtail Engage, Salesforce Financial Services Cloud with Einstein, Practifi, Smarsh, Global Relay, Catchlight, SmartAsset, and Microsoft Copilot — collectively reports roughly two hundred distinct usage and outcome metrics across the stack. Vendor dashboards trend toward exhaustive: minutes saved, prompts run, tokens consumed, summaries generated, meetings recorded, action items extracted, NIGOs prevented, scoring confidence. The practice owner who tries to consume all of them produces nothing actionable. The CCO who tries to supervise all of them produces no supervisory clarity. The buyer in diligence wants exactly five numbers that survive the practice-economics question.

The five metrics that matter for a wealth practice in 2026, in the order they appear on the dashboard: (1) hours recovered per advisor per week; (2) households per advisor; (3) meeting-to-follow-up SLA; (4) NIGO rate; (5) close rate. Each maps to a specific operational lever, a specific source system, and a specific category of revenue or cost. Each is instrumented inside the CRM, the time-tracking stack, the custodian portals (Schwab Advisor Services, Fidelity Wealthscape, Pershing NetX360+, BNY Mellon), and the AI tool's own usage telemetry. None requires a custom data warehouse. All five are auditable from raw source data — meaning they survive both a buyer's due diligence and an SEC examiner's books-and-records request under Rule 4511 and SEC Rule 204-2.

Metric 1 — Hours Recovered per Advisor per Week

Hours recovered is the headline. Zocks' published claim of 10+ hours/week saved, the Kitces time-and-task research's roughly 36% of advisor hours in meeting prep and servicing, and the Schwab 2026 RIA Benchmarking Study's finding that adoption more than doubled vs. 2023 all point at this single number. It is the number the partner will ask first, the buyer will underwrite, and the advisor will feel.

The clean measurement is the difference between the pre-deployment baseline (captured in the L4 Ch5 L1 weeks 1-2 measurement sprint) and the rolling 30-day moving average post-deployment, segmented by advisor cohort. The sources: time-tracking tools (HoursTrackable, Toggl, the time fields inside Wealthbox / Redtail / Salesforce FSC), CRM meeting-prep-and-follow-up activity logs, and a quarterly advisor self-report (the explicit narrative on where hours actually went). The combination is necessary because each source alone is partial — time-tracking captures explicit logged work, CRM captures workflow-bound work, and the self-report captures the unlogged "background" work (worrying about a household at 11pm, drafting a follow-up at the kitchen table on Sunday) that AI tools relieve.

The honest target for a well-deployed AI stack on a 200-household book in 2026 is 8-12 hours/advisor/week recovered. Practices that report 20+ are usually mis-attributing — either the baseline was unmeasured, or the recovery includes administrative reorganization the AI didn't cause, or the self-report inflates. Practices that report 2-3 typically have a deployment problem (under 60% utilization, missing prompt library, low-trust champion role) the L4 Ch5 L1 curve framework will diagnose. The dollar value of an hour recovered, for the typical $1M-revenue solo or $5M-revenue ensemble RIA, is in the $150-$400 range when allocated to billable advisory work, prospect-meeting capacity, or paraplanner relief.

Metric 2 — Households per Advisor

Households per advisor is the structural metric. The industry standard for a comprehensive-planning RIA practice in 2024 was roughly 80-120 households per lead advisor before AI; the practitioner press in 2026 is now citing the well-deployed AI stack as supporting 150-200 households per lead advisor at the same service quality. The metric is the ratio of currently-served households to producing-advisor headcount, measured at quarter-end. It is auditable from Wealthbox / Redtail / Salesforce FSC household counts cross-referenced against the producing-advisor list.

The strategic significance of this metric is not just productivity — it is the capacity to absorb growth without adding advisor headcount, which is the single biggest determinant of operating leverage in an RIA. A practice that grows AUM 15% and households 12% without adding an advisor in the same period has demonstrated AI-driven operating leverage; buyers underwrite this as the "premium attribute" criterion in L4 Ch8. The Mercer Capital Q3-Q4 2025 and ECHELON data placing top-quartile RIAs at roughly 8x-10x adjusted EBITDA (premium-top ~11.6x) is significantly influenced by demonstrated operating-leverage history, which households-per-advisor is the cleanest proxy for.

The honest target is dependent on practice complexity. A mass-affluent practice (households $250K-$2M) can credibly reach 180-220 households per advisor on a deployed AI stack. A UHNW practice (households $10M+) where each relationship requires deep coordination across estate, tax, equity comp, and entity structures typically tops out around 50-80 households per advisor regardless of AI. The metric's value is in tracking the trajectory inside the practice's own niche, not in chasing a benchmark.

Metric 3 — Meeting-to-Follow-Up SLA

The SLA from meeting end to follow-up communication sent (recap email, action-item summary, trade authorization DocuSign, planning-software update, CRM activity log, Smarsh archive seal) is the operational metric. The pre-AI baseline at most practices is 2-5 business days; many follow-ups never happen at all. The well-deployed Zocks-to-Wealthbox-to-Smarsh pipeline (L3 Ch10 L1) targets 9 minutes from meeting end to fully archived. The realistic 2026 target across all meeting types is under 4 business hours.

The measurement is the time delta between the meeting end (from calendar, Zocks/Jump/FinMate AI/Sybill/Zeplyn timestamps) and the sent timestamp on the follow-up communication (from Outlook/Gmail, Wealthbox/Redtail/Salesforce FSC activity log, or DocuSign envelope creation). The metric is segmented by meeting type — discovery meeting, annual review, quarterly check-in, ad-hoc — because each has different follow-up complexity. The discovery meeting follow-up that converts a prospect into an engagement (L2 Ch2 L3) has the highest revenue leverage and the tightest target (under 4 hours).

The SLA matters for three reasons. First, prospect conversion data shows the prospect-to-engagement conversion rate drops materially when same-day follow-up slips. Second, the supervisory architecture under FINRA Rule 3110 reasonable design and Rule 2210 principal review of AI-drafted communications requires the follow-up to be archived in Smarsh or Global Relay under Rule 4511 and SEC Rule 204-2 — a follow-up that never happens isn't archived, and the Reg BI file under §240.15l-1 is incomplete. Third, the metric is the most-direct adoption signal — when the SLA improves, the advisor is actually using the tool; when it doesn't, the advisor is producing AI drafts and shelving them.

Metric 4 — NIGO Rate — The Leading Indicator

NIGO ("Not In Good Order") is the custodian's classification for any account-opening, ACAT, trade authorization, beneficiary form, or distribution request that arrives at the custodian missing a signature, with a name mismatch, with an unfunded field, with an outdated document, or with any condition that prevents the custodian from processing the request without an additional advisor touchpoint. NIGO is the single best leading indicator of operational AI maturity — meaning a practice's NIGO rate trend predicts its overall AI deployment health 30-60 days before the other metrics catch up.

The mechanism: the AI workflows that produce the highest leverage in a practice (meeting AI feeding CRM feeding planning software feeding custodian forms) all converge on the custodian's intake desk. When the AI workflows are running cleanly — accurate transcripts, correctly-categorized action items, properly-routed planning updates, custodian-form pre-checks running before submission — the NIGO rate drops. When the workflows are under-trained, under-supervised, or producing unedited AI drafts that the advisor pushes through without verification, the NIGO rate climbs as the operational gaps compound. The custodian portals at Schwab Advisor Services, Fidelity Wealthscape, Pershing NetX360+, and BNY Mellon all publish per-firm NIGO rates in their advisor dashboards; the raw counts are auditable.

The honest target is to track the practice's own baseline (captured during the L4 Ch5 L1 measurement sprint) and to require steady-or-better NIGO rate as a precondition for the deployment moving to maintenance mode. A worsening NIGO rate alongside improving hours-recovered is the canonical warning sign — the advisor is using AI to draft faster, but the verification layer (L1 Ch2 L3 Cardinal Rule three-tier check) is not being applied; the operational gaps are compounding inside the AI productivity layer. The L4 Ch6 L1 AI Risk Register treats NIGO regression as a Tier 1 risk indicator. The L4 Ch8 buyer diligence pack lists NIGO trend as a top-line metric.

Metric 5 — Close Rate (Prospect-to-Engagement Conversion)

Close rate is the revenue metric. It is the percentage of qualified prospect meetings (first-touch discovery meetings with a household meeting the practice's qualification criteria) that convert to an engaged client within 90 days. The pre-AI baseline at most growing RIA practices in 2024 was 25-40% depending on niche, lead source, and advisor skill. The well-deployed AI stack — particularly the same-day discovery follow-up workflow (L2 Ch2 L3), the Catchlight or SmartAsset prospect-research integration, the Holistiplan tax-return same-day analysis turnaround, and the AI-generated personalized planning hook — moves close rate up by a measurable percentage in the trade-press data.

The measurement is the count of new engagements that started in the prior 90 days divided by the count of qualified discovery meetings 90 days prior. The source is the CRM (Wealthbox household statuses or Salesforce opportunity stages) cross-referenced against the discovery meeting calendar. The metric must be segmented by lead source — Catchlight referrals, SmartAsset leads, COI referrals, organic, paid search — because each has a different baseline conversion rate, and aggregating them obscures the lift the AI workflow delivers.

The Marketing Rule 206(4)-1 consideration is critical here. Any close-rate claim used in marketing — to peers, on a podcast, in a marketing brochure — triggers the substantiation file under Rule 206(4)-1(d) and the "clear and prominent" disclosure standard. The L4 Ch7 L2 lesson on testimonial / third-party rating / ADV / performance disclosure operationalizes this. The dashboard tracks close rate for internal management purposes; using it externally requires the substantiation and disclosure discipline.

Instrumenting the Five Metrics Inside the Existing Stack

None of the five metrics requires a custom data warehouse or a new BI tool. The instrumentation is inside the systems the practice already runs.

CRM (Wealthbox, Redtail Engage, Salesforce FSC, Practifi)

Households per advisor, meeting-to-follow-up SLA, close rate, and segmented lead-source data all come from CRM activity exports. Wealthbox's activity API exports meeting records, activity timestamps, household ownership, and pipeline stages. Redtail Engage's reporting layer produces the same. Salesforce FSC + Einstein has the most mature reporting layer with native productivity dashboards out of the box; the AI activity log adds another layer. Practifi's reporting suite is built for ensemble RIAs and supports per-advisor segmentation natively. The practice's monthly dashboard pulls a standard set of exports — meeting count by advisor, household count by advisor, time-from-meeting-to-activity-log timestamp, pipeline-stage transitions for prospect-to-engagement.

Time Tracking (HoursTrackable, Toggl, CRM Time Fields)

Hours recovered is measured against the weeks 1-2 baseline. The ongoing measurement uses the same time-tracking tool as the baseline. The discipline is twofold — the advisor logs explicit work time, and the CRM activity log captures workflow-bound work. The quarterly advisor self-report covers the unlogged background work. The combination produces a defensible hours-recovered number — defensible meaning it survives a buyer's diligence question about how it was measured.

Custodian Portal (Schwab, Fidelity, Pershing, BNY Mellon)

NIGO rate is exported from the custodian's advisor portal NIGO dashboard. Schwab Advisor Services publishes monthly NIGO rates with categorization (signature missing, beneficiary mismatch, ACAT-related, etc.); Fidelity Wealthscape produces an equivalent report. The practice tracks the trend month-over-month and segments by NIGO category to surface which workflow gap is responsible for any uptick.

AI Tool Telemetry (Jump, Zocks, Holistiplan, FP Alpha, etc.)

License utilization (the day-90 checklist item from L4 Ch5 L1), prompt library version, meeting-AI capture rate, and extraction-tool processing volume come from the AI vendor's own admin dashboard. Each named tool exports the data; the dashboard pulls a standard set of usage signals per tool. The cross-check matters — if Zocks reports 95% meeting capture but Wealthbox reports only 70% activity-record creation, there is a sync failure to investigate.

Archive Coverage (Smarsh, Global Relay)

The archive vendor reports coverage rates — what percentage of meetings, AI summaries, advisor edits, supervisor signoffs, and AI-generated client communications are sealed in the tamper-proof timestamp archive under WORM compliance per SEC Rule 17a-4(f) for BDs and the practical equivalent for IAs under Rule 204-2. This isn't a productivity metric per se but it is the audit-grade companion that lets the practice say to a buyer or examiner: "Here is what we shipped, here is what we archived, and here is the gap." The L4 Ch5 L1 day-90 checklist requires ≥99% archive coverage on AI-touched artifacts.

The Monthly Board-Ready ROI Dashboard

The dashboard is one page. The recipients are the practice owner, the COO, the CCO, and (for ensemble RIAs) the partner group. The cadence is monthly, with a quarterly deep-dive that adds segmentation. The 2026 board-ready template has six sections.

Section 1 — Five Headlines. The five metrics in a single row each, with the current month value, the prior month value, the prior year value (or the pre-deployment baseline if year-over-year not yet available), and a single-color trend indicator. No charts, no decoration, just the numbers and the deltas.

Section 2 — AI Spend. Per-tool seat cost, total monthly spend by tool, per-advisor allocated cost, and the rolling 12-month spend trajectory. The line items: Jump or Zocks (meeting AI), Holistiplan plus FP Alpha plus Wealth.com (planning AI), Wealthbox or Salesforce FSC + Einstein (CRM AI), Orion Eclipse (portfolio AI), Smarsh or Global Relay (archive), Catchlight or SmartAsset (prospecting AI), Microsoft Copilot Enterprise (horizontal). Total monthly spend for a 10-advisor ensemble typically runs $4,000-$12,000/month.

Section 3 — Productivity Lift. Hours recovered times advisor blended rate equals dollar value recovered. AUM/advisor trend. Households/advisor trend. Total productivity lift expressed in dollars on a monthly run-rate.

Section 4 — Revenue Lift. AUM growth attributable to capacity expansion (the structural argument from households-per-advisor). COI referrals trended over time. Time-to-onboard trended over time (a leading indicator of new-revenue ramp). Close rate by lead source. New client revenue closed in the period.

Section 5 — Operational Health. NIGO rate, meeting-to-follow-up SLA, archive coverage rate, principal review queue exception rate (under FINRA Rule 2210), Reg BI documentation completeness (sampled), Smarsh archive coverage. The CCO's pillar on the dashboard.

Section 6 — ROI Calculation. The headline calculation in one line: ($ productivity lift + $ revenue lift) ÷ $ AI spend, expressed as a multiple. For a well-deployed practice in 2026 the multiple typically runs 4x-8x — meaning every dollar of AI spend produces $4-$8 in measurable productivity-and-revenue value. The multiple becomes the case for next year's budget and the headline figure the buyer underwrites in diligence.

How the CCO Uses the Dashboard (and Why FINRA Cares)

The CCO reads the dashboard differently from the practice owner. Sections 1 and 6 confirm the practice's investment thesis. Sections 4 and 5 carry the supervisory weight. The NIGO rate trend, the meeting-to-follow-up SLA, the archive coverage, and the principal review exception rate are the operational signals the FINRA 2026 Annual Regulatory Oversight Report (informed by Reg Notice 24-09) framed as the supervisory layer for AI deployments under Rule 3110 reasonable design.

The CCO's monthly memo to the AI Governance Committee (L4 Ch6 L1) attaches the dashboard, names any threshold breaches (NIGO worsening, SLA slipping, archive coverage gap, principal review exception spike), and proposes mitigations. The memo and the dashboard become books-and-records under FINRA Rule 4511 and SEC Rule 204-2 — the practice's audit-grade documentation that the AI deployment is supervised and measured. In an SEC exam or FINRA cycle, the 12 trailing monthly dashboards + governance committee memos + WSPs constitute the supervisory architecture exhibit. There is no substitute for it; firms that can't produce this set are exactly the firms the 2025-2026 AWC pattern is identifying.

The Buyer-Diligence Version of the Dashboard

The dashboard's second audience is the buyer in any M&A scenario. The L4 Ch8 lessons develop the full diligence framework; the operational input is the trailing 12-36 months of monthly dashboards. The buyer is underwriting the operating leverage the practice has demonstrated and the AI-maturity-driven multiple premium (0.5-1.5x on the top-quartile 8x-10x range, per Mercer Capital and ECHELON Q3-Q4 2025; premium-top ~11.6x adjusted EBITDA).

Two specific dashboard reads matter in diligence. First, the trajectory shape — a clean ramp from baseline to a steady-state target across 12 months reads as institutionalized adoption; a noisy trajectory with reversals reads as fragile and discounts the premium attribute. Second, the supervision exhibit — the operational health pillar (Section 5) is what the buyer's CCO underwrites; clean coverage, low exception rate, steady NIGO, complete Reg BI files combine to support the multiple. The practice that runs the dashboard for 24+ months walks into diligence with the artifact set already produced; the practice that starts the dashboard six weeks before the buyer arrives is exactly the discount-multiple pattern.

Month 13 — The Second-Year Conversation

The dashboard's twelfth month produces the annual review. The recipients add the partners (in an ensemble), the AI Governance Committee (L4 Ch6 L1), and — for any practice planning a transition or near-term sale — outside counsel. The conversation has three threads. First, the spend trajectory — what worked, what didn't, what to retire (L5 Ch2 L2 develops the discipline of killing under-performing tools). Second, the next-year intervention — what tool category to add, what workflow to upgrade, what training cohort to expand. Third, the ADV Part 2A reading — has the tool stack changed materially since the prior annual amendment, and if so, the L4 Ch7 L2 off-cycle amendment lesson governs the timing.

The L4 Ch1 lesson on the three-year roadmap is where the multi-year arc is anchored; the annual dashboard review is the operational mechanism. Year 1 is productivity. Year 2 is integration. Year 3 is differentiation. The dashboard's role across the arc is to produce the numbers that survive partner scrutiny, buyer diligence, regulator exam, and the practice owner's own honest assessment of whether the AI investment is paying.

Key Takeaways

• Five metrics, not fifteen. Hours recovered/advisor/week, households/advisor, meeting-to-follow-up SLA, NIGO rate, close rate. Each maps to a source system, a revenue or cost lever, and an auditable raw-data trail under FINRA Rule 4511 and SEC Rule 204-2.
• NIGO rate is the leading indicator of operational AI maturity. NIGO trend predicts deployment health 30-60 days before the other metrics catch up. Worsening NIGO alongside improving hours-recovered = AI is being used to draft faster without the L1 Cardinal Rule verification layer.
• Hours recovered honest target is 8-12/advisor/week on a well-deployed stack. 20+ usually means mis-attribution. 2-3 means a deployment problem. Dollar value of a recovered hour: $150-$400 in 2026 typical practice economics.
• Households per advisor is the operating-leverage metric. 80-120 pre-AI standard; 150-200 well-deployed-AI standard for mass-affluent / mass-HNW. UHNW practices top out around 50-80 regardless. Buyers underwrite the trajectory as the premium-attribute criterion (L4 Ch8).
• Meeting-to-follow-up SLA target is under 4 business hours. 9 minutes per the Zocks-to-Wealthbox-to-Smarsh pipeline (L3 Ch10 L1) for routine follow-ups. Discovery follow-up under 4 hours converts prospects at materially higher rates. SLA is the most-direct adoption signal.
• The dashboard is one page, six sections. Five headlines, AI spend, productivity lift, revenue lift, operational health, ROI multiple. Cadence monthly, deep-dive quarterly. Audience: practice owner + COO + CCO + partners. Multiple typically 4x-8x in a well-deployed practice.
• CCO uses the dashboard for FINRA Rule 3110 supervisory architecture. Operational health pillar (NIGO, SLA, archive coverage, principal review exceptions under Rule 2210 + Marketing Rule 206(4)-1, Reg BI completeness under §240.15l-1) is the supervision exhibit. Monthly memos to the AI Governance Committee (L4 Ch6 L1) attach the dashboard and become books-and-records under Rule 4511 + Rule 204-2.
• The dashboard is the pre-diligence artifact for L4 Ch8. 24+ months of trailing dashboards demonstrate institutionalized adoption and support the AI-maturity premium (0.5-1.5x on the top-quartile 8x-10x range; premium-top ~11.6x per Mercer Capital and ECHELON Q3-Q4 2025). Six-week pre-sale dashboard production reads as fragile and supports the discount multiple.