Will Zoom disclose a quantitative Custom AI Companion paid metric (attach rate or ARR) by April 2027?

Twelve-month window with 4 quarterly earnings + likely Zoomtopia Investor Day (Oct-Dec) + 10-K (Feb 2027) + 8-K opportunity creates multiple independent disclosure venues, but the dominant signal is the 4-quarter pattern of qualitative-only framing. Management has had 6+ disclosure opportunities across FY26 (Q1-Q4 calls, Zoomtopia Dec 2025 Investor Day, 10-K Feb 2026) and chose 'strong adoption' / free-tier MAU framing every time. Per Insider Investigator analog, this is informationally meaningful asymmetry — companies typically disclose paid AI metrics when material (MSFT Copilot, ServiceNow Now Assist Pro Plus). Zero discretionary insider buys at ~$84 corroborates that the behavioral signal is not consistent with imminent material AI monetization disclosure. Black Swan Beacon assigns 30-40% probability to attach <10% AND ARR <$100M by FY28 (failed monetization scenario), which would make management strategically prefer continued silence. Resolution criteria explicitly EXCLUDES management's preferred fallback paths (aggregated AI revenue, qualitative descriptors, free-tier MAUs) — raising the bar. However, the YES case is meaningful: (1) 12-24 months post-launch (CY2025 launch → FY27) matches MSFT Copilot's natural disclosure cadence; (2) competitive disclosure pressure escalates as peers monetize and disclose; (3) Investor Day at Zoomtopia is a natural high-stakes venue; (4) analyst Q&A pressure compounds quarterly and deflection becomes harder; (5) bull case incentive to disclose IF attach is strong is high (validates pricing power, supports multiple expansion). Weighted reasoning: P(attach strong enough that disclosure validates bull thesis) ~0.40 × P(disclose | strong) ~0.80 = 0.32; P(attach weak) ~0.60 × P(disclose | weak) ~0.15 = 0.09; sum ~0.41. Add small additional weight (~0.01-0.03) for forced-disclosure paths (Anthropic event, regulatory pressure, competitor catalyst). Settling at 0.42 — slightly below 50/50 because pattern continuation is the strongest single signal and the resolution bar excludes management's preferred linguistic fallbacks, but the 12-month window with multiple venues and natural disclosure timing prevents going meaningfully lower.

Decompose into a mixture over the latent state of paid Custom AI Companion attach. Bayesian setup: (i) prior P(attach is sub-threshold) ~ 35% (BSB failed-monetization scenario at 30-40%) vs P(attach is on-track-or-strong) ~ 65%. (ii) Conditional disclosure probabilities over the 12-month window: if attach is weak, management has direct incentive to NOT quantify (P_disclose ~ 8-12%); if attach is strong, voluntary disclosure to validate bull narrative is strategically valuable and competitive pressure (MSFT Copilot paid seats, NOW Pro Plus attach) compounds (P_disclose ~ 60-70%). Mixture: 0.35 x 0.10 + 0.65 x 0.65 = 0.035 + 0.423 = 0.458, rounded toward 0.44 after applying a downward adjustment for the strongest single signal in the dossier — 4 consecutive quarters of qualitative-only AI framing across Q1-Q4 FY26 plus Investor Day plus 10-K, which is a revealed-preference signal that management's default is non-disclosure regardless of trajectory. Also de-weighting the mixture upper-bound: comparable disclosure cadence (MSFT 12-15 months post-launch) is real but ZM management style is notably more qualitative than peers (signature 'AI-first system of action' framing without numbers). Counter-pressures toward YES: (a) 4-lens convergence (Moat, Gravy, Myth, Atomic) means analyst Q&A pressure escalates each cycle; (b) 6+ disclosure venues in window (4 earnings calls, Zoomtopia investor day, FY27 10-K); (c) 4 acceptable disclosure pathways in resolution criteria (attach %, seat count, ARR, revenue %); (d) Zoomtopia (Oct-Dec 2026) is typically when product-AI metrics roll out; (e) Anthropic IPO/strategic event (~40-60% probability per BSB) could force AI commercial-relationship transparency tied to Zoom Ventures Fund $1.6B mark-to-market position. Counter-pressures toward NO: (a) zero discretionary insider buys at ~$84 across 4-month Form 4 corpus — behavioral confirmation insiders don't expect imminent material AI disclosure; (b) common deflection pattern where management discloses aggregated 'AI revenue' or 'AI Companion 3.0 MAUs' (free tier) without Custom AI Companion attribution, which explicitly does NOT resolve YES; (c) FY27 10-K won't be filed until ~Feb 2027, meaning the most binding venue is back-loaded near window close; (d) 4-quarter avoidance is the strongest base-rate signal — Bayesian update against a 'flip in FY27' hypothesis. Net: market price 0.50 anchor reasonable; my read is slightly bearish at 0.44 due to revealed-preference dominance over the comparable-peer cadence argument. Confidence MEDIUM because the binary outcome hinges on a single strategic management decision and the latent attach-rate truth is meaningfully unobservable from outside.

Independent superforecaster reasoning: Decompose into base rate, asymmetric incentive structure, and window mechanics. (1) BASE RATE FROM REVEALED PATTERN. Management has had 6 explicit disclosure venues across FY26 (Q1-Q4 calls, Investor Day at Zoomtopia Dec 2025, 10-K Feb 2026) and chose qualitative framing in every one. With four prior consecutive non-disclosures and an Investor Day specifically designed to showcase the AI strategy, the empirical hazard rate per quarter is ~0%. A naive Laplace-rule extension over 4 more quarterly opportunities gives a per-quarter implied probability that bounds the FY27 result well below 50%. This is the strongest single anchor. (2) ASYMMETRIC PAYOFF FOR MANAGEMENT. Disclosure is a one-way signal: once a number is on the tape, it becomes the QoQ benchmark and any deceleration is punished disproportionately. Eric Yuan and Michelle Chang have repeatedly chosen MAU framing on the free tier — a deliberate substitution of a fast-growing free metric for a slow-growing paid metric. This pattern is consistent with paid attach being below internal target. Behavioral economics of corporate disclosure: managers disclose when the metric is at or above expectations and helps the multiple; they suppress when below. The MAU substitution is itself evidence the paid number is unflattering at threshold. (3) INSIDER BEHAVIOR CORROBORATES. Zero discretionary insider buys at ~$84 over 4 months. If a material AI disclosure were near, insiders typically front-run modestly. Absence is a soft confirmatory signal that nothing imminent is teed up. (4) FORCES TOWARD YES. (a) Microsoft/ServiceNow precedent creates analyst-Q&A pressure each quarter; (b) Anthropic IPO/strategic event (40-60% in FY27) could force disclosure on Zoom Ventures Fund mark and AI relationship economics — but this is conditional and partial; (c) Zoomtopia 2026 (~Oct-Dec 2026) is the highest-leverage venue, and a year of additional adoption could push the metric to a level worth disclosing; (d) the resolution criteria are reasonably generous — ANY of attach rate, paid seat count, Custom AI Companion ARR, or revenue contribution % counts. Management could split the difference with a 'we now have X paid seats on Custom AI Companion' soundbite without committing to ongoing cadence. (5) RESOLUTION LENIENCY MATTERS. Four acceptable disclosure forms across 4 earnings + 1 Investor Day + arbitrary 8-K = ~6 high-probability venues with multiple satisfying formats per venue. The disjunction inflates probability above what a strict 'attach rate %' standard would. (6) WHAT KEEPS PROBABILITY BELOW 50%. The non-disclosure pattern through 4 quarters PLUS Investor Day is strong. Management has the motive (asymmetric punishment) and the cover (free-tier MAU framing as substitute). For YES to occur, attach must reach a number management is willing to commit to publicly AND they must overcome the disclosure-suppression equilibrium. Given the bull thesis depends on this number, if it were strong, Zoomtopia 2025 would have been the natural venue — and management chose not to disclose there. That revealed-preference moment is heavily informative. (7) NUMERICAL SYNTHESIS. Per-quarter hazard if attach is on-track: ~12-15%. Per-quarter hazard if attach is sub-threshold: ~3-5%. Weighting these by my prior on attach being on-track (~40%): expected per-quarter hazard ~7-9%. Compounding over ~5 venues (4 quarters + Zoomtopia + 8-K window) with venue-correlation: P(YES) approximately 1 - (1-0.08)^5 ~= 34%, then add ~6-8 points for Anthropic-event forcing function and analyst pressure escalation = ~40-44%. FINAL: 0.42. MEDIUM confidence — outcome is genuinely path-dependent on attach trajectory and Zoomtopia framing, but base rate of management non-disclosure is the dominant evidence and pulls the estimate below 50%.

The dominant signal is management's demonstrated pattern of non-disclosure across six major opportunities (Q1-Q4 FY26 earnings calls, Zoomtopia Investor Day Dec 2025, 10-K filed Feb 2026) — each time choosing qualitative framing ('strong adoption', 'leading AI assistant', free-tier MAU growth) over any quantitative paid metric. This is not oversight; companies that are hitting material paid thresholds have strong investor-relations incentive to quantify. The six-opportunity silence is strong Bayesian evidence that the metric is sub-threshold. Against this, the 12-month window (4 earnings cycles + Investor Day + FY27 10-K) provides meaningful optionality, and the Microsoft Copilot analog suggests natural disclosure cadence at 12-24 months post-launch — Custom AI Companion is in that window. Competitive pressure from analyst Q&A will escalate, but management has standard deflection available ('we don't break out individual product lines'). Resolution criteria are strict, explicitly excluding free-tier MAUs, qualitative descriptors, and aggregated AI revenue without Custom AI Companion attribution — management's current disclosure posture would not resolve YES even if slightly expanded. Zero discretionary insider buys at ~$84 across the Form 4 corpus is a mildly negative confidence proxy. Net assessment: ~30% probability of a qualifying quantitative disclosure in the 12-month window.

The dominant signal is the established 4-quarter non-disclosure pattern: management had Q1-Q4 FY26 earnings, Zoomtopia Investor Day (Dec 2025), and the FY26 10-K (Feb 2026) — six disclosure opportunities — and chose qualitative framing ('strong adoption', '+3-4x MAUs on free tier') over any quantitative paid metric every single time. This pattern is not accidental; it reflects either sub-threshold attach rates or deliberate management optionality preservation. The base rate for a company that has consistently withheld a metric continuing to withhold it is high. Against this, the YES case has three credible inputs: (1) the 12-month resolution window provides 4 additional earnings plus Zoomtopia FY27 plus the FY27 10-K — materially more opportunities than any prior quarter; (2) the Microsoft Copilot analogy is credible (paid seat disclosure ~12-15 months post-launch, which aligns with Custom AI Companion's timeline for FY27); and (3) the resolution criteria cast a wide net — any of attach rate, paid seat count, ARR, or revenue contribution % across any disclosure venue resolves YES, reducing the bar. However, zero discretionary insider buying at $84 is a mild negative — if management had high confidence in a strong paid AI disclosure coming, insiders would be deploying personal capital. The most likely NO path is continued aggregated qualitative AI framing or 'AI Companion MAU' disclosures that do not cross the strict resolution threshold. The most likely YES path is escalating analyst Q&A pressure forcing a directional quantification at Q2 or Q3 FY27 earnings. Net: non-disclosure pattern dominates, but 12-month window and broad resolution criteria prevent this from going below 35%.

The 4-quarter consecutive non-disclosure pattern (Q1-Q4 FY26, Investor Day, 10-K) is the dominant base rate signal. Management has had 6+ opportunities to disclose a quantitative paid metric for Custom AI Companion and chosen qualitative framing ('strong adoption', MAU growth on free tier only) at each one. The simplest prior is continuation. However, three countervailing forces move probability toward YES: (1) the resolution threshold is low — any quantitative paid metric qualifies, not a minimum level, meaning even a modest disclosure (e.g., '5% enterprise attach') resolves YES; (2) the 12-month window contains 4 more quarterly earnings cycles plus Investor Day and FY27 10-K, and competitive pressure from Microsoft Copilot and ServiceNow Now Assist Pro Plus disclosures will escalate analyst Q&A pressure, particularly by Q3-Q4 FY27; (3) the Microsoft Copilot analog (paid seat count disclosure ~12-15 months post-launch) suggests the FY27 window is the natural inflection period for Custom AI Companion, launched earlier in 2025. The absence of insider buys at ~$84 is a mildly negative signal that management lacks conviction in imminent positive catalysts. Net: non-disclosure base rate dominates but the accessible resolution threshold and long window create meaningful YES probability. I shade 3 points below 50% to reflect the strength of the non-disclosure pattern as a deliberate management choice, not an oversight.

Management demonstrated preference for qualitative AI framing (free-tier MAU growth) across 4 quarters of FY26 earnings despite multiple disclosure opportunities (Q1-Q4 earnings, Investor Day, 10-K). This pattern suggests actual Custom AI Companion paid attach rate remains sub-threshold (<10% enterprise base or <$100M ARR). Comparable SaaS companies disclose paid AI metrics when achievement is material; reluctance to quantify typically signals weak monetization. However, 12-month FY27 window provides 4 quarterly cycles + Investor Day + FY27 10-K filing + potential Anthropic catalysts. If Custom AI Companion attach matures to >10% or >$100M ARR (threshold defined in market), disclosure becomes strategically valuable to support multiple expansion and validate bull thesis. Insider behavior (zero discretionary buys despite bull narrative) is mildly negative for imminent disclosure strength. Resolution criteria are strict: requires quantitative attach rate %, paid seat count, or ARR disclosure—free-tier MAUs and vague qualitative descriptors explicitly excluded, reinforcing structural headwind.

Management's 4-quarter pattern of qualitative-only AI framing (Custom AI Companion attach undisclosed through Q1-Q4 FY26 earnings cycles, 10-K, Investor Day) despite multiple disclosure opportunities is a strong signal of either sub-threshold attach rate (incentive to avoid quantification) or strategic caution. Base-rate continuation of non-disclosure into FY27 is the simpler forecast. However, 4-lens convergence on disclosure as upgrade trigger, Custom AI Companion's 12-24mo post-launch window (natural disclosure point), and escalating competitive pressure (Microsoft Copilot, ServiceNow Now Assist) create material YES scenario. Zero insider buying at ~$84 is mildly negative for bull thesis confidence. The 12-month window (Q1-Q4 FY27, Investor Day, 10-K) provides multiple opportunities, but the committee's own 'informationally meaningful asymmetry' framing (Insider Investigator analog) suggests management may continue qualitative-only path to avoid setting quantified expectations on a potentially sub-threshold metric. Estimate reflects genuine 48/52 split rather than 50/50: slight lean toward continued non-disclosure as the base-rate continuation, but high probability either outcome resolves fairly.

Disclosure hinges on two competing asymmetries: (1) Bull case (YES): Management has incentive to disclose IF attach has materialized; 12+ months post-launch is natural window for investor pressure + competitive disclosure pressure to escalate. Anthropic IPO/strategic event could force transparency on AI commercial relationship and Zoom Ventures Fund mark. (2) Bear case (NO): Four-quarter pattern of qualitative-only framing is strongest base rate predictor. If actual attach is below threshold (<10% or <$100M ARR), management has incentive to avoid quantification. Aggregated 'AI revenue' without Custom AI Companion attribution does NOT resolve YES, creating escape hatch. Free-tier MAU updates remain path of least resistance. Materiality of actual attach rate (unknown; could be <5% explaining non-disclosure, or 8-12% approaching threshold) is binding variable. Inertia slightly favors continued avoidance.