AI Hallucination ResearchFindings by audienceSectorsInternational / MultilateralSoftware & SaaSCompliance › Promoting the Harmonisation of Application Programming Interfaces to Enhance Cross-Border Payments: Recommendations and Toolkit
Software & SaaS × Compliance — International / Multilateral · Last updated 11 Jun 2026 · methodology v2.3 · Hallucination Register
Share / Print X LinkedIn Email

AI Hallucination on Promoting the Harmonisation of Application Programming Interfaces to Enhance Cross-Border Payments: Recommendations and Toolkit for Compliance teams at Software & SaaS firms in international jurisdictions

Software & SaaS Compliance teams: documentation and reporting gaps possible from AI reading of CPMI Cross-Border API Harmonisation 2024

Compliance officers at software and SaaS firms building cross-border payments platforms aligned to the CPMI API harmonisation programme are increasingly using AI to draft customer-facing CPMI alignment statements, generate regulatory-horizon-scan summaries for client compliance teams, prepare board-paper compliance annexes on the SARB pre-validation workstream, update product-level CPMI mapping documents against the 10 recommendations, and validate ISO 20022 address-format commitments against regulator-issued source text. The RLB Specialist Panel tested how that AI usage performs against the regulator's own primary text on CPMI's October 2024 d224 report and the related CPMI Brief and speech series.

The audit surfaced four substantive failure modes that the AI subjects delivered with regulator-fluent confidence.

Stakeholder Taxonomy Fabrication and Fabricated Date-and-Format Commitment on CPMI API Harmonisation for Cross-Border Payments. Two frontier AI models tested by the RLB Specialist Panel returned confident, citable answers across the panel's CPMI substrate-bound question set on the October 2024 d224 report and the related CPMI Brief and speech series. The panel binds each AI finding to verbatim regulator-issued source text held as primary substrate.

Across the 2 findings in this Compliance teams at Software & SaaS firms briefing, the AI subjects built a recommendation-by-recommendation stakeholder breakdown from category names rather than the regulator's actual recommendation text; introduced a specific November 2026 cutover commitment for structured ISO 20022 addresses that does not appear in the regulator's text.

A customer-facing CPMI alignment statement that records a November 2026 structured-ISO-20022 cutover as a CPMI mandate gives the customer's procurement and compliance team a regulator commitment that does not exist. A product-level CPMI mapping document built on AI per-recommendation stakeholder taxonomy carries fabricated assignments into the platform's roadmap and into customer-facing collateral.

The findings are published with immutable RLB Citation IDs: RLB-H-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q008-Opus47, RLB-H-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q009-Sonnet46. The full audit is published at the CPMI API Harmonisation for Cross-Border Payments hub on RegLegBrief.com.

Compliance leads at payments-API SaaS firms touch CPMI material at three points: the regulatory-applicability statement the sales team uses with bank prospects, the SOC 2 control narrative tied to payments-schema obligations, and the customer-facing roadmap commitment on ISO 20022 readiness. Two AI failures on this regulation hit those exact deliverables. Opus 4.7 returned a per-recommendation stakeholder taxonomy reconstructed from category labels rather than the d224 recommendation text, and Sonnet 4.6 committed to a November 2026 ISO 20022 structured-address cutover the d230 source does not state.

Either error, lifted into a vendor-facing deliverable, hands a bank-customer counsel the exact misallocation they will challenge at procurement review.

What the AI got wrong, and why it matters here

Both failures land on artefacts SaaS compliance teams ship out of the company, not artefacts kept internal. The review window is the customer's procurement counsel, not the SaaS firm's own second line.

Finding 1: Reconstructed stakeholder taxonomy

Opus 4.7 returned a clean stakeholder taxonomy across d224's 10 recommendations, built from category labels rather than the recommendation text. A SaaS regulatory-applicability statement or SOC 2 narrative written off that taxonomy assigns the vendor obligations the regulator did not assign, which procurement counsel at the bank customer will flag.

Citation: RLB-H-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q008-Opus47.

Finding 2: Fabricated November 2026 ISO 20022 cutover

Sonnet 4.6 committed to a hard November 2026 structured-address-only cutover for ISO 20022 cross-border payment messages. The d230 source describes only standardisation and regulatory developments since 2023 and a separate technical annex; the November 2026 cutover is not there. A customer roadmap commitment quoting the AI line commits the product to a regulator deadline that does not exist.

Citation: RLB-H-INT-BIS-CPMI-API-HARMONISATION-CROSS-BORDER-2024-Q009-Sonnet46.

When this hits the SaaS compliance calendar

SaaS compliance pulls CPMI material on three artefacts: the regulatory-applicability statement for the sales pack, the SOC 2 control narrative, and the customer-facing roadmap commitment to ISO 20022 readiness.

Standing artefact Where the AI risk surfaces Failure mode
Regulatory-applicability statement Stakeholder-obligation routing Finding 1
SOC 2 control narrative Stakeholder-obligation routing and cutover commitments Findings 1 and 2
Customer roadmap commitment on ISO 20022 readiness Cutover dates Finding 2

Aggregate impact on the team

Both failures show up in customer-facing deliverables. The downstream risk is procurement-counsel challenge and reputational hit when the misallocation is caught externally rather than internally.

Risk ImpactCountAffected findings
0

What this team should do

Tag the d224 stakeholder taxonomy and the d230 ISO 20022 cutover date as known-failure outputs. Any AI draft headed for a customer-facing deliverable must be returned through a primary-source check (d224 recommendation text and d230 source) before it ships externally.

Detection patterns to add to AI-review

  • Stakeholder-obligation mapping on d224 must be verified against the recommendation text.
  • ISO 20022 cutover-date assertions against d230 must be verified against the d230 text and technical annex.

How RLB can help

RLB tracks AI failures on d224 and d230 and refreshes the catalogue against live AI subjects on rotation. SaaS compliance teams can wire the catalogue into the customer-deliverable review step so these two failure shapes are caught before the language ships to a bank or PSP customer.

Every finding on this page compares an AI subject's account of the rule against the regulator's verbatim text from the regulator's own portal. Both are linked. Each delta, its root causes, and impact analysis are documented and published with immutable Citation IDs.