Comparison · ISO 20022 Structured Addresses · November 2026 Deadline

StructX vs ioNova ARS
— an LLM structurer on a platform, and a compliance engine

Purple Fabric StructX (IntellectAI) is a narrowly scoped, LLM-based structurer that converts unstructured input into ISO 20022 PostalAddress24 format, running on a general enterprise AI platform oriented to insurance and wealth. ioNova ARS is a deterministic, end-to-end SWIFT / ISO 20022 compliance engine — and a full application. Here is the honest comparison.

ioNova ARS · Recommended StructX · Narrow scope only Source: 25-criterion analysis, March 2026
4.90 vs 1.36
Composite score / 5.00
4.90 vs 0.00
Risk-adjusted score / 5.00
StructX floors at 0.00 after −2.75 in penalties
25/25 vs 5
Requirements met
StructX: 5 met · 9 partial · 11 not met
High vs Low
Evidence confidence
28 of 70 research questions unanswered
A Capability vs. An Application

An LLM structurer on a platform — and the engine that runs the lifecycle

Both output ISO 20022 PostalAddress24. The decision is “one integrated deterministic engine” versus “an LLM capability plus the review workflow, latency mitigation and AI governance you build around it.”

Purple Fabric StructX

IntellectAI · LLM multi-agent + RAG + rules validation

A single-purpose ISO 20022 structurer on a broad enterprise AI platform — LLM-based multi-agent architecture with RAG (MongoDB Atlas Vector Search) and a rules engine for validation.

  • Outputs ISO 20022 PostalAddress24 — 14 structured fields claimed; 99%+ format-compliance claim (unvalidated).
  • Choose-your-own-LLM via Model Optimization Hub; Enterprise Knowledge Garden for RAG grounding.
  • Platform-level governance: ISO/IEC 42001, ISO 27001, SOC 2.
  • Professional-services engagement; no public API reference, schema or per-parse pricing.

ioNova ARS

Deterministic rule engine · 30-year payments heritage (ACE)

A comprehensive, deterministic compliance engine — and a full application — that resolves, corrects and returns compliant ISO 20022 XML in a single call, production-ready for November 2026.

  • All 14 PostalAddress24 elements plus BIC / LEI / IBAN entity fields; 50+ financial identifiers ring-fenced pre-parse.
  • Bit-identical output; 30 reason codes citing EPC153-22, PMPG v1.11, EU Reg 2023/1113, CBPR+ SR2026 by provision.
  • Full application: Exceptions Workbench, live dashboards and Developer Portal — not just a model.
  • Six integration mechanisms; start in days, live in 2–4 weeks; sub-50ms median latency; batch 10,000 addresses per call.
Side by Side

Dimension by dimension

The gap is not effort — it is architecture, scope and evidence.

DimensionStructXioNova ARS
ArchitectureLLM multi-agent + RAG + rules orchestrationDeterministic rule engine + reference data + entity extraction
ScopeStructuring only — PostalAddress24 outputFull lifecycle: resolve, correct, return compliant XML in one call
DeterminismNon-deterministic LLM core — “deterministic enterprise knowledge governance” refers to data governance, not parse outputBit-identical output for identical input
LatencyLLM inference typically 100ms–2s; no mitigation documentedSub-50ms median
Financial identifiersNo documented BIC/IBAN/LEI extraction50+ types extracted and preserved pre-parse
EvidenceTier-3 marketing; 28 of 70 questions unansweredDocumented API, latency and scheme coverage
Operational tooling NewNone documented for StructX specificallyExceptions Workbench + live dashboards, built in
Onboarding New~18 weeks incl. mandatory PoC, latency mitigation and AI-governance setupDeveloper Portal sandbox in days; live in 2–4 weeks
Composite score1.36 / 5.004.90 / 5.00
Risk-adjusted score0.00 / 5.00 (floored)4.90 / 5.00 — no penalties
VerdictConditionally recommended — narrow, tactical scope where Purple Fabric already runsRecommended — primary compliance engine
Category View

Behind in every category — near zero on signature capabilities

Average score per category (out of 5). StructX’s best category is enterprise operations at 2.2 — carried by platform-level governance — while the signature capabilities that carry 48% of the weight score 0.7.

Signature capabilities48% of composite · financial IDs, historical names, dedup, explainability…
ioNova
5.0
StructX
0.7
Enterprise operations25% · determinism, security, sovereignty, deployment
ioNova
4.7
StructX
2.2
Payments domain15% · SWIFT MT, sanctions, correspondent banking, ISO 20022
ioNova
5.0
StructX
1.7
Parsing quality12% · coverage, normalisation, robustness to messy input
ioNova
4.8
StructX
2.0
The Application Advantage

StructX is a capability on a platform.
ioNova is a working application.

Adopting an LLM structurer means adding your own review workflow, latency mitigation and AI-governance wrapper. ioNova ships the operational layer on day one — with no LLM in the critical path to govern.

Exceptions Workbench

Low-confidence resolutions route to maker-checker review queues with per-field confidence, the proposed repair and the rule citation behind it — every action audit-logged.

StructX: no review UI documented; human-override details undisclosed (ENT-4 3/5).

Live Dashboards

Real-time STP rate, exception ageing, latency percentiles, reason-code distribution and per-country compliance posture — examiner-ready evidence, always current.

StructX: health checks platform-level only; metrics not documented.

Developer Portal

Self-service onboarding, API keys, sandbox with SWIFT MT / ISO 20022 test corpora, OpenAPI specs and integration guides. First message parsed the day access is granted.

StructX: no public API reference, schema or implementation documentation — engagement starts with professional services.

Six ways to integrate — documented, versioned, self-serve

REST API
24 endpoints · OpenAPI 3.x
SFTP
Secure batch file transfer
IBM MQ
Enterprise message queue
Database
Direct database integration
Kafka
Event streaming at scale
MCP
Model Context Protocol · agent-ready

ioNova ARS — start in days, live in 2–4 weeks

  1. D1Developer Portal sign-up: sandbox keys, test corpora, OpenAPI specs — first message parsed same day.
  2. W1Integrate via the mechanism that fits your stack (REST API, SFTP, IBM MQ, Database, Kafka or MCP).
  3. W2–4Parallel-run against production traffic in the Workbench; tune thresholds; go live.

StructX — ~18 weeks including mandatory PoC

  1. W1–4PoC validation — accuracy, determinism, identifier handling and latency must be proven on your corpus first.
  2. W5–12Integration plus latency mitigation (3 wks) and AI-governance setup (3 wks).
  3. W13–18UAT and go-live — with LLM non-determinism still in the critical path.

No equivalent found in IntellectAI’s published documentation as of the March 2026 analysis; to be confirmed against current releases before external publication. File uploads are also supported directly in the application UI.

Credit Where Due

Where Purple Fabric StructX is genuinely capable

A real product on a governed enterprise platform, with a clear design focus. None of this makes it weak; it makes it a narrow, platform-bound structuring capability rather than a compliance engine.

Focused ISO 20022 structuring

PAY-4 at 4/5 is its best score — unstructured input to PostalAddress24 is its core purpose, and the 99%+ format-compliance claim shows real design focus on the target schema.

Enterprise AI governance

ISO/IEC 42001, ISO 27001 and SOC 2 at platform level (ENT-5 4/5) — one of the stronger AI-governance postures in the evaluated set, inherited from Purple Fabric.

Multi-agent RAG architecture

Choose-your-own-LLM via the Model Optimization Hub, RAG grounding through MongoDB Atlas Vector Search and postal-DB validation to mitigate model drift.

Platform-level PII handling

Entity redaction and masking are built into the Purple Fabric platform — sensitive-data controls come with the estate rather than needing separate assembly.

Strong fit where the platform runs

For institutions already on Purple Fabric, StructX leverages existing IntellectAI contracts, governance frameworks and in-house skills — a low-friction tactical add.

Honest evidence caveat

Several low scores reflect undocumented capability, not proven absence — for an existing Purple Fabric customer, a scoped PoC is the fair next step.

The Structural Gaps

Out of scope or out of evidence — and two gaps the LLM core cannot close

Some gaps are design scope, some are evidence gaps — and two are structural consequences of the architecture. Scores out of 5, from the March 2026 analysis.

SIG-1 · 0/5 Financial-ID preservation

No documented BIC, IBAN or LEI capability — identifiers in address text become address tokens. The defining gap for payment compliance. ioNova extracts and preserves 50+ types before parsing begins.

ENT-2 · 1/5 Determinism

The LLM core is non-deterministic by architecture — the make-or-break gap for reconciliation and replay. ioNova returns bit-identical output for identical input on versioned, rule-cited logic.

PAY-3 · 0/5 Correspondent banking

No multi-party or party-role awareness. ioNova distinguishes ordering, beneficiary, intermediary and ultimate parties with role-aware rules.

SIG-5 · 1/5 Explainability

No per-field rule citations — a structural consequence of the LLM core. ioNova’s 30 reason codes cite EPC, PMPG, EU Reg 2023/1113 and CBPR+ by provision.

ENT-9 · 1/5 Latency

LLM inference typically runs 100ms–2s against sub-100ms real-time payment budgets, with no mitigation documented. ioNova’s median is sub-50ms.

ENT-3 · 2/5 Hallucination risk

RAG grounding and postal-DB validation mitigate but cannot eliminate hallucination — and no hallucination rate is published for the structuring task.

Full Transparency

All 25 criteria, scored side by side

No ties. The gaps cluster in the signature capabilities and payments-domain criteria that carry most of the weight for payment compliance.

CriterionStructXioNova ARSΔ
Signature capabilities · 48%
Financial-ID preservationSIG-1 · BIC / IBAN / LEI / ABA0 / 55 / 5+5
Geographic disambiguationSIG-22 / 55 / 5+3
Historical name resolutionSIG-30 / 55 / 5+5
Data de-duplicationSIG-4 · cross-field, multilingual0 / 55 / 5+5
Full explainabilitySIG-5 · rule citations, audit trail1 / 55 / 5+4
Confidence scoring & routingSIG-61 / 55 / 5+4
Parsing quality · 12%
International coveragePQ-1 · countries, scripts2 / 55 / 5+3
Component taxonomyPQ-22 / 55 / 5+3
Normalisation & abbreviationPQ-33 / 55 / 5+2
Robustness to messy inputPQ-41 / 54 / 5+3
Enterprise operations · 25%
Regulatory explainabilityENT-1 · FCA / OCC / BaFin / MAS2 / 55 / 5+3
Determinism & consistencyENT-21 / 55 / 5+4
Hallucination resistanceENT-32 / 55 / 5+3
Human overrideENT-4 · Exceptions Workbench3 / 54 / 5+1
Bank-grade securityENT-54 / 55 / 5+1
Governance & change managementENT-63 / 54 / 5+1
Data sovereigntyENT-7 · no external API2 / 55 / 5+3
Deployment flexibilityENT-8 · SaaS / VPC / on-prem / air-gapped2 / 55 / 5+3
Performance & scalabilityENT-91 / 54 / 5+3
Payments domain · 15%
SWIFT message handlingPAY-1 · fields 50 / 52 / 56 / 57 / 592 / 55 / 5+3
Sanctions & compliance integrationPAY-21 / 55 / 5+4
Correspondent bankingPAY-3 · multi-party chain0 / 55 / 5+5
ISO 20022 migration readinessPAY-44 / 55 / 5+1
Payment routing & enrichmentPAY-51 / 55 / 5+4
Batch & real-time processingPAY-62 / 55 / 5+3

Source: ioNova 25-criterion competitive analysis, March 2026 (70-question protocol). StructX’s −2.75 in penalties floors its risk-adjusted score at 0.00; its scores rely heavily on platform-level inference, and evidence confidence is low — 28 of 70 research questions could not be answered from public sources. ioNova’s 4/5 ratings were documentation gaps at analysis time; the Exceptions Workbench and live dashboards in the current application release address the ENT-4 and monitoring notes. Note: Purple Fabric StructX (IntellectAI) is a different product from Catalyst Data Intelligence (IntellectEU), which is compared separately.

The report’s bottom line
“A deterministic engine compounds. A scoped structurer plateaus. Both are real choices. Only one improves over time.” StructX is conditionally recommended only where the institution already runs Purple Fabric, the scope is narrow and tactical, and a structured PoC validates latency, determinism and identifier handling first.
FAQ

What teams ask about StructX vs ioNova ARS

Not directly. StructX is a single-purpose structuring capability on a broad enterprise AI platform oriented to insurance and wealth; ioNova ARS is an end-to-end SWIFT / ISO 20022 compliance engine and full application. They overlap on one task — converting unstructured input into PostalAddress24 — and the composite scores (1.36 vs 4.90) measure how much of the surrounding compliance problem each one actually solves.

For institutions that already run Purple Fabric, with a narrow, tactical structuring scope — and the willingness to validate latency, determinism and identifier handling in a structured PoC first. In that setting it leverages existing IntellectAI contracts, governance and skills. As a primary compliance engine for payments, it is not the fit.

Flexible on messy input, yes — but the flexibility is paid for in non-determinism, 100ms–2s inference latency and the absence of field-level rule citations. Payment reconciliation and audit depend on bit-identical replay: the same input must produce the same output, with a citable reason for every change. That is exactly what an LLM core cannot guarantee.

No — it is a scope-and-evidence verdict, not a quality one. 28 of 70 research questions could not be answered from public sources, and several zeros record undocumented capability rather than proven absence. But an evaluation against a compliance mandate has to score what can be evidenced — and the burden of proof sits with the vendor.

No. Purple Fabric StructX is from IntellectAI — an LLM-based structurer, composite 1.36. Catalyst Data Intelligence is from IntellectEU — a supervised-ML structurer, composite 2.72. Different companies, different architectures, compared separately in our series. The similar names cause real confusion in procurement shortlists — check which one is actually on yours.

Sub-50ms median latency, a 24-endpoint versioned API, batch at 10,000 addresses per call, 41 SEPA country profiles, 25+ message types, 11 schemes — plus the Exceptions Workbench, live dashboards and Developer Portal.

Meet the mandate with the engine that compounds,
not the structurer that plateaus.

Deterministic SWIFT MT parsing, 50+ identifiers preserved, corrected ISO 20022 XML, 25 of 25 criteria met — plus the Exceptions Workbench, live dashboards and Developer Portal. Start in days, live in 2–4 weeks — no PoC required.

See structured resolution in action