RiskSonnar vs Featurespace (ARIC)
Adaptive Behavioural Analytics platform with strong fraud-detection ML; heavy enterprise sales motion.
Side-by-side feature matrix
Sixteen dimensions, one row each. Where Featurespace (ARIC) has the stronger story we mark the row as their advantage; where we win we mark ours; where the products are equivalent we mark parity.
| Dimension | Featurespace (ARIC) | RiskSonnar |
|---|---|---|
| Pricing transparency | Enterprise contracts only; pricing fully bespoke. | Public pricing at /pricing — Starter / Pro / Enterprise with EUR amounts in plain view. |
| Deployment time | Multi-quarter implementations typical, especially in tier-1 environments. | Sandbox spins up in minutes at /sandbox; production tenant onboarded in days, not quarters. |
| Audit verifiability | Model output traceable inside the platform; not cryptographically chained or externally reproducible. | Hash-chained per-event receipts; manifest SHA-256 verifiable at /v1/compliance/evidence-pack. |
| No-code scenarios | Detection is model-driven; rule authoring is a secondary surface, often consultant-led. | No-code scenario studio at /scenarios/new (csn_* IDs) with JSON export + version history. |
| Hash-chained receipts | Not part of the platform. | Every state-changing action emits a hash-linked receipt — tamper-evident by construction. |
| Sandbox availability | POC environments arranged with sales and SI; no public self-serve sandbox. | Public synthetic-tenant sandbox at /sandbox, 60-minute sessions, no signup. |
| Public trust dashboard | Not offered. | Opt-in public trust dashboard at /trust/<slug> — integrity status visible to anyone. |
| Federated-learning support | Research investment in federated and privacy-preserving ML; productisation varies by deployment. | Federated-learning support on the roadmap (rev-54); cross-tenant typology sharing without raw data movement. |
| Perpetual KYC | Not the primary focus — Featurespace is detection-centric, not KYC-centric. | Perpetual KYC included in Pro+: continuous re-screen on watchlist deltas and PEP/adverse-media changes. |
| LLM SAR drafting | Not native; SAR authoring is downstream of detection. | LLM SAR drafting in-product (MLRO signs) with the model output appended to the audit chain. |
| Crypto screening | Not a primary surface. | Crypto screening at /watch/crypto — chain-analytics signals fused with off-chain identity. |
| Open API | Integration via consultancy-led delivery; not a self-serve API product. | OpenAPI 3.1 spec public; cockpit and SDKs build against the same surface a customer integrates with. |
| Regulator-replay (byte-exact) | Model explainability features exist; byte-exact replay of a historical decision is not the product posture. | Regulator-replay endpoint reproduces any decision byte-exactly from the chained inputs. |
| Self-hosting option | On-prem and private cloud common for tier-1 banks. | Self-hosted tenancy available on Enterprise (BYO KMS + PrivateLink). |
| Source available | Proprietary closed-source. | Source-available components published under the RiskSonnar evaluation license. |
| License model | Multi-year enterprise license + maintenance + SI services. | Annual subscription with public list-price brackets; no per-feature lock behind quotes. |
When to pick Featurespace (ARIC)
Pick Featurespace if you are a tier-1 bank or a card scheme with a high-volume payments-fraud problem where behavioural-ML uplift is the dominant driver of value, and you have the systems-integrator engagement budget to stand up ARIC properly. The research depth and detection quality in card-fraud are real.
When to pick RiskSonnar
Pick RiskSonnar if your problem set is broader than card-fraud detection — if AML, sanctions, DSAR, link-analysis, and SAR-filing are all on the table, and if you want explainable, auditable, no-code scenarios alongside ML-assisted detection. We are honest that a tier-1 card-fraud shop with a multi-quarter SI engagement is the wrong customer for us; everyone else likely fits better here.
“Featurespace would have been right if our only problem was card fraud. We had six problems, and we wanted one chain.”
Next steps
The fastest way to evaluate RiskSonnar against Featurespace (ARIC) is to spin up a synthetic-data sandbox and verify the claims yourself — audit chain, scenario authoring, screening, cases, SAR drafting. Sixty minutes, no signup.