Modules/Insight CoreCore

Kernel: the analytical core

Every module rests on a single foundation that keeps the chain consistent: module coordination, full traceability of every operation, connectivity, 100% local AI and document grounding.

KernelIRRBBCSRBBGAPLCRNSFRALMMFlowQuant ALMSolvency
Role, inputs, outputs

The foundation shared by every module.

Inputs
Natural-language requestModules & enginesReferences (RAG)
Kernelmodule coordination
Outputs
Chain executed end to endTimestamped, reproducible traceNatural-language output
Module coordination

The Kernel coordinates the modules and their agents: a single request can call on several modules, from cash flows to metrics and analysis, each agent handling the step of its own domain under the Kernel's coordination. Changing an assumption and measuring its impact on every metric takes a single request, every step traced.

Full traceability

Every result is timestamped, sourced and reproducible; agent, tool, parameters and calculation steps are identified for every operation: no black box.

Connectivity & mapping

Intelligent ETL chain and standardised files; semantic mapping proposed by the platform and validated by the business.

Local AI, orchestration role

The AI never computes a metric and never improvises an analysis: it orchestrates the engines, models, agents and analysis routines, and reports in natural language. 100% local AI: the models run in the chosen deployment environment, on-premise, in the bank's private cloud or hosted by RIALTO. No data is sent to any external AI provider.

Document grounding (RAG)

The bank's internal and regulatory references become a queryable knowledge base, every answer citing its sources.

ICAAP · ILAAP

Contribution to the ICAAP and ILAAP

End-to-end traceability of both assessments.

Functional scope

What the module covers.

Multi-module orchestration in one requestTimestamping, sourcing, reproducibilityStep-by-step audit panelIntelligent ETL chain and standardised filesBusiness-validated semantic mapping100% local AISourced knowledge base (RAG)Scheduler and historyPermissions and roles
In conversation

One request, four modules, every step traced.

The expert focuses on risk analysis. RIALTO chains the calculations and outputs, and keeps each step's parameters, timestamp and results.

rialto-kernel · bridge NSFR
NSFR · bridge 30/06 → 30/09 · in % · truncated axis
130131132133134132.530/06/2024+0.49ASF effectEffet ASF-1.26RSF effectEffet RSF131.730/09/2024
30/06/2024 → 30/09/2024 -0.8Click ASF effect or RSF effect for the drill-down.
Largest movements · items30/06/202430/09/2024VariationNSFR weight
ASF from retail deposits ›37,224.2 M€37,577.9 M€+353.8 M€+0.66 pt
ASF from financial customers and central banks ›14,689.8 M€14,415.1 M€-274.7 M€-0.51 pt
ASF from other non-financial customers (excl. central banks) ›9,611.6 M€9,703.0 M€+91.3 M€+0.17 pt
ASF from own funds items and instruments ›9,091.2 M€9,177.6 M€+86.4 M€+0.16 pt
ASF from other liabilities ›480.6 M€485.2 M€+4.6 M€+0.01 pt
Other (1)381.2 M€384.8 M€+3.6 M€+0.01 pt
Total · ASF71,478.6 M€71,743.6 M€+265.0 M€+0.49 pt
Largest movements · items30/06/202430/09/2024VariationNSFR weight
RSF from loans ›47,370.6 M€47,820.8 M€+450.2 M€-1.11 pt
RSF from other assets ›2,797.2 M€2,823.8 M€+26.6 M€-0.07 pt
RSF from liquid assets ›2,020.8 M€2,040.0 M€+19.2 M€-0.05 pt
RSF from securities other than liquid assets ›730.0 M€736.9 M€+6.9 M€-0.02 pt
RSF from off-balance-sheet items ›496.1 M€500.8 M€+4.7 M€-0.01 pt
Other (2)298.5 M€301.3 M€+2.8 M€-0.01 pt
Total · RSF53,713.2 M€54,223.7 M€+510.5 M€-1.26 pt
RIALTO · Kernel
Decrease of 0.8 pt (132.5% → 131.7%): the ASF effect (+0.49 pt) does not offset the RSF effect (-1.26 pt), driven by loans.
Illustrative example · fictitious data
4
stages: data, cash flows, metrics, analysis, simulation and reporting
3
combinable entry levels →
100 %
local AI, no data shared
0
black box

See RIALTO reason over your numbers.

We simulate a full analysis with your data types and the scenario of your choice.

Request a demo Our publications