Reconciliation, verified
by intelligence.
The reconciliation platform for global banks and financial institutions — where agentic AI clears breaks, hash-chained audit seals every action, and your data never leaves your perimeter of trust.
Estimate your annual reconciliation savings
Drag the sliders — see the impact FVI's 99.4% auto-match rate delivers.
* Indicative model based on $95/hr analyst cost and 1,800 hrs/yr capacity. Request a tailored business case.
The Enterprise Reconciliation Buyer's Guide
12 pages covering: architecture trade-offs, FVI vs Duco vs legacy comparison, implementation timeline, and a self-scoring RFP checklist. Trusted by enterprise procurement teams at Tier-1 banks.
- FVI vs Duco feature-by-feature matrix
- Total cost of ownership model
- Implementation and migration playbook
- Regulatory checklist: SOC 2, DORA, ISO 27001
The falcon's eye — total visibility across cash, securities, intercompany and GL flows.
Deterministic rules, probabilistic matching, and tolerance-aware clustering, at millisecond latency.
Agentic AI that reasons about breaks, drafts fixes, and writes back under human control.
Scrub the pipeline. Watch reconciliation resolve in real time.
A live sample of 7 bank statement rows against 7 ledger entries. Move the slider through each stage — exact keys, tolerance, AI alias resolution — and see matched and unmatched outputs update instantly.
Five stages. One continuous intelligence.
SFTP · S3 · Kafka · SWIFT · ISO 20022 · webhooks — normalized at the edge.
One platform. Every reconciliation.
Cash & nostro, securities & trades, intercompany & GL — matched, triaged and explained by AI agents under enterprise-grade governance.
1:1, 1:N, N:N, fuzzy, tolerance, composite keys
Auto root-cause, NL rule builder, AIDP for PDFs
OpenAI · Anthropic · Azure · cost & token caps
Cron schedules, retries, 90d+ run history
42 transforms, watchers, quality rules
Kanban, side-by-side inspector, 4-eyes
MTTR, ageing, dashboards, scheduled exports
Presence, comments, attachments, timeline
Email digests · Slack · Teams · in-app
Groups, per-rekon perms, custom tenant URLs
Immutable, tenant-isolated, tamper-proof
TOTP · SSO/SAML · IP allowlist · PII redaction
Agents that ingest, match, resolve and audit — autonomously.
Design reconciliations in natural language. Let the Resolution Agent triage breaks, propose root causes, and draft fixes. Every action is policy-checked, four-eyes approved, and cryptographically chained.
- Design Agent — author reconciliations from a single sentence
- Resolution Agent — cluster and triage breaks with confidence scores
- AIDP — extract structured data from PDFs, statements, confirms
- Break Knowledge Graph — learn from every resolution, forever
Enterprise-grade by default.
Multi-tenant isolation, hash-chained audit, configurable data residency, tamper-proof triggers, and zero-trust auth — on day one.
How FVI stacks up against Duco and legacy reconciliation
| Capability | FVI | Duco | Legacy |
|---|---|---|---|
| Agentic AI triage | — | — | |
| BYO-LLM / data sovereignty | — | — | |
| Hash-chained audit | — | — | |
| Multi-tenant RBAC | — | ||
| 20+ matching rule types | — | ||
| AIDP PDF extraction | — | — | |
| Streaming / near-real-time | — | — | |
| Zero-trust + SSO/SAML | — |
Get the full 20-capability comparison sheet including TCO model and migration assessment.
Download full comparisonEvery signal, every source, one intelligent graph.
Hover any node to trace how data, signals, and decisions travel through the Veritas Core — from ingest to write-back, under hash-chained governance.
Built for scale. Designed for clarity.
Ingestion Hub
12 connectors — SFTP, S3, Azure, Snowflake, BigQuery, Kafka, API, webhooks.
Near-real-time rekon
Streaming engine with watermarking and exactly-once semantics.
Dev → UAT → Prod
Sign and promote reconciliation packages with four-eyes and one-click rollback.
10× better. Not 10% better.
Legacy tools bolted AI on top. FVI was architected the other way around — agents first, audit chain woven in, your LLM your choice. That's not an upgrade. That's a replacement.
From sentence to live rec — in hours
FVI Copilot authors reconciliation rules from a plain-English description. Average time to first match: under 4 hours.
99.4% auto-match — before a human sees the queue
Deterministic + probabilistic + AI fuzzy matching in a single engine. The 0.6% that remains is real breaks, not noise.
Zero vendor lock-in — including your LLM
BYO-LLM: OpenAI, Anthropic, Azure OpenAI, or your private endpoint. Per-tenant cost & token caps. PII masked pre-inference.
Recognised across reconciliation, RegTech & AI
What enterprise teams say
“We eliminated 94% of manual cash-break touches within the first quarter. The agentic triage doesn't just surface breaks — it tells you why they happened before your team even opens the queue.”
Everything you need to know
Stop reconciling. Start verifying.
Launch your tenant in minutes. Bring your data, your LLM, your rules — keep your sovereignty and your certainty.