Question 1

What stage is FVI at today?

Accepted Answer

FVI is a fully architected, enterprise-grade platform preparing to onboard its first design partners. The product, security model and audit chain are built and testable end-to-end; we are intentionally selective about early clients so each launch partner receives white-glove implementation and direct engineering access.

Question 2

How long does it take to build a reconciliation on FVI?

Accepted Answer

The AI Copilot generates rules from a plain-English description, producing a working configuration for review within an hour. A complete reconciliation — data prep, tolerances, agent policies — is typically ready in under a day. No mandatory professional-services engagement.

Question 3

Do I have to use your AI model?

Accepted Answer

No. FVI is Bring-Your-Own-LLM by design: OpenAI, Anthropic Claude, Azure OpenAI, Google Vertex, or a private endpoint you control. Per-tenant cost caps and token limits apply. PII is masked before any payload leaves your boundary, and providers can be switched without rebuilding any reconciliation.

Question 4

How does the hash-chained audit work?

Accepted Answer

Every match decision, tolerance override, agent action and human approval is SHA-256 hashed and linked to the previous record, forming a tamper-evident chain. Any mutation is detectable on verification. The full lineage is exportable for regulators and internal audit without requiring FVI access.

Question 5

What reconciliation types does FVI support?

Accepted Answer

One-sided, two-sided and tri-party. Covers cash & nostro, securities & trade confirmation, custodian + prime + internal, intercompany, GL-to-subledger, and any structured or unstructured data pair. PDFs, bank statements and trade confirms are parsed by the document extraction engine before matching.

Question 6

What does multi-tenant mean in practice?

Accepted Answer

Each client is an isolated tenant with dedicated RLS boundaries, its own tenant URL, RBAC groups and reconciliation library. Data is never co-mingled across tenants. Pooled (cost-efficient) and siloed (maximum isolation) deployment modes are both supported.

Question 7

Can FVI connect to my existing data stack?

Accepted Answer

Planned and supported connectors include SFTP, S3, Azure Blob, Snowflake, BigQuery, Kafka, REST/webhooks, and financial-messaging formats (SWIFT MT940, ISO 20022). The Data Prep engine handles normalisation and transforms in-platform — no external ETL middleware required.

Question 8

How does FVI approach compliance and certifications?

Accepted Answer

The architecture is designed with reference to DORA, SOC 2, ISO 27001 and SEC 17a-4 — covering governance, immutable audit, security controls and operational resilience. Formal certifications and third-party attestations are on the roadmap and will be pursued in step with our first production deployments.

Question 9

Can I see FVI with my own data before committing?

Accepted Answer

Yes. Book a working session and we will stand up a proof-of-concept against a representative sample of your data — showing match rates, break triage, the audit chain and BYO-LLM routing against your real workflow.

Question 10

What does an early-partner engagement look like?

Accepted Answer

Direct access to the founding team, jointly-defined success criteria, preferential commercial terms, and influence over the product roadmap. Onboarding, tenant provisioning and initial reconciliation build are handled by FVI engineering at no additional charge.

Question 11

Which deployment model should we choose — SaaS or VPC / Private Cloud?

Accepted Answer

SaaS (multi-tenant) is the fastest path to production and is appropriate for most business units: fully managed, continuously updated, hash-chained audit and per-tenant RLS isolation. VPC / Private Cloud (single-tenant) is designed for regulated workloads that require the FVI control plane to run inside your own AWS, Azure or GCP account — with your VPC, your KMS keys, your egress policy and your identity provider. Both models share the same product, APIs and audit chain; only the operating boundary and a small set of controls differ (see the deployment matrix above).

Question 12

Where is customer data stored and can we pin a region?

Accepted Answer

On SaaS, tenants are pinned to a home region at provisioning time (EU, UK, US, APAC and others on request) and data does not leave that region for processing, backup or DR. On VPC / Private Cloud, the deployment lives entirely in the region(s) of the cloud account you nominate. Cross-border transfers, if any, are documented in the DPA and gated by tenant configuration.

Question 13

How is client data stored and encrypted at rest?

Accepted Answer

Structured data lives in a hardened Postgres cluster; files and evidence artefacts live in per-tenant object storage prefixes. Everything is encrypted at rest with AES-256 using envelope encryption backed by cloud KMS — database, object storage, backups, snapshots and log archives. Field-level encryption and PII redaction can be enabled per column for the most sensitive data (PANs, IBANs, tax IDs). Backups inherit the same encryption and residency controls as the primary store.

Question 14

How is data segregated between tenants?

Accepted Answer

Every row, file, job and log line is stamped with a tenant_id and enforced by Postgres Row-Level Security — not just by application code. Object storage uses per-tenant prefixes with short-lived signed URLs, queues and workers execute in isolated contexts, and search indexes are partitioned per tenant. Tenant identity is resolved from the authenticated session, never from client input, so a compromised UI cannot cross a boundary. In VPC / Private Cloud, each tenant additionally runs in its own database, storage, network and KMS scope.

Question 15

Who at FVI can access client data, and under what controls?

Accepted Answer

Standard operations do not touch tenant data. Support and engineering access is gated by role, requires SSO + MFA, and — for anything beyond read-only diagnostics — a time-boxed, four-eyes-approved break-glass grant that is fully session-recorded. Every access event lands in the same hash-chained audit log the tenant sees, so customers can independently verify who touched what and when. Customer-managed keys (BYOK/CMEK) can be revoked to cut FVI access at the cryptography layer.

Question 16

How are audit logs generated?

Accepted Answer

Every meaningful event — authentication, permission change, data read on sensitive objects, rule edit, match/unmatch, tolerance override, agent action, human approval, export, admin action — is written synchronously to an append-only audit store. Each record captures who (user, tenant, session), what (action, target, before/after where applicable), when (UTC timestamp), where (IP, region, deployment) and why (justification / ticket reference where required). Records are SHA-256 hash-chained to the previous entry, so any tampering breaks the chain and raises an alert.

Question 17

How long are audit logs retained and can we export them?

Accepted Answer

Default retention is seven years, aligned to SOX, BCBS 239, SEC 17a-4 and DORA expectations, and configurable per tenant to match your firm's records-management policy. Logs can be streamed in near-real-time to your SIEM (Splunk, Microsoft Sentinel, Elastic, Chronicle) with per-tenant filtering, or exported on demand as signed, verifiable evidence bundles for regulators and internal audit — with no dependency on FVI staff to produce them.

Question 18

What does the first 30–60 days of onboarding look like for a Tier-1 client?

Accepted Answer

A structured five-phase pathway with named owners on both sides. Days 0–7: contracts, security packet, SSO/SCIM, tenant provisioning and role model. Days 8–14: source-system connectivity, sample data ingestion, PII masking and residency confirmation. Days 15–30: AI-assisted mapping, tolerance and rule configuration, agent policies and four-eyes approval flows. Days 31–45: pilot reconciliation in shadow mode against production data, match-rate tuning, break-triage playbooks and SIEM streaming. Days 46–60: parallel run, sign-off on RPO/RTO drills, regulator-ready evidence bundle, and formal go-live with 24×7 support. The full checklist is shown above.

Question 19

What security artifacts can FVI provide to enterprise prospects during due diligence?

Accepted Answer

FVI maintains a Security Overview Pack that prospects can request under NDA. It includes: (1) a Security & Architecture Overview covering tenant isolation, identity, encryption, key management and the hash-chained audit design; (2) completed due-diligence questionnaires (SIG Lite and CAIQ) reflecting current, in-product controls; (3) a sample Audit Log & Evidence Bundle showing the event schema, hash-chain verification and export format; (4) a Data Retention & Deletion schedule with the default seven-year audit retention and per-tenant configuration options; (5) a Data Processing Addendum (DPA) template with standard contractual clauses, the current subprocessor list and residency commitments; (6) an Incident Response & Breach Notification summary with contact channels and target notification windows; and (7) a Shared Responsibility Matrix separating FVI, customer and cloud-provider obligations. These artifacts describe controls that are built and testable in-product today. Formal third-party attestations (SOC 2 Type II, ISO 27001) are on the roadmap and will be published as they complete — FVI does not claim certifications it has not yet earned.

Question 20

What is in-scope for the first Tier-1 client engagement?

Accepted Answer

The first client engagement is intentionally focused so both sides can move quickly and evidence outcomes cleanly. In scope: one production reconciliation (typically cash & nostro, custodian-vs-internal, or GL-to-subledger) in a single home region; SSO/SCIM against your identity provider; BYOK/CMEK against your KMS; the full hash-chained audit with SIEM streaming to one destination; the Security Overview Pack and DPA above; a shadow run followed by a parallel run and formal go-live within the 30–60 day pathway; and 24×7 support with a named FVI Delivery Lead and Executive Sponsor. Out of scope for the first engagement (available in follow-on phases): additional reconciliations, additional legal entities or regions, custom connectors beyond the supported catalogue, on-premises deployment, and bespoke regulatory reporting formats. Scope, exit criteria and commercial terms are documented in a joint Statement of Work signed before Day 0.

Question 21

How does FVI fit our Third-Party Risk Management (TPRM) programme?

Accepted Answer

FVI is designed to slot into a Tier-1 TPRM lifecycle without bespoke work. Inherent-risk tiering: we accept your classification (typically High / Material given reconciliation touches books-and-records) and align controls accordingly. Due diligence: the Security Overview Pack ships with completed SIG Lite and CAIQ questionnaires, architecture diagrams, a subprocessor register with change-notification commitments, financial viability disclosures, and BCP/DR evidence. Contracting: our MSA, DPA and SCCs cover audit rights, sub-processor consent, right-to-terminate for material change, source-code escrow on request, and defined SLAs with service credits. Ongoing monitoring: SOC 2 / ISO reports (as they complete), quarterly control attestations, annual pen-test summaries, real-time SIEM streaming of tenant audit logs, and a named Vendor Relationship Manager on the FVI side. Exit: a documented exit plan with data-return format, retention wind-down and certificate of destruction is executed on notice — no lock-in on your data.

Question 22

How does FVI support DORA, PRA SS2/21 and OCC third-party expectations?

Accepted Answer

FVI is architected to be a well-behaved ICT third-party provider under DORA and equivalent regimes (PRA SS2/21, OCC 2023-17, MAS TRM, HKMA SA-2). Concretely: (1) the standard contract includes the DORA Article 30 mandatory clauses — service description, locations of processing, security levels, assistance during incidents, cooperation with competent authorities, termination rights and exit strategy; (2) FVI is registerable in your ICT third-party register with all required fields (LEI, function criticality, substitutability assessment inputs); (3) incident classification and reporting timelines are aligned to the DORA RTS with a documented notification workflow to your operational-resilience team; (4) TLPT / threat-led penetration testing participation and evidence sharing is supported for in-scope entities; (5) concentration-risk and substitutability information (data portability, standard export formats, exit runbook) is provided so you can evidence a credible exit strategy to your regulator.

Question 23

How do you handle security assessments, penetration testing and our right to audit?

Accepted Answer

FVI supports a full enterprise assurance cadence. Annual independent penetration tests are performed by a CREST/OSCP-accredited third party against the production platform and the tenant boundary; executive summaries and remediation status are shareable under NDA, and Tier-1 clients may review the full report on-site or via secure data room. Continuous controls: SAST, DAST, SCA and container/IaC scanning gate every release; secrets scanning and dependency-vulnerability SLAs (Critical 24h, High 7d) are enforced. Customer-led testing: subject to a Rules-of-Engagement document and scheduling, clients may pen-test their own tenant in a dedicated test environment. Right to audit: contractual audit rights include an annual on-site or remote audit of controls relevant to the services, access to SOC 2 / ISO reports as they complete, and unrestricted rights for competent regulators under DORA / PRA / OCC to examine FVI as a material third-party.

Question 24

What are your operational resilience commitments — RTO, RPO, BCP and exit?

Accepted Answer

Target service levels for Tier-1 SaaS tenants: 99.95% monthly uptime for the control plane; RTO ≤ 1 hour and RPO ≤ 5 minutes for the primary region, with cross-AZ high availability by default and cross-region DR on request. Quarterly DR drills are executed with tenant-visible evidence in the audit log; annual full-scenario BCP exercises cover regional outage, key-provider failure and severe cyber events, mapped to your Important Business Service definitions and Impact Tolerances under UK/EU operational-resilience rules. Exit: a documented, tested Exit Plan guarantees data return in open formats (Parquet/CSV for data, JSON for configuration, verifiable hash-chain evidence bundles for audit), a transition-assistance window of up to 12 months, and a certificate of secure destruction — all specified contractually before signature.

Question 25

How does the procurement and legal review typically run for a Tier-1 bank?

Accepted Answer

A predictable, parallelised path designed to compress a typical 4–6 month enterprise procurement into 6–10 weeks. Week 0: NDA and Security Overview Pack released; joint kickoff with Procurement, InfoSec, Legal, Data Privacy, Operational Resilience and the business sponsor. Weeks 1–3: SIG Lite / CAIQ / bespoke questionnaire responses; architecture and threat-model deep-dive with InfoSec; DPA and SCC review with Data Privacy; TPRM inherent-risk and concentration-risk assessment. Weeks 2–5: MSA / Order Form / DORA addendum redlines with Legal; SLA, service-credit and exit-plan negotiation; sub-processor and residency confirmations. Weeks 4–8: pricing, commercial approvals, and internal risk-committee sign-off (Vendor Risk, Model Risk if AI features are in scope, Operational Risk). Weeks 8–10: signature, tenant provisioning kickoff, Day-0 of the 30–60 day onboarding pathway. A named FVI Deal Desk coordinates all workstreams so your procurement lead has a single point of contact.

Question 26

How is the AI governed and validated under our Model Risk Management (SR 11-7 / SS1/23) framework?

Accepted Answer

FVI treats every AI capability as a model in the SR 11-7 / PRA SS1/23 sense and provides the artefacts your Model Risk function will ask for. Per-model documentation: intended use, training / prompting methodology, data sources, limitations, known failure modes, and out-of-scope uses. Validation evidence: reproducible test suites, benchmark match-rates on representative data, confidence-score calibration, drift monitoring and champion/challenger comparisons. Human-in-the-loop by default: AI is advisory for auto-mapping and break triage; four-eyes approval is enforced for anything that mutates books-and-records or overrides a tolerance. Explainability: every AI decision records the model, version, prompt/inputs, output, confidence and the human who approved or rejected it — all inside the hash-chained audit. Controls: per-tenant model allow-lists, BYO-LLM with your provider and key, PII masking before egress, cost and token caps, and a documented model-change management process with tenant notice before any material change.

Question 27

Who are FVI's sub-processors, and how do you manage concentration risk?

Accepted Answer

The current sub-processor register is published as part of the Security Overview Pack and is versioned: cloud infrastructure (AWS / Azure / GCP — customer-selectable per tenant), managed Postgres and object storage within the same cloud, an SIEM/log-archive provider, a transactional email provider, and — where the tenant opts in — the customer's own choice of LLM provider (OpenAI, Anthropic, Azure OpenAI, Google Vertex, or a private endpoint). Each sub-processor is contractually bound to security and confidentiality terms flow-down, is assessed annually, and appears in your DORA ICT third-party register with the correct sub-contractor classification. Change control: at least 30 days' prior written notice for any new or replaced sub-processor with a documented right to object and terminate for the affected services without penalty. Concentration risk: tenants may pin their deployment to a specific cloud and region to avoid stacking on top of another critical vendor, and the exit plan guarantees data portability in open formats so substitutability is credible rather than theoretical.

Question 28

What are your incident notification SLAs and how do you coordinate with our CSIRT?

Accepted Answer

Security-incident classification follows a documented severity matrix aligned to NIST SP 800-61 and the DORA RTS on major-incident reporting. Notification SLAs to the affected tenant's designated security contact: Severity 1 (confirmed material incident affecting confidentiality, integrity or availability of tenant data) — within 24 hours of confirmation, with the initial notification and a follow-up within 72 hours containing a preliminary root cause and containment status; Severity 2 — within 72 hours; Severity 3 — in the next regular service report. A named FVI Security Duty Officer is reachable 24×7 via a dedicated secure channel; a joint CSIRT-to-CSIRT bridge can be established on request. Post-incident: a full RCA with corrective and preventive actions is delivered within 20 business days, and tenants receive the underlying hash-chained audit slice to independently verify scope and timing.

Question 29

How do you manage supply-chain security, SBOM and vulnerability disclosure?

Accepted Answer

Every production build produces a signed SPDX / CycloneDX Software Bill of Materials that is retained and shareable with Tier-1 tenants on request. Dependencies are continuously scanned (SCA) with remediation SLAs of 24 hours for Critical, 7 days for High and 30 days for Medium, tracked in a public-to-tenant vulnerability register. Container and infrastructure-as-code images are scanned and signed (Sigstore / cosign) with provenance attestations aligned to SLSA level 3 as a target. A responsible disclosure programme with a published security.txt accepts researcher reports; validated findings are triaged within 2 business days. Third-party components critical to the reconciliation path (matching, extraction, LLM SDKs) are tracked with named owners so tenants can evidence the full supply chain in their own vendor register.

Question 30

What insurance cover and financial viability information can FVI share?

Accepted Answer

Under NDA, FVI provides the standard financial-viability packet Tier-1 vendor risk teams expect: Certificates of Insurance evidencing Cyber Liability, Professional Indemnity / Technology Errors & Omissions, Commercial General Liability and Employer's Liability at limits appropriate for regulated financial-services engagements (specific limits confirmed in the Order Form and scaled to contract value where required); latest audited or reviewed financials, funding position and 12-month cash-runway statement; corporate structure, beneficial ownership and sanctions/PEP screening; a going-concern statement from the CFO; and — for critical engagements — source-code and configuration escrow with an independent agent (NCC Group or equivalent), with release conditions tied to insolvency events and unremedied material breaches so continuity of your reconciliation is protected.

Question 31

How are data-subject rights, retention and deletion handled?

Accepted Answer

FVI acts as a data processor on tenant instructions. Data-subject requests (access, rectification, erasure, restriction, portability under GDPR / UK GDPR / CCPA / equivalents) are supported through tenant-side self-service APIs and admin tooling — no ticket to FVI required for the common cases. Where FVI assistance is needed, requests are actioned within 30 days by default and 72 hours for urgent regulatory requests. Retention: per-tenant policies for operational data (default 7 years for audit / books-and-records aligned to SEC 17a-4 and BCBS 239, configurable), with automated purging on schedule; PII fields can be shortened or tokenised earlier under a per-field policy. Deletion on contract exit: within 30 days of exit-plan completion, all tenant data (primary, replicas, backups, log archives) is cryptographically erased via KMS key destruction where applicable and physically purged from all systems, with a signed Certificate of Destruction issued to the tenant and to the tenant's regulator on request.

Question 32

How do we contact FVI's Security, Privacy and Vendor-Risk teams during procurement and in-life?

Accepted Answer

Named, monitored channels — no shared inbox. During procurement: a Deal Desk lead coordinates all workstreams; a Security Solutions Engineer owns the InfoSec deep-dive; the DPO owns DPA / SCC / privacy questions; a Legal Counsel owns MSA redlines. In-life: a named Customer Security Manager and Vendor Relationship Manager are assigned per Tier-1 tenant, with quarterly business reviews covering security posture, incident summary, sub-processor changes, roadmap and control-effectiveness metrics. 24×7 contacts: Security Duty Officer for confirmed incidents, Support Duty Manager for P1 service events, and an executive-escalation ladder up to the CEO / CTO with target response times documented in the Order Form. All contact points, PGP keys and escalation paths are published in the Security Overview Pack and refreshed on personnel change.

Question 33

What is the data lifecycle — backup frequency, versioning, and the exact deletion process when a tenant requests removal?

Accepted Answer

FVI operates a documented, contractually-committed data lifecycle so tenants can evidence recoverability and destruction end-to-end. Backups: the primary Postgres store is protected by continuous WAL streaming (point-in-time recovery to any second within the retention window), automated full snapshots every 24 hours, and cross-AZ replication in real time; cross-region DR snapshots are taken every 6 hours where enabled. Object storage (evidence artefacts, extracted documents, export bundles) uses versioned buckets with change-data replication to a secondary region on request. Backup retention: 35 days rolling for PITR, 90 days for daily snapshots, 12 months for monthly archives (aligned to BCBS 239 / SEC 17a-4 expectations), configurable per tenant. All backups inherit the primary store's AES-256 encryption, KMS keys (including BYOK/CMEK), and regional pinning — a backup never crosses a region boundary the primary is not permitted to cross. Restore is tested quarterly with tenant-visible evidence in the audit chain and target RTO ≤ 1 hour / RPO ≤ 5 minutes. Versioning: reconciliation definitions, mapping rules, tolerance policies, agent policies and RBAC configuration are versioned with immutable revision history, diff view, who/when/why metadata and one-click rollback; object-storage artefacts retain prior versions for the tenant's configured window (default 90 days). Every write to a versioned surface is captured in the hash-chained audit so any change — including reverts — is verifiable. Deletion on tenant request: (1) the tenant admin (or an authorised counterparty via signed request) opens a deletion ticket scoped to a dataset, a reconciliation, a legal entity, or the entire tenant; (2) FVI confirms scope, legal hold status and any regulatory retention obligations that would block deletion, then issues a written deletion plan with target completion date; (3) on execution, the primary records are soft-deleted and moved out of query paths immediately, then hard-deleted from live storage on a 7-day cool-off (extendable at tenant request) that protects against accidental or malicious requests; (4) all replicas, secondary indexes, caches, search shards and queue payloads are purged in the same window; (5) backups and snapshots containing the deleted data expire on their normal rolling schedule, and — for BYOK/CMEK tenants — cryptographic erasure is available immediately by destroying the tenant's KMS key, rendering all encrypted copies (including backups) permanently unreadable; (6) on full-tenant exit the same process runs across primary, replicas, backups, log archives and evidence stores within 30 days of exit-plan completion; (7) a signed Certificate of Destruction is issued to the tenant and, on request, to the tenant's regulator, listing datasets, timestamps, key IDs destroyed, backup expiry dates and the operator who executed each step — all cross-referenced to the hash-chained audit records so the deletion itself is independently verifiable.

Question 34

How are audit logs accessed and exported — UI, API, and what tenant-level authorization applies?

Accepted Answer

FVI exposes the same hash-chained audit stream through three tenant-scoped surfaces, each gated by the tenant's RBAC and RLS policies so an operator can only ever see records belonging to tenants they are entitled to. (1) In-product UI: the Audit Explorer under Governance → Audit lets authorised roles (default: Tenant Admin, Security Officer, Auditor, Compliance Reviewer) filter by actor, tenant, business unit, resource type, resource ID, action, correlation ID, IP, session, time range and hash-chain verification status, then export the current result set as CSV, NDJSON or a signed Evidence Bundle (ZIP containing NDJSON + Merkle proof + chain-head signature + manifest). Every export is itself an audited event with the exporter, filter hash, row count and destination recorded. (2) REST API: GET /v1/audit/events (cursor-paginated, filter parity with UI) and POST /v1/audit/exports (asynchronous, returns a signed download URL scoped to the caller's tenant and expiring in 15 minutes) — both require an API token minted under an RBAC role with the audit.read or audit.export permission, and both enforce the same RLS predicate (tenant_id = auth.tenant_id() AND has_role(auth.uid(), 'auditor'|'admin'|'security_officer')). Tokens are tenant-bound at issue time; a token minted for Tenant A physically cannot query Tenant B, and any attempt is itself logged as a tenant_isolation_violation. (3) Streaming: authorised tenants can subscribe their SIEM (Splunk, Sentinel, Chronicle, Elastic) via signed webhook or Kafka-compatible topic; delivery is at-least-once with per-event hash and monotonic sequence so the SIEM can independently verify no gaps. Event schema (stable, versioned): event_id (uuid), event_time (ISO-8601 UTC, µs), tenant_id, actor_type (user|service|agent|system), actor_id, actor_email, session_id, ip, user_agent, action (verb, e.g. recon.match.approve), resource_type, resource_id, resource_version, business_unit_id, correlation_id, request_id, before (JSON, redacted per DLP), after (JSON, redacted per DLP), outcome (success|denied|error), reason_code, prev_hash, hash (SHA-256 over canonical JSON of the record + prev_hash), chain_id, signature (Ed25519 over the daily chain head), retention_class and legal_hold flag. Authorization checks applied on every read/export: (a) authenticated principal with an unrevoked session or token; (b) RBAC role grants audit.read (view) or audit.export (bundle/stream); (c) RLS predicate binds the query to the caller's tenant_id and, where configured, business_unit; (d) ABAC predicates can further restrict by resource_type, sensitivity label or data-residency region; (e) step-up MFA is required for exports over a configurable row threshold (default 10,000) and for any query touching PII columns; (f) four-eyes approval is required to export another tenant's records under a Break-Glass grant, and the Break-Glass session itself appears in that tenant's audit within seconds. Retention defaults to 7 years (configurable up to 10) on WORM storage with legal-hold override; deleted tenant data is purged per the data-lifecycle FAQ but audit records of the deletion itself are preserved for the full retention window as regulator-facing evidence.

Control	SaaS · multi-tenant	VPC / Private Cloud
Tenant isolation	Row-Level Security + per-tenant storage prefixes	Dedicated database, storage, queues and workers
Encryption at rest	AES-256, FVI-managed KMS envelope keys	AES-256, customer-managed KMS in your account
BYOK / CMEK	Optional — pinned to your cloud KMS	Default — every key lives in your KMS
Data residency	Pinned to a home region at provisioning	Any region(s) your cloud account permits
Network boundary	Shared, TLS 1.2+, IP allow-list, private egress	Private VPC, PrivateLink / Peering, no public egress
Identity & SSO	SAML / OIDC / SCIM against your IdP	Same — plus optional bind to internal-only IdP
Audit chain	SHA-256 hash-chained, tenant-isolated exports	Same chain, streamed to your SIEM in-account
Upgrade cadence	Continuous — controlled release rings	Scheduled windows agreed with your CAB
Break-glass access	Time-boxed, four-eyes, recorded	Same — plus optional customer approval gate
DR / RPO / RTO targets	Multi-AZ, RPO ≤ 15m, RTO ≤ 2h target	Configurable to your DR standard
Shared responsibility	FVI runs the platform end-to-end	FVI runs the platform; you own the cloud account
Commercials	Subscription, usage-metered	Subscription + platform-ops fee for in-account run

Clear reconciliation breaks 10× fasterwith agentic AI.

Estimate your annual reconciliation savings

The Enterprise Reconciliation Buyer's Guide

Scrub the pipeline. Watch reconciliation resolve in real time.

Five stages. One continuous intelligence.

One platform. Every reconciliation.

Matching engine

Agentic AI triage

BYO-LLM sovereignty

Rekon Orchestrator

Data Prep pipelines

Break Manager

Reports & SLA

Live collaboration

Smart notifications

Multi-tenant RBAC

Hash-chained audit

Zero-trust security

A team of specialised agents — under human control.

Enterprise-grade by default.

Three disciplines. One source of truth.

Aperture

Precision

Intelligence

Built for scale. Designed for clarity.

Veritas Core

Ingestion Hub

Near-real-time rekon

Dev → UAT → Prod

Engineered to be an order of magnitude better. Not incrementally better.

From sentence to live rec — in hours, not weeks

95–99% auto-match target — before a human sees the queue

Zero vendor lock-in — including your LLM

Built for organizations operating in highly regulated environments.

One platform. Many banks. Zero shared blast radius.

Every row, file, and job is bound to a tenant.

Least-privilege by default, RBAC + ABAC, four-eyes on sensitive actions.

Encrypted in transit, at rest, and — where required — under your own keys.

Hash-chained, tenant-isolated, tamper-evident — exportable on demand.

Engineered to satisfy Operational Risk, InfoSec, and Sourcing — not to work around them.

Second-line assurance pack

Deployment topology & data residency

Legal, commercial & supplier onboarding

Governance, records & auditability

Integration & change governance

Operational resilience & TPRM

From sponsor alignment to controlled production in 6–9 weeks.

Everything you need to know

A Tier-1 pathway from contract to go-live

Run FVI the way your risk model requires

Tell us where you are — we'll bring a plan.Dr. Falcon answers product questions and can book time with our team. Or leave your details and we'll reach out within one business day.

Tell us where you are.

Stop reconciling. Start verifying.

How to read the numbers on this page

Clear reconciliation breaks 10× faster
with agentic AI.

Tell us where you are — we'll bring a plan.
Dr. Falcon answers product questions and can book time with our team. Or leave your details and we'll reach out within one business day.