average monthly penalties paid by T2S participants for late matching and settlement failures through 2024 - ranging from €42m in December to €108m in September
Firebrand Research and DTCC
of transactions did not settle as scheduled by volume, and around 2.5% by value
Annual public disclosure of settlement fails, as reported by SIX
of 2024 settlement failures across surveyed firms were counterparty shorts; a further 21% were data issues, including incorrect or stale SSIs Firebrand Research and DTCC
Firebrand Research and DTCC
UK, EU and Switzerland T+1 go-live, with ESMA's October 2025 settlement discipline amendments shifting emphasis from penalising fails to preventing them through automation
ESMA Final Report on amendments to the RTS on Settlement Discipline, 13 October 2025
Over nine in ten reported failures trace to two causes – a counterparty short, or bad data. Both leave a signature in your own history before they become a fail. ESMA’s October 2025 amendments to the settlement discipline RTS moved the emphasis from penalising fails to preventing them, prioritising automation and standardisation. That is exactly the shift Phase 2 is built for.
Every open break carries a settlement risk weight. The trades most likely to fail by cutoff sit at the top of the operator's screen.
Incomplete data entries and unmet conditions are surfaced before they cause a failure.
For a specific failed transaction, TradeIQ reads your replica database plus system and interface logs, identifies issue behind it, and returns recommended remediation - not a generic explanation.
As operators resolve exceptions, TradeIQ learns the pattern. Recurring failure types move from being investigated to being pre-empted.
Settlements and recon, risk and margin, finance, and senior management each get an actionable view.
Your regulatory templates are mapped to data already inside NOVA, so recurring extracts populate for review instead of being rebuilt in spreadsheets by several people.
Escalation reserved for judgement, so L2 and L3 handle what genuinely needs them.
Configurable recon engine across trades, payments, cash, static data & corporate actions - beta targeted H2 2026.
Entitlements posted straight to the holding - no interface, no reconciliation.
Cross-desk visibility in one queue across multi-desk settlement flows.
Specific cause and remediation per failed transaction, drawn from your replica and logs.
Claims and excess entitlements identified and controlled automatically.
Live operations board ranking open breaks by risk of failing at cutoff.
Full resolution history retained as an auditable record of who decided what.
Full reverse and redo with a complete audit trail on every event.
Escalation reserved for judgement, so L2 and L3 handle what genuinely needs them.
Cross-desk visibility in one queue across multi-desk settlement flows.
Live operations board ranking open breaks by risk of failing at cutoff.
Configurable recon engine across trades, payments, cash, static data & corporate actions - beta targeted H2 2026.
Specific cause and remediation per failed transaction, drawn from your replica and logs.
Full resolution history retained as an auditable record of who decided what.
Entitlements posted straight to the holding - no interface, no reconciliation.
Claims and excess entitlements identified and controlled automatically.
Full reverse and redo with a complete audit trail on every event.
No market NOVA CMX serves is waiting for an AI act. Existing frameworks apply now, and the answer expected is human oversight, explainability and an audit trail. That is why the final call always sits with your operator.
MAS consulted on Guidelines on AI Risk Management (13 Nov 2025), covering decision support as well as decision-making, with a proposed 12-month transition after issuance.
The SFC's generative AI circular puts senior management oversight first; SFC and HKMA both issued AI-related circulars on 2 June 2026.
ASIC's REP 798 found governance lagging adoption across 624 use cases; APRA's 30 April 2026 letter named governance immaturity and inadequate assurance.
BNM's revised RMiT took effect 28 Nov 2025 and its AI discussion paper was published August 2025; the SC's technology risk guidelines already carry an AI appendix.
The FCA plans no AI-specific rules, applying the Consumer Duty, SM&CR and its governance expectations, with AI Live Testing running through 2026.
CBUAE, SCA, DFSA and ADGM's FSRA jointly issued Guidelines for Financial Institutions Adopting Enabling Technologies, covering AI and analytics.
01
Breaks are worked in the order they will cost you, and the predictable ones stop becoming fails.
02
Risk-weighted queues and precondition flags replace the overnight buffer that no longer exists.
03
Escalations that only needed a diagnosis are resolved before they reach your experts.
04
Isolated replica, controlled query layer, cited output, human approval - mapped to what your supervisor has published.
05
TradeIQ sits with NOVA RPA and NOVA Citadel inside NOVA Intelligence Hub - automation, intelligence, resilience on one operating model.