Client authority
PMS, O/EMS, positions, entitlements, limits, compliance policy, workflow history, orders, fills and books of record remain client-controlled.
Platform
Execution Intelligence Infrastructure
Avarrai brings together market context, execution workflow, institutional knowledge and governed AI to support smarter trading before, during and after execution.
Built for institutional execution
Trading decisions increasingly depend on fragmented market data, workflow context, liquidity signals, counterparties, policy and institutional knowledge.
Avarrai is being built as the intelligence layer that connects that context, evaluates execution choices and returns explainable, governed recommendations into existing workflows.
What the platform is
Avarrai is designed as an execution intelligence layer for institutional trading.
It is not intended to replace O/EMS, venues or existing trading workflows. It sits across the execution stack, bringing together context, intelligence and governance to support better decisions while preserving existing systems of record and control.
Avarrai helps execution teams answer the questions that shape execution outcomes: whether to trade, when to act, how to execute, where to execute, with whom to engage, at what price, and why that decision is supported.
This turns fragmented execution context into structured, explainable and governed outputs that can flow back into trader interfaces, O/EMS workflows, APIs, audit records and analytics.
What it helps with
The Avarrai platform is being built around six core capabilities:
Market & liquidity intelligence
Interpret market conditions, liquidity signals and execution context to understand what is actionable.
Execution decision intelligence
Evaluate timing, price, counterparties, venues, protocols and execution strategies.
Execution orchestration
Coordinate routes, workflow steps, escalation paths, fallback behaviour and decision hand-offs across execution channels and systems.
Governed recommendations
Turn market and workflow context into explainable recommendations while keeping accountability with the trading desk.
Execution evidence
Capture the context, reasoning and outcomes behind execution decisions to strengthen oversight and auditability.
Execution learning
Connect recommendations, overrides, workflow actions and outcomes so execution intelligence can improve through use.
How Avarrai fits into the execution stack
Avarrai is not designed to replace the systems institutions already rely on.
It connects to the client’s execution environment, combines client-owned workflow context with market, venue, dealer and reference data, and returns structured execution intelligence into the systems and interfaces where decisions are made.
The principle is simple – integrate where the client is authoritative. Connect directly where the market is authoritative.
Market authority
Venues, dealer APIs, RFQs, streaming prices, IOIs, axes, runs, inventory, consolidated tapes, APAs, TRACE, curves and reference data provide the market context that informs execution intelligence.
Avarrai intelligence
Avarrai normalises fragmented context into canonical views, evaluates execution choices and produces structured execution plans, reason codes, confidence and evidence.
Avarrai outputs can be consumed through trader interfaces, O/EMS workflows, APIs, audit records and analytics.
Control plane
The control plane governs permissions, policy checks, workflow progression, model boundaries, fallback behaviour, audit and replay. Models can support recommendations, ranking and explanation. Deterministic controls govern what the platform is allowed to do.
Models may change. The control boundary does not.
Applied to fixed income
Avarrai's execution intelligence infrastructure is being applied first to fixed income.
Fixed income combines fragmented liquidity and multiple execution protocols with an abundance of market, pricing and workflow signals. Axes, IOIs, runs, consolidated tape data, executions, evaluated and composite pricing, alongside other market signals, all contribute context, but often remain scattered across systems and workflows.
Avarrai brings that context together to support more informed execution decisions.
The same platform architecture is designed to extend across markets as execution workflows evolve.
Multiple execution workflows. One intelligence layer.
Designed around human judgement
Avarrai is designed to strengthen human judgement, not replace accountability.
augment the trader, strengthen the workflow, preserve control.
Assist, with accountability
AI can interpret context, evaluate options and make recommendations while keeping accountability with the trading desk.
Explain, don't obscure
Recommendations should be supported by understandable reasoning, evidence and the context from which they are derived.
Learn, without losing control
Execution outcomes and institutional knowledge improve future intelligence within defined governance and control frameworks.
Designed for institutional deployment
Avarrai is SaaS-first, with deployment models designed for different institutional, operational and sovereignty requirements.
The platform supports multi-tenant SaaS, dedicated single-tenant and sovereign SaaS configurations, allowing institutions to choose the appropriate level of isolation, control and operational resilience.
Avarrai is designed to integrate with existing O/EMS environments and institutional infrastructure. It is API-first and integration-aware, adding intelligence without forcing unnecessary replacement programmes.
Interoperability, data isolation, security, auditability, operational resilience, portability and sovereignty are design requirements, not afterthoughts.
Standard Enterprise SaaS | Dedicated Enterprise SaaS | Sovereign SaaS
Built for institutional market participants
Avarrai is being designed for:
- buy-side trading and execution teams
- sell-side trading and liquidity teams
- trading technology and architecture teams
- market-data and workflow partners
- venues and execution platforms
Interested in shaping the platform?
Avarrai is engaging with market participants, trading teams, technology partners and institutions interested in the next generation of execution intelligence.
We are particularly interested in conversations around design partnership, platform integration, fixed-income execution workflows, governed AI, market-data strategy and early access.