Errors caught before they leave the desk. Not in next-day reconciliation.
Neith evaluates trading intent before execution — deterministic controls, a clear decision, and a complete record. It starts in shadow mode and never touches the execution path.
[ SHADOW MODE · 5 PLACES THIS QUARTER ]
WHY NEITH
Trading has execution infrastructure. It has risk systems. It has surveillance. What is missing is a control layer that understands trading intent before execution — regardless of where that intent originates. Electronic orders already pass through structured controls. Voice, chat and RFQ trading often does not.
ONE CONTROL LAYER · EVERY TRADING WORKFLOW
NEITH CONTROL LAYER
Neith connects through browser or API, evaluates the trade against controls you configure, records the decision, and passes it on. Today it operates in advisory / shadow mode and does not interfere with execution.
WHY NOW
Trading decisions increasingly originate from software as well as people. An autonomous agent trades with no lunch break and no second pair of eyes — one bad instruction repeats at machine speed. Neith applies the same independent, deterministic check whether a person or a model sent the trade, and records what happened.
WHY IT HAS TO BE INDEPENDENT
NOT THE MODEL, NOT THE OMS
A regulator won't accept a black box checking itself. Neith is a separate, third-party control.
NO AI IN THE VERDICT
Every decision traces to a named rule — explainable and repeatable, the opposite of the agent it supervises.
BUILT FOR REGULATED TRADING
Market-specific rules and a tamper-evident audit trail, by someone who traded the desk.
Every generic guardrail wraps the model. A bank needs the layer that doesn't.
START SMALL · SCALE ACROSS THE ESTATE
START SMALL
Start with one desk, one channel or one workflow, through the browser or API. Run Neith non-invasively and prove the control model without disrupting execution.
GO WIDE
Extend the same control layer across desks, trading channels, systems and increasingly automated trading workflows — one policy model, one evidence layer.
THE GAP
The pre-trade checks firms run are hard-coded thresholds built for typed electronic orders under MiFID II RTS 6 — set once, rarely retuned. They miss errors, and the 41% of FX agreed by voice, chat and RFQ barely passes through them at all.[1]
For twenty years those checks were static rules bolted to trades already typed into a system. They could not read a voice line, a chat or an RFQ — so on that flow they ran blind, and errors surfaced a day later, if at all. That is what changed — and it is why this is buildable now.
“In certain cases, ownership of pre-trade and post-trade controls was poorly defined and not documented… compliance staff had a lack of oversight of pre-trade and post-trade controls… a weak understanding of the controls and how they functioned.”
— FCA, Multi-firm review of algorithmic trading controls, 21 August 2025[3]All ten firms in that review had adequate pre-trade controls. What the regulator could not find was anyone who owned them, documented them, or could explain how they worked.
The controls exist. Owning them — and catching what they miss — is the gap.
WHY THIS EXISTS
I spent years broking FX. In all that time, not one bank ever rejected a trade I sent them — they took it, booked it, confirmed it every time. The call came the next morning, after reconciliation: wrong rate, wrong value date, wrong counterparty. By then the market had moved, and the only record of what was agreed was a chat window and somebody's memory.
Acceptance was never validation. It only ever meant the error had not been found yet.
Nothing was built to catch it in the moment; the checks ran the next day, long after the cheapest minute to fix it had passed. Neith is that check, moved to the point of the trade — and because a control nobody can evidence may as well not have run, it records what it found, who saw it, and what they decided.
REFERENCES
Regulatory references are provided for context and do not constitute legal or compliance advice. Neith supports these obligations; it does not discharge them.
One desk, six weeks, shadow mode — no execution risk. Five places this quarter.
Start a shadow-mode pilot →