Equalsfive. Start hereDeliverablesPositionEvidenceAudienceCommercialComparisonThe corpusThe brief
The argument · and what's at stake

The failure mode of this job is silence

Why the accountability argument is available, what their buyer actually carries, and the ten ways it goes wrong — each drawn from Fund Recs' own product library, turned from feature into failure.

⚠ Pre-human

No human has read a word of this. Machine output, published raw so the reasoning can be attacked before anyone polishes it. Judge the chain, not the copy — if a numbered step is wrong, say which.

What's contested and what isn't

The category has spent eighteen months arguing about agentic AI in the back office. Four of six competitors now make the same supervised-autonomy argument — Duco on a PostTrade360 stand, AutoRek in the ARIA factsheet, Xceptor in a post four days old, SmartStream via Celent.

So the argument is contested. The audience is not. Every one of them is talking to the trading floor and the middle office, where an error costs time and money. Nobody is talking to the person who is personally answerable — and that person is Fund Recs' entire customer base.

Duco's buyer is a bank's head of middle office. He is not named on a licence. They cannot make this argument credibly even if they want to, which is what makes it defensible rather than merely unclaimed.

What their buyer actually carries

A Responsible Entity holds the licence. A depositary is strictly liable for loss of assets under AIFMD/UCITS. A ManCo director is named. The boutique picks the stocks; the person buying this software is the one a regulator writes to.

If the position and cash records are wrong, the NAV is wrong. If the NAV is wrong, someone bought or sold at the wrong price with real money, and somebody has to make them whole. From Fund Recs' own blog: "Investors may transact at the wrong price · Fund boards may raise questions about control adequacy · Auditors or regulators may flag the absence of documented oversight."

The cost, published by them in 2015 and never used since: "In 2009 GlobeOp paid out US$43.5 million to Regents Park Capital Management, a UK hedge fund, as a result of claims that GlobeOp miscalculated the NAV."

And a real example from their 2025 blog: a 3.8% NAV swing on a multi-currency fund, traced to stale FX rates that weren't updated because an overnight job failed. A cron job didn't run and the fund was mispriced by 3.8%.

Ten ways to lose the licence

Fund Recs sells twenty-odd things called "Controls". Each exists because something specific goes wrong and somebody carries it. Here they are as failures rather than features. Every one is described on their own site — the only change is the direction.

#The failureWhat it costs · and what catches it
1The NAV goes out wrong and people trade on itUS$43.5m in the documented case. Everything below, uncaught — the NAV is downstream of all of it. NAV Oversight: 14 checks before the price goes out.
2A price didn't move and nobody noticedThe 3.8% swing. "Stale prices create valuation risk, distort NAVs, and can trigger regulatory issues." Stale Price Control.
3Something priced at zeroRedeemers get less than they're owed; buyers get a bargain at everyone else's expense. Both compensable. Zero Price Control.
4The FX rate was wrongOne bad rate misvalues every non-base-currency holding at once — and looks like a market move. FX Rates Control.
5A real break written off as timingTheir own before-state: free-text comments "created a lot of inconsistency due to differing users completing the reconciliations day on day." Pending Trades Control matches breaks to the trades that explain them.
6Trades never reported at all"The absence of a trade report isn't just an oversight — it's a regulatory failure… the absence passes silently." No alert fires. Reconcile what you should have reported against what the repository holds.
7Reports went in late, repeatedly"Repeated lateness suggests deeper control weaknesses." The regulator stops looking at the reports and starts looking at you. Automated overnight ingestion; timeliness as a tracked metric.
8A payment went to a fraudsterSS&C sued over US$5.9m — and Meaney's damning line: "processed a further 5 times across 21 days without being caught." Investor Static Data Reconciliation.
9Fees charged outside the prospectusOvercharging investors is a compensation event and a disclosure breach in one, found three years later across every fund at once. Expenses vs Prospectus.
10It was all fine — and you can't prove itASIC: "some trustees not carrying out any checks in a month despite a 75% adverse finding rate… comfortable with limited, almost entirely manual indicators." Absence of documentation is the finding. Every control produces audit-ready output as a by-product of running.

The pattern

Almost nothing on that list announces itself. A stale price looks like a stable price. An unreported trade generates no alert. A wrong FX rate looks like a market move. A real break looks like a timing difference. A missing control looks like a clean month.

The failure mode of this job is silence.

Two consequences. It explains why the buyer feels no urgency — nothing has gone wrong yet, because nothing that goes wrong here makes a noise. And it defines the product honestly: Fund Recs is not selling speed, it is selling the ability to notice, and the ability to prove you noticed.

Not a comprehension gap — a placement gap. The buyers understand the risk perfectly. But every control page opens with what it automates and assumes the reader supplies the fear. The consequence material exists, is theirs, and is filed where nobody buying will find it. That is cheaper to fix than a repositioning.
Built by Equalsfive from public sources only, 31 August 2026. No human has reviewed the copy on this site. Evidence base: content-corpus.jsonl — 1,205 records. Grades: measured counted or quoted · inferred derived from a small sample · domain industry practice, unverified.
Equalsfive · private

Fund Recs

Pre-human working papers. Not for distribution.