Frequently asked questions.

No. SmartSettle AI doesn't create settlement instructions; ALERT is the system of choice, and managers, custodians, and middle-office providers can input them far more accurately than we could. What we do is detect when instructions are wrong (e.g. a PSET/PSAF mismatch) by subscribing to custodian settlement-instruction packs and, in progress, ingesting ALERT instructions to identify which side is incorrect. This matters because sometimes the manager's PSET matches the custodian's instruction and it's the sell-side that's wrong.

Once the accurate instructions are known, one side needs to correct and re-enrich the trade; the SSAI Root Cause and Recommended Action fields flag this. Humans or agentic AI inside the customer's firm source the correct instructions (e.g. from ALERT) and re-enrich, and the orchestration layer then sees the corrected trade come through, at which point it typically settles.

Yes, we're building this. We're developing a way to trap failing “box moves” and “box compressions” for inter-PB movements directed by hedge funds: consuming a transaction report, mapping the transactions, and identifying the matrix of movements that mismatch (i.e. moves that did or didn't happen as planned).

Yes. We plan to subscribe to cash-break files from managers' IBOR reconciliations and apply root causes to non-receipts and non-delivers of cash collateral, honing in on the purpose indicator, wire-instruction accuracy, and ideally agreed calls married to the breaks. The goal is to supply enough intelligence that automated emails, users, and AI agents can resolve these across the Roundtable, escalating only aged or risky items. We aim to do the same for high-grade securities collateral (US Treasuries, Agencies, many sovereigns), including collateral maturing “out of the box” and needing to be returned as cash.

Building on our prime-broker box-move work, we intend to work with sub-custodians and global custodians to pull detailed inventory and transaction files and identify Free-of-Payment movements of shares between depositories and locations, especially in Europe. Consuming those detailed feeds for split positions and box moves is core to our mission of acting as the utility that resolves failed trades and movements between firms.

Not yet. We don't have this functionality or the sub-custodian feeds today, but non-ticketed events (especially FOP) are prone to fail, so their identification and resolution is on the roadmap. We're starting with prime-broker box moves and compressions by consuming activity reports to match the directed “grid” of moves, and by subscribing to IBOR breaks we intend to apply our root-cause layer and agentic orchestration to resolve cash-break fails alongside the DVP fails we handle today.

No. We're not replacing cash transaction or position reconciliation. For in-kind and FOP movements, though, we believe we can subscribe to activity reports, corporate-action feeds, and directed activity from various sources and, at the transaction level, identify failures and apply clear root-cause analysis so humans or agents can resolve them through the orchestration layer.

We haven't needed minimum-lot-size validation yet (that root cause hasn't arisen with current clients), but it would be straightforward to code by asset class and clearing venue. We don't instruct trades; our job is to flag the reason for failure as close to trade time as possible. If a trade were too small to clear (e.g. certain sub-$25,000-par mortgage issuances, or an AustraClear 500k minimum), SSAI would raise the root cause to the manager and broker, recommend breaking the trade with any error proceeds going to the client, and suggest a gating constraint in upstream pre-trade compliance to prevent recurrence.

We're building real-time resolution for PSET/PSAF mismatches with a decision waterfall that mirrors current market practice as agents and humans negotiate around the Roundtable: the seller's place of safekeeping takes first precedence on where to settle; if the buyer isn't a member there, the seller moves the securities to the buyer's location; and if both are members of both, the manager's client-custodian gets preference, even on a buy. Parties can also reinstruct to a shared third venue such as Euroclear. Because pools of agents resolve trades in real time, once the location is set the trade is reinstructed by whichever side flexed, then verified as settled, with agents taking somewhat free paths within the workflow each root cause requires.

Absolutely. We intend to consume mismatches from ClearPar for new issuances, secondary trades, drawdowns, and revolver trades, and to subscribe to IBOR cash breaks to flag failed income receipts. Bank-loan income often fails because the agent bank distributes coupons at the “lender” level rather than the “sleeve” level, leaving custodians and outsourcers to clean up money movements; our aim is to apply root causes, let agentic AI split and move bulk income across sleeves with humans in the loop, and anonymize root-cause themes to push upstream fixes. Tax withholdings in non-taxable accounts are another common cause. By engaging brokers, syndicators, and agent banks on the Roundtable, agents can help organize tax forms (e.g. W-8BENs) to prevent recurring breaks.

Possibly, though the system works a little differently than that implies: SmartSettle AI isn't a dashboard of all trades and positions. It's a dashboard of failing or not-yet-settled trades, with root causes so users can fix them. As feeds arrive through the day we marry the data, retaining the initial place of safekeeping and appending new data like custodian fail reason codes. So a one-time initialization generally isn't needed; we pick up PSAF on new unsettled and failing trades as they come. And if a custodian can't supply inventory but a middle-office provider can, we can take that provider's location feed and append it to the fails data easily.