Optimization, not just calculation
Eligibility, haircuts and fees checked before an asset is pledged, not after it's rejected.
Effective optimization requires current inventory, accurate eligibility, informed decisioning and a result that can actually be executed.
- ~$30Bprocessed daily by Baton's technology
- 100+accounts in 17 currencies live for G.H. Financials
- 8 weeksto key account connectivity and visibility at G.H. Financials
Inventory visibility
A single institution can be connected to more than 70 counterparties, custodians, CCPs and settlement venues, rising into the thousands for a G-SIB, each reporting on its own schedule and in its own format.
Source: ISDA, ISLA and ISSA market surveys.
- Expected against confirmed
Balance Manager and Exposure Manager consolidate real-time balances and reconcile expected against confirmed settlement activity.
- Filtered to what the call needs
Filter by currency, legal entity and segregation to confirm funds are in the right currency, location and accessible window.
- Forward-looking and historical
Dashboards show peaks and troughs through past volatility on demand, rather than as an ongoing manual task.
Knowing you hold the currency isn't the same as knowing you hold it where and when it's usable.
Eligibility rules
Each venue applies its own eligibility criteria, haircuts and fees. Comparing them has traditionally taken manual analysis, which is slowest exactly when a margin call lands and volatility is running high.
- Venue-specific criteria
Eligibility is evaluated automatically against each venue's criteria, using real-time eligibility information pulled from the venue itself.
- Compared across every venue
Haircuts, fees and eligibility are compared automatically across every connected CCP, protecting margins through volatility.
- Cash and non-cash composition
Cash is clearly separated from eligible non-cash assets, so government bonds, high-quality corporate debt and equities stay in view.
Regulatory and CCP reporting becomes a by-product of the same real-time, normalised data, rather than a separate aggregation exercise.
Decision and executability
Mobilising non-cash collateral carries real operational drag under intraday cut-offs: slower settlement and higher fail risk. So cash becomes the default, even though it leaves higher-yielding assets sitting unused.
- The decision automated
Reclaim idle collateral by automating the allocation decision itself, not just the visibility around it.
- Governed AI access
With Model Context Protocol support, institutions apply their own AI models and agents to collateral, balance and eligibility data and instruct actions within governed constraints.
- Logged and overseen
Every model-derived decision is logged, with human oversight retained throughout.
“If you don't have a real-time view, then by definition you've got a suboptimal liquidity scenario, you have cash sitting in places it doesn't necessarily need to be.”Steve Plestis, Head of Sales, G.H. Financials
The cost of disconnected optimization
70% of firms surveyed report daily settlement matching and delivery issues, and 35% post more than half their collateral overnight to ensure timely delivery.
Sector-wide research: Nasdaq and The Value Exchange, Making the Case for Tokenised Collateral.
- Inventory
- Held, but not usable
Funds exist in aggregate but sit in the wrong currency, location or accessible window.
- Eligibility
- Comparisons made by hand
Without automated comparison, higher-quality eligible assets stay locked in custodian accounts.
- Decision
- Rejected after the event
A rejected or sub-optimal placement is discovered only after the instruction has gone out.
- Execution
Go deeper
Examine the operating model that connects inventory, eligibility, optimization and execution.
A collateral decision is not a completed movement
Speak with a collateral specialist
Tell us what you'd like to discuss and we'll arrange a time.