Twenty days. Recommend only. Nothing changes on your rails.
For collateral operations, securities finance and treasury desks.
A shadow run sits beside your existing process and proposes. It instructs nothing, changes no contract, and touches no venue. At the end you compare our proposal with what your desk actually did — and if it did not beat your current path, that is the finding.
- Week 0
- You choose a deployment option. Nothing runs yet.
- Days 1–5
- Adapters read your extract shapes. Disagreements between your own sheets surface here — usually the first thing worth seeing.
- Days 6–20
- A daily proposal, recommend-only, scored against what your desk actually pledged.
- Close
- A scorecard, the disagreement log, and a decision. If it did not beat your current path, we say so.
Requirement per account and the venue profile it runs under
Inventory by firm, encumbered and long box
Identifiers, lot size and increment
Permitted pairings with margin or haircut factor
Exclusions with a reason
Concentration caps as available value, with scope
These files never leave your environment — the software runs where you put it, and we receive no copy. See how that works →
| Measure | Compared against what your desk actually pledged |
|---|---|
| Cash posted | Where a cheaper eligible security was available and cash went out instead. |
| Rejects | Lines the triparty agent or counterparty refused, and whether we would have avoided them. |
| Substitutions | Movements proposed against the cap your desk set. |
| Specials preserved | Names trading wanted back that were pledged anyway. |
| Time on short accounts | Minutes spent on the accounts that actually needed work. |
The baseline is your own book over the same twenty days. We do not score against a benchmark we chose.
If it does not beat your current path, keep your current path.
The shadow runs beside your existing process for twenty business days and ends there. There is no production change to unwind, no contract to exit, and no rail to disconnect. A shadow that fails is still a useful result, and we will say so plainly.
This trial evaluates the allocation file. If your question is the documentation behind it — agreement packages, operative scope, extracted terms — see Agreement Intelligence, which is evaluated separately and with a different team.
The first call is technical before it is commercial: which book, which venue, and what your extracts look like. Terms are scoped from that, once the work is understood. Security and model-risk material is on the Trust page and can be sent ahead.