Skip to content
Broker news. Connected. RSS feed
FXBROKERS

A clearer view of the broker industry.

Gresham separates AI-assisted configuration from the reconciliation engine

Analysis of the 1 October launch: Control Studio turns business knowledge into reconciliation rules, making configuration review central to adoption.

Gresham’s Control Studio, launched on 1 October, brings generative AI into the construction of reconciliation controls. The company says business users can describe requirements and create configurations for its deterministic rules engines. It aims to reduce the work needed to translate knowledge of accounts and transactions into controls.[1]

That architecture gives the AI and the reconciliation engine different jobs. The AI helps express the intended matching and validation logic; the engine applies the configured rules. Gresham describes suggestions for matching keys, validation checks and enrichment expressions, together with testing against sample data before a control goes live.[1]

The wider Control platform supports reconciliation across three or more sources. Gresham identifies uses in prime brokerage, fund accounting and collateral operations, where checking two records may leave another system’s disagreement unresolved. It also describes tracking breaks by age, value or counterparty and assigning work to the responsible team.[2]

Consider a broker comparing its trade record, clearing statement and cash ledger. Matching two is insufficient if the third records a different amount or settlement status. A control spanning all three can expose the inconsistency, although the firm must decide which differences matter and which tolerances it will accept.[2]

The implementation question consequently moves towards configuration review. A sample-data test should contain both records expected to match and deliberate exceptions; otherwise a rule that suppresses a real problem could appear successful.[1]

For operations teams, easier configuration is valuable when it shortens the route from a newly identified discrepancy to a controlled, repeatable check. The continuing work is to maintain those rules as source files, transaction types and responsibilities change. Faster setup has to feed a reliable process for investigating the differences that remain.[2]

Follow the evidence

Sources & further reading

Report an error or submit an update