Duco, the data automation company, has launched a suite of next-generation AI-powered automation tools, aimed at helping financial institutions manage growing data volumes, accelerate regulatory compliance, and prepare for the transition to T+1 settlement in the UK and Europe. The launch includes three core innovations: an Agentic Rule Builder, AI-Native Data Prep, and T+0 Assurance Controls.
The move comes as firms grapple with the operational shifts required by compressed settlement cycles, which render many legacy processes and manual workflows untenable. With the North American transition to T+1 providing a stark lesson in the need for enhanced automation, Duco’s new toolset is designed to address the entire data lifecycle, from ingestion and normalisation to reconciliation and exception management.A key challenge for firms is managing disparate data sources, which often use different formats and terminology. Before data can be reconciled, it must be standardised – a process that is traditionally complex and time-consuming. Duco’s Enhanced AI-Native Data Prep addresses this head-on, explains Steve Walsh, Managing Director, Reconciliation at Duco, to TradingTech Insight.
“Clients often find it very difficult to bring disparate data together. For example, in a custody reconciliation, I might have data from my fund administrator and my internal books and records. It could be in various file formats with different attributes. There could be nuances and differences in terminology, such as ‘nominal’ versus ‘notional’ or ‘buyer’ versus ‘payer’. Before attempting to reconcile the data, I want to normalise it and create a standard way for it to be represented. The AI-Native Data Prep allows you to bring everything together to create standardisation, reduce complexity, and be in a position to scale and trust your data before it enters the reconciliation stage. You can normalise all those different terms – buy/sell, pay/receive, nominal/notional, BPS versus percentage – before you reconcile them. This ensures you have real trust that you are bringing the right data elements together for the reconciliation.”
Building on its existing Natural Rule Language (NRL), which allows non-coders to create rules in plain English, Duco’s new Agentic Rule Builder introduces a prompt-based interface. This enables business users to describe a desired outcome as they would to a colleague, with the tool’s agents then generating the formal rule for human validation.
“The Agentic Rule Builder enables you to use simple prompt language,” states Walsh. “For example, you could input, ‘Write a rule that flags all transactions over $500k as high value, adds a comment that says ‘needs 4-eye check’, and assigns them to the high value team.’ The Agentic Rule Builder will then propose the NRL to cover your requirements. So, you have entered a request in plain English, and it creates the NRL. The rule is rendered in the UI, and the operator or subject matter expert can review and accept it, ensuring a human is in the loop at all times.”
This human-in-the-loop approach, combined with the clarity of plain English rules, is crucial for maintaining transparency and satisfying audit requirements – a key concern for regulatory compliance.
“A core benefit of DUCO is auditability,” Walsh notes. “At any given time, if I’m creating a reconciliation, I can extract a PDF that provides a blow-by-blow account of exactly what has been done: who accepted it, when they accepted it, and how they created it. One of the key business benefits we hear from prospective clients is that because it’s in plain English, internal and external auditors can read it.”
Ultimately, these advancements in data preparation and rule creation are critical enablers for the operational model required by T+1. The 80% reduction in post-trade processing time means that intra-day reconciliation and exception management must become the norm, demanding unprecedented levels of automation and data integrity on the day of the trade (T+0).
As Walsh concludes, “By utilising data ingestion, data normalisation, and reconciliation, you can do much more on the trade date (T+0). And you need to be comfortable with complete data integrity on trade date, because there is no time to agree on anything afterwards. Effectively, by 9 p.m. Central European Time, you’ll need to have checked, affirmed, confirmed, and validated the settlement instructions for T+1 settlement.”
Subscribe to our newsletter
