Money20/20 Europe 2023: AI data reconciliations – built to perform

Jethro MacDonald, product manager, artificial intelligence (AI) and machine learning (ML) at SmartStream, runs over the capabilities of its latest data reconciliation and cash management solution.

SmartStream Air: a robust, reliable and cost effective solution for data reconciliations and cash balances

Cloud-native and AI-enabled technology are both frequently mentioned as the ideal; a destination to get to.

But a proven case use lies in SmartStream Air. It leverages both the cloud and AI to produce something that is robust, reliable, cost effective whilst providing time and operational efficiency at scale for data reconciliations, and now for cash balances.

The latest version, V.8, extends the existing functionality out to onboard cash balances faster and more accurately, whilst further improving cash reconciliations. Indeed, cash balances can be tricky to negotiate and getting it wrong can result in financial losses and even regulatory penalties. The latest version mitigates the risk of either of those two things happening by translating existing AI-enabled technology and exception management capabilities to the cash management space.

SmartStream Air was initially launched at Sibos in 2019; its first iteration focused on the need for an efficient and cost-effective means to manage data reconciliations. The concept was that the AI can autonomously build and learn data matching behaviour, transforming typical data tasks – effectively reducing the time it takes to do such tasks to just seconds and it is proven to deliver highly accurate results.

Function wise the initial aim was to tackle intersystem risk – where data is transferred from one system to another. That evolved to support different file formats and data structured, using AI to map and direct the journey of data from one system to another and to get the matching rates to new levels, which the industry had never seen before.

The next task was to tackle the day-to-day issues of matching and using past decisioning to inform future decisions that needed to be made about where a piece of data should go. This is a massive efficiency gain as well as a knowledge gain, as all the information is retained within the software; something that is very relevant at a time when people move roles frequently and they take the knowledge with them when they leave a firm.

Beyond data matching the classifications of exceptions are now a possibility; labels can be applied to data and priorities set for dealing with exceptions. Using the power of AI the process becomes more efficient over time as it learns from previous exceptions that contains similar transactions.

In addition, to extending the function of the solution, the firm is also extending its reach in other industry sectors with use cases for how SmartStream Air’s AI capabilities can process data with more accuracy and efficiency.

“We feel that the time is right to be extending the reach; the solution is proven and reliable at a time when cultural acceptance of AI and ML is growing. People see its use within their everyday life and from there it is not too much of a step to imagine how it could be applied to business functions,” MacDonald comments.

Of course, of the biggest issues that financial institutions are facing is getting the data right in the first place. SmartStream Air can help with that too, the solution can take in unstructured data, normalise it, and then apply the AI technology to deliver usable data sets. Indeed, the solution is agnostic to structured data formats and can easily integrate with existing systems to perform additional analytics on reconciled data. In addition, it can provide business intelligence by connecting to third-party platforms which are already deployed in-house.

“We have also invested heavily in the concept of explainability so that people are not unnerved by a black box approach. Instead, we aim to be transparent and to be able to show what brought the solution to a decision and how likely that is in percentage terms to be the correct decision. For example, the observational learning will highlight strings between various records and then say how close a match they are. It is demonstrable and auditable and this is comforting in an era where people are willing but sometimes not confident enough to use AI to its full potential,” Macdonald explains.

Another key facet to the solution is the cost of ownership – and this still holds true today. A multi-cloud approach makes for a reliable, scalable, and cost-effective means to tackle large volumes of data; scaling up and down as required.

Speed of go-to-market is also important; SmartStream Air has been designed to be live the same day as it is installed and is intrinsic to use.

“At the end of the day we have developed something that is proven, secure, cost efficient, reliable and accurate that will seriously enhance operational efficiency and do all the heavy lifting. There’s nothing not to like,” concludes MacDonald.

Sponsored by SmartStream