Treasury's Manual Data Problem

Spending my career as a treasury practitioner and a technology provider has given me the opportunity to see this industry from several different perspectives: as someone buying and using treasury software, as someone building it, and as part of a much larger financial technology business.

On the surface, the industry is unrecognisable from the early days of my career. Much of the treasury's activity is now highly automated, bank statements are delivered electronically, payments can be generated and transmitted without anyone touching a spreadsheet, and cash reporting, reconciliation and forecasting have all benefited from APIs, better connectivity and significantly more capable software.

However, one aspect that's struck me throughout that time is how little one of the most basic treasury data problems has changed. When you look at many of the financial instruments a treasury team is responsible for, the data entry process can still be surprisingly manual. Loans, bonds, trade finance facilities, guarantees and other instruments are often entered into the TMS by a treasury professional who has taken the information from a bank portal, confirmation or document and recreated it in the system.

Once the data is there, the software can do a great deal with it. It can calculate interest, generate cash flows, support forecasting and produce reporting – but the problem is that someone still has to get the information into the system in the first place.

We've automated the payment, but not the instrument

The TMS market has made considerable progress in automating what happens after an instrument has been entered. The leading vendors offer sophisticated functionality for managing loans, investments, debt, derivatives, guarantees and other financial instruments throughout their lifecycle. However, what we've overlooked is the step before all of that: how the underlying instrument data gets into the system.

A treasury team shouldn't have to act as the intermediary between its banks and its TMS, and if a bank already has structured information about a loan or bond, a natural question is why the treasury team should have to re-key that information somewhere else. The same applies to trade finance transactions and other instruments where the underlying data already exists electronically.

This is exactly this problem we're working on. The objective is to take financial instrument data directly from banks in structured formats and make it available in the TMS without requiring someone in treasury to enter every detail manually. It should be no different to logging onto a bank portal – the user enters their credentials, and the information is there for them to see.

We've already put this into practice with several customers, where equity, bond and trade finance data ingestion has been fully automated. Instead of asking the treasury team to maintain those positions manually, data flows into xfolio to be used by the platform immediately.

AI should take work away from treasury

There's a lot of excitement about AI in treasury, and much of the discussion focuses on areas like forecasting, anomaly detection, recommendations and natural-language interfaces. While these are certainly useful applications, there's another opportunity that is potentially more important: using AI to remove much of the manual work involved in the process.

In fact, the best AI experience for a treasurer may be one they barely notice. If a system can identify information from a bank, understand what it represents, validate it, map it to the right instrument and update the treasury position, the user doesn't need another screen or another workflow. The system has simply taken care of something that previously required human intervention.

Instead of building an intelligent assistant that helps treasury professionals complete manual processes, we can start designing software that handles more of those processes itself.

The TMS should know what is happening

When I started TreasuryXpress in 2009, connectivity was one of the biggest challenges in any implementation. Connecting banks, configuring formats and establishing reliable data flows could take a significant amount of time, particularly for treasury teams working across multiple banks and countries.

A lot has changed since then: APIs are a much more common part of the banking landscape, and treasury platforms have far more options for establishing secure, automated connections. The industry has invested heavily in solving the connectivity problem, which allows us to tackle the next layer of manual activity. The goal should be a treasury system that's continuously building an accurate picture of the company's financial position, instead of one that depends on treasury staff to keep feeding it information.

Imagine logging in each morning and seeing current cash balances, bank transactions, loans, bonds, trade finance facilities and other instruments already reflected in the system. Your forecasts can work from that information, reporting can be generated from it and changes can be identified without someone spending the first part of the day checking different bank portals and updating spreadsheets.

I've always believed that treasury technology should reduce the operational burden on the treasury team – not just digitise it. There's a difference between giving someone a better interface for entering a transaction, and removing the need to enter the transaction at all. We've spent decades making the former more sophisticated, and the next phase of treasury technology innovation is about making the latter possible.

We built TreasuryXpress around the idea that treasury software could be easier to deploy and more accessible to companies without large treasury or IT teams, and my experience at Bottomline reinforced how closely treasury, payments and banking infrastructure are connected.

xfolio takes these lessons and applies them to a different generation of technology, where AI and connectivity aren't features added around the edges of the product, but part of the way the platform is designed.

Ultimately, the idea of manually entering a financial instrument into a TMS ought to feel as dated as manually typing a bank statement into a spreadsheet, and while we're not at that point across every type of treasury data yet, the technology now exists to move much closer to it.

Our goal, which we hope to achieve before the end of this year, is that a treasurer using xfolio can log in and see a far more complete picture of the business, with the majority of the underlying data arriving automatically rather than being keyed in by the treasury team.

For me, that is what the future of treasury management looks like: less time collecting and maintaining data, and more time using it to make better decisions.