Because of its position in the company, Treasury has always been the custodian of a treasure-trove of corporate financial data. But the scale and depth of the data at their disposal today far outstrips anything they dealt with a decade ago.
As infrastructure becomes more connected and real-time, organisations are gaining sharper visibility over cash positions, liquidity flows, FX exposures, and payment activity across markets and systems.
Treasury must consider how best to use this visibility: how can they move past forecasting to apply richer insights in ways that strengthen decision-making?
Technology plays a key role in answering this question. Advances in treasury infrastructure – from integrated treasury platforms to API connectivity – have made it easier to consolidate and access financial data at scale, and activate it as a continuous source of insight.
Case in point: Artificial Intelligence (AI) tools that leverage Machine Learning (ML) and Natural Language Processing (NLP) to turn raw, structured, or unstructured data into interactive visualisations and actionable, real-time insights.
Transformer architecture-based big data AI models – like Ant International’s Falcon Time Series Transformer Model – take billions of parameters of historically observed, timestamped data to predict future values. To increase output accuracy, these models dynamically adjust look-back horizons – shorter windows of historical data during periods of high volatility, and longer windows in low-volatility environments.
Areas such as FX and liquidity forecasting, where outcomes are shaped by multiple variables and changing conditions, are particularly strong candidates for improvement through AI models.
For instance, AI models can train on historical cash flow data, map behavioural patterns in payment timing against changing market conditions, and produce an automatically-updated cash flow forecast rather than a periodic snapshot.
Another AI model might analyse intraday volume ratios at minute-level granularity. By refining its look-back context on a case-by-case basis, transformer-based models can statistically forecast volume spikes and liquidity fluctuations with greater accuracy than traditional point-prediction models in certain market conditions.
In either example, the outcome is the same: treasury teams can take a more dynamic approach to forecasting. They can react in real time to emerging developments, and concentrate their analytical effort on genuine forecast variances rather than routine reconciliation.
The real-time insights yielded by new treasury technologies translate to concrete, practical benefits for teams.
Application Programming Interfaces (APIs) enable instantaneous communication between internal systems and banking partners, in place of traditional batch transfers.
This enables quicker decision-making for treasury: As APIs update information in real time, treasury professionals can make up-to-the-minute decisions regarding funding, investments, and risk mitigation based on live liquidity positions rather than end-of-day bank reports or manual spreadsheets.
Treasury Management Systems (TMS) consolidate fragmented data into a “unified source of truth” with real-time visibility across global bank accounts, currencies, and entities.
Treasury professionals using TMSs can be more proactive in planning future actions: advanced TMS platforms allow teams to model diverse financial scenarios and run what-if analyses with snapshot comparisons. These simulations empower teams to build targeted contingency plans and robust risk-mitigation strategies well in advance of actual market volatility.
Finally, AI-powered platforms can automate a wide variety of tasks in the treasury function, from invoice matching to cash forecasting to reporting. Some advanced systems use Machine Learning (ML) and fuzzy logic to automatically match bank transactions against general ledger records, handling the high-volume, repetitive work of identifying discrepancies while achieving much higher matching rates relative to human output.
This shifts productive time away from aggregation and toward analysis. By automating data aggregation and transaction matching, AI-powered systems allow treasury professionals of companies earning between $1 billion and $10 billion of revenue to reclaim up to 52% of their time previously lost to manual data collection, according to industry research. [1]
Treasury is entering a new phase of maturity. Access to data is no longer the main constraint: treasurers must now consider how effectively that data can be translated into insight that improves funding, liquidity, and risk decisions.
This is where technology proves its value: connected systems, real-time data flows, and AI models help treasury teams replace static processes with continuous visibility and faster action.
By carving out the resources needed for more thorough analysis and stronger decision-making, technology allows treasury to be more strategic and more responsive to their principals’ needs – taking on a more central role in shaping organisational strategy.
At Bettr, our Falcon Time Series Transformer technology is designed to support that evolution. With approximately 8.5 billion parameters and growing, it analyses large volumes of market and transactional data to help treasurers better anticipate FX exposure, sharpen forecasts, and make better-informed decisions.
Contact us to learn more about gaining better insight through technology.
1. PwC, “2025 Global Treasury Survey”, https://www.pwc.com/us/en/services/consulting/business-transformation/library/2025-global-treasury-survey.html
This article is intended for informational purposes only and does not constitute legal, financial, investment, or other professional advice, nor does it constitute a recommendation of any product or service. This article should not be regarded as constituting an offer or a solicitation to buy or sell any regulated or financial products or services. It has not been reviewed by any regulatory authority in any jurisdiction. AI-based forecasting models are subject to inherent limitations, including the possibility of inaccurate outputs, and past performance is not indicative of future results. Bettr makes no representations or warranties regarding the accuracy, completeness, or applicability of the content, and readers are encouraged to consult with legal, financial or other professionals for advice tailored to their specific situation. Bettr does not guarantee the accuracy and completeness of this article and expressly disclaims any and all liability to any person in respect of the consequences of anything done or omitted to be done wholly or partly in reliance on this article.
This advertisement has not been reviewed by the Monetary Authority of Singapore or any other regulatory authority in Singapore.
