The use of AI in corporate treasury is evolving fast. What began with traditional machine learning or rules-based algorithms used for fraud detection or basic forecasting has now developed into generative AI and advanced analytics. This can be used to interpret vast datasets to generate real time insights and simulate complex scenarios beyond humans’ capabilities.

One of the most important applications is automating the mundane. For example, AI can analyse payment patterns, counterparty performance and flag potential late payments even before they occur, enabling treasurers to proactively manage liquidity or trigger collection alternatives. AI-driven reconciliation tools can also match invoices and payments in seconds – a process that used to take hours or even days and was sometimes only semi-accurate.

“It’s difficult to put an absolute number on it, but organisations that are leveraging AI in treasury operations have seen up to 70% reduction in manual processing time and a 30–40% decrease in payment errors,” says Narang.

The role of AI in risk mitigation is another systemic change. From real-time fraud detection like spotting anomalies in payment patterns to scenario analysis for stress-testing liquidity, AI helps treasurers identify and address risks before they escalate. Many financial institutions are already using AI-driven tools to monitor transactions across jurisdictions, ensuring compliance and reducing exposure to operational and market risks.

Another benefit of AI is a new level of precision to decision-making. Cash flow forecasting is frequently inaccurate due to the quality of disparate data or input in different formats, but because AI can smooth these data anomalies with its standardised machine learning based platform it is bringing a new level of precision to cash flow forecasting. “This will be a tectonic shift when it comes to cash flow forecasting which will have an immediate positive impact on the headroom numbers,” he says.

AI doesn’t just make sense of internal data. It also excels at synthesizing internal forecasts with external data like live FX rates or geopolitical news. Imagine a scenario where AI identifies a potential currency risk based on forecasted cash flows and market volatility. It can then recommend strategies like currency swaps, reducing certain exposures or flag opportunities to optimise working capital. This kind of proactive decisioning is becoming increasingly common, especially in volatile markets.

And AI will soon go beyond just suggesting actionable insights. Full automation, where the technology executes the trade or transaction end-to-end is just around the corner. It still requires robust governance and human oversight, especially for high-value or complex transactions, but Narang points to an industry consensus that suggests that within the next three to five years this will be normal. “We’ll see more semi-autonomous treasury operations, with humans focusing on oversight and exception management,” he predicts.

Steps to efficacy

AI’s effectiveness hinges on access to clean, consolidated data. That’s why creating a single source of truth, or a data lake, has become essential. It allows AI models to work with comprehensive, up-to-date information, driving more reliable insights. It’s why leading banks (and corporates) are investing heavily in integrating data across their networks, so clients benefit from a unified, trustworthy view of their financial landscape.

AI adoption isn’t just about upgrading technology – it’s about evolving mindsets and capabilities too. Narang counsels on the importance of treasury teams embracing data-driven decision making and new skills in analytics, digital tools and AI literacy. “Upskilling is a top priority for successful AI integration. At Standard Chartered, we’ve invested in regular training and cross-functional workshops to ensure our teams and clients are ready to leverage AI’s full potential, and we see this trend accelerating across the industry,” he says.

AI is only as good as the business logic and context it’s given, he continues. Treasury professionals play a vital role in defining use cases, validating model outputs and ensuring that AI solutions align with real-world needs. This partnership between domain expertise and technology is what drives successful AI adoption.

In another step, Standard Chartered has implemented a Responsible Artificial Intelligence Standard to ensure every AI use case adheres to fairness, ethics, transparency and accountability. As AI becomes more advanced, explainability features are increasingly important, he explains. “This transparency is vital, not just for compliance and auditability, but for building trust among treasury professionals and regulators.”

Corporate treasurers must also balance building AI capabilities in-house with the benefits of control and customisation that brings, with the downside of added cost and resource-intensity. Narang suggests partnering with trusted fintechs and banks to bring scale, security and access to the latest innovations. Data security is paramount, so it’s essential that any data sharing is governed by strict protocols and compliance standards, especially given the diverse regulatory environments in which multinational organisations operate.

“AI isn’t just a technology upgrade – it’s a strategic enabler,” he concludes. “The future of transaction banking is intelligent, connected, and above all, human centric. By combining deep local expertise with global innovation, banks and leaders like Standard Chartered are helping clients turn data into decisions, and decisions into growth.”