Cash visibility in 2026: why CFOs are finally getting serious about treasury architecture

A global manufacturer with operations in 14 countries recently discovered it had approximately $340 million sitting idle across subsidiaries in Southeast Asia, largely because its treasury team in Frankfurt had no real-time visibility into those accounts. The money was there. The company just could not see it clearly enough to deploy it. This is not a fringe situation. According to research from PwC's Global Treasury Survey (one of the more rigorous independent benchmarks in this space), over 60% of large corporates report that cash forecasting accuracy remains their single biggest treasury pain point, even after years of investment in financial systems.

The gap between what CFOs believe their treasury function can do and what it actually delivers is often significant, and it tends to widen precisely when conditions get harder: rising rates, tighter credit markets, geopolitical disruption to payment flows.

The structural problem most treasury teams are still sitting with

Treasury technology has improved materially over the past decade. Treasury Management Systems from vendors like Kyriba, ION, and SAP (all commercial platforms, and their marketing materials should be read accordingly) have added forecasting modules, bank connectivity layers, and APIAPIApplication Programming Interface: a standardised interface that lets applications communicate and exchange data without knowing each other's internal workings.View full definition →-based integrations. Yet the underlying architecture problem persists in most organizations: cash data is still assembled rather than live.

The typical mid-to-large corporate treasury team pulls bank statements via SWIFT MT940 files, reconciles them against ERP ledgers, adjusts for intercompany positions, and then produces a cash position that is, at best, 24 hours old. In volatile FX environments or during liquidity stress, that lag is not a minor inconvenience. It is a decision-making handicap.

The companies that have genuinely moved past this model tend to share a few characteristics. They have consolidated their banking relationships enough to make API connectivity practical, typically working with a core group of four to six relationship banks rather than twenty-plus local banks with inconsistent reporting standards. They have separated the data aggregation function from the analysis function inside treasury, so analysts are not spending the first three hours of each morning building a position. And they have made a deliberate choice about where treasury sits in the ERP stack: either deeply embedded or cleanly separated, not awkwardly straddling both.

Where AI is genuinely useful, and where it is mostly noise

There is a lot of vendor-driven enthusiasm right now around AI-powered cash forecasting. Kyriba, for instance, has been marketing machine learning-based forecasting capabilities as a core differentiator (a claim worth scrutinizing independently, since the accuracy gains depend heavily on data qualitydata qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition → and transaction volume). The honest picture is more nuanced.

For companies with high transaction volumes and relatively stable business models, machine learning can meaningfully improve short-term (one to thirteen week) cash forecasting. Pattern recognition across payables, receivables, and payroll cycles does add value at scale. For companies with lumpy, project-based cash flows, or with significant unstructured inflows, the models often underperform a well-maintained rolling forecast built by an experienced treasury analyst.

The more durable AI use case in treasury is exception detection: flagging unusual transaction patterns, identifying bank fee anomalies, or surfacing intercompany mismatches that would otherwise be buried in reconciliation files. This is less exciting than "AI forecasting" but it is where the technology consistently delivers measurable time savings.

What this means for the CFO

The conversation CFOs need to have internally is less about which treasury platform to buy and more about what problem they are actually trying to solve. These tend to collapse into two distinct priorities, and they require different responses.

The first is liquidity risk visibility. If the CFO cannot answer, within two hours, what the company's net cash position is across its five largest currencies, that is a governance failure, not a systems limitation. The fix often starts with banking structure rationalization before it touches technology. Getting from 30 banking relationships to 12, implementing a notional or physical cash pooling structure, and standardizing account naming conventions across subsidiaries are the enabling steps. The technology investment makes more sense once the underlying structure is simpler.

The second is working capitalworking capitalWorking capital is the difference between a company's current assets and current liabilities, measuring short-term liquidity and the funds available to run daily operations.View full definition → intelligence. Treasury increasingly owns or influences the conversation about receivables DSO, payables DPO, and supply chain financing programs. Companies like Unilever and Siemens have built treasury functions that actively manage working capital levers as a liquidity source, not just a finance metric. This requires treasury to have access to operational data (order management, procurement, customer payment behavior) that traditionally sits in other systems. The integration work here is real and takes time, but the payoff in terms of released cash is often larger than anything achievable through investment optimization alone.

A third pressure point worth flagging: counterparty risk monitoring. After several high-profile bank stress events in 2023, many treasury policies were updated to include stricter concentration limits and more frequent credit monitoring. Maintaining those practices in 2026, when the immediate memory of those events has faded, requires active governance, not just a policy document.

Practical priorities for treasury leadership

• MapMapUsing software to automate repetitive marketing tasks and campaigns, enabling personalisation at scale across channels like email, web, and social.View full definition → your actual cash visibility lag today: how old is your morning cash position, and what decisions are being made based on it?
• Before investing in treasury technology, run a banking structure audit. Platform ROIROIReturn on Investment: the ratio of net profit to the cost of an investment. A 300% ROI means each dollar invested returns $3.View full definition → drops sharply when the underlying account structure is fragmented.
• Distinguish between forecasting accuracy (a data and methodology problem) and forecasting discipline (a process and governance problem). Most organizations have the second problem, not the first.
• Build the working capital connection explicitly into treasury's mandate. If treasury does not have a formal seat in DSO and DPO conversations, that is an organizational design issue worth raising at the CFO level.
• Review counterparty concentration limits annually, not just after a market event forces it.

The CFOs who are getting the most out of their treasury functions in 2026 are not necessarily the ones with the most sophisticated platforms. They are the ones who have been rigorous about simplifying the structures those platforms sit on top of. Complexity is the hidden cost of treasury underperformance, and it rarely shows up as a line item until it becomes a crisis.