A joint perspective from the banking and corporate sectors
For decades, treasury operations were clearly structured: daily closings, cut-off times, periodic forecasts and historically established management frameworks formed the backbone of liquidity and risk management – both in bank treasury departments and in the treasury functions of large corporations.
In a world with clear time windows, predictable cash flows and sufficient reaction time, this model works well. Yet it is precisely these framework conditions that are currently breaking down.
Real-time payments, 24/7 availability of payment infrastructures, shortened settlement cycles such as T+1, volatile markets and new digital asset classes are forcing treasury functions into a new role. Treasury is evolving from a periodic function into a permanent control centre.
In this new reality, artificial intelligence is no longer a playground for innovation, but a structural enabler for controllability, speed and the quality of decisions.
A new concept of liquidity for banks and corporates
With the establishment of instant payments and global real-time payment rails, the concept of liquidity has changed fundamentally. Liquidity is now relevant at all times, not just at the end of the day. For banks, this means a massive increase in the demands placed on intraday and nostro management. Correspondent accounts must be continuously monitored, intraday funding efficiently managed and liquidity made available in real time – particularly to meet customer expectations of immediate payment processing.
Corporates face a similar challenge. In-house banks, cash pooling structures and working capital strategies must be coordinated in real time. Payment flows are no longer linear but event-driven. Traditional, rule-based models quickly reach their limits here. They are too slow, too static and not adaptable enough to reflect the dynamics of modern payment flows.
AI-based approaches address precisely this issue. They are capable of continuously analysing large volumes of heterogeneous data – from ERP systems, payment platforms, market information and historical patterns – and deriving reliable forecasts from it. Not only for balances, but also for timing, volatility and the probability of deviations occurring.
Inter- and intraday liquidity management as a key use case
One of the most effective applications of AI in treasury is intraday liquidity management.
AI models process payment transaction data, calendar information, market volatility and historical patterns almost in real time. This forms the basis for dynamic liquidity forecasts that are continuously updated.
For banks, this means significantly more efficient nostro account management. Over- and under-liquidity in correspondent accounts are identified at an early stage, intraday funding can be planned in a targeted manner, and costly ad hoc measures can be reduced. For corporates, this opens up the possibility not only of pooling liquidity across the group, but also of allocating it in a needs-based and forward-looking manner – particularly within centralised in-house banking structures. In-house banks, in particular, benefit disproportionately from AI-supported management.
The consolidation of payment transactions, cash pooling, financing and currency management creates a level of data complexity that is virtually impossible to manage efficiently by manual means. AI can act as an intelligent control element here: it anticipates payment flows, dynamically adjusts pooling structures and generates proposals for internal interest rates, refinancing or external investments. The in-house bank thus evolves from an administrative service centre into an active liquidity and risk manager.
Dynamic currency management instead of static hedges
Another key use case lies in currency management. In an environment of heightened FX volatility and shorter settlement cycles, accurate cash flow forecasts across currency borders are becoming increasingly important. AI models cluster cash flows by currency, forecast exposure trends and dynamically derive hedging requirements from these.
The key difference from traditional approaches is also its decisive advantage: hedging strategies are no longer based exclusively on static planned figures, but are continuously adjusted to actual payment trends. This reduces risks, avoids over-hedging and ties up less liquidity. Both banks and corporates benefit from lower costs and greater flexibility.
An end-to-end approach: Treasury in 24/7 operation
From a consulting perspective, the added value of AI becomes particularly clear when treasury workflows and risks are considered from end to end. In a combined banking and corporate scenario, the AI continuously processes the bank’s nostro positions, the corporates’ internal cash pools, and expected customer and supplier payments.
If the system detects short-term increases in outflows in a particular currency, it automatically simulates options for action: internal liquidity transfers, short-term refinancing or the use of existing FX lines – including cost and risk assessments in each case.
Treasury managers receive prioritised recommendations for action rather than static reports. Humans remain the decision-makers, but on the basis of significantly more reliable and rapidly available information.
AI and blockchain as complementary enablers
The growing importance of digital assets and blockchain-based settlement models is reinforcing this trend. Tokenised deposits, stablecoins and delivery-versus-payment mechanisms promise faster and more transparent settlement processes, but at the same time alter the mechanics of liquidity management. When payments and securities settlement take place almost in real time, the demands on precise liquidity management continue to rise.
This is where the combination of blockchain and AI demonstrates its particular strength. Distributed ledger technologies provide high-quality, virtually delay-free transaction data. AI translates this data into intelligent control logic: optimising cash buffers, prioritising payments and making informed decisions regarding on-chain versus off-chain settlement.
AI is not an IT project – it is a transformation initiative
Despite all the technological possibilities, one thing remains clear: the successful deployment of AI in treasury is not merely an IT project. It requires clean data environments, clearly defined governance structures and, above all, in-depth treasury expertise. AI does not replace experience – it amplifies it.
This is precisely where business and treasury consultancy comes in. It is not about the rapid deployment of a model, but about the sustainable transformation of control logic, processes and decision-making pathways.
However, the decisive factor for success does not lie in technology alone. Sustainable added value arises where treasury expertise, clean data management, clear governance and AI models are consistently brought together. AI does not replace experience – it makes it scalable, consistent and usable in real time.
Banks and corporates that view AI in isolation will be disappointed. Organisations that consistently integrate expertise, data and technology, on the other hand, create a treasury function that is both more efficient and strategically capable of taking action.
Conclusion
Treasury is rapidly evolving from a functional area operating on a periodic basis into a permanent management function. Real-time payments, T+1 settlement, rising market volatility and digital assets are significantly increasing complexity – and with it the demands on precision, speed and the quality of decision-making.
Artificial intelligence is no longer a promise for the future, but a central component of modern liquidity, risk and cash flow management for banks and corporates.
However, the decisive factor for success does not lie in the technology alone. Sustainable added value is created where treasury expertise, clean data management, clear governance and AI models are consistently brought together. AI does not replace experience – it makes it scalable, consistent and usable in real time.