HAL, how much money is left for our company after inflation, wage increase and euro weakness?

What challenges will the treasury department and treasury management systems face in the future? Will the treasurer soon ask his treasury system about the future liquidity development of his company like other digital assistants and only need it if the HAL 9000* computer equipped with artificial intelligence is unable to make headway on its own despite predefined sets of rules?



Digitisation and Refactoring in All Alleys.


It is well known that many treasury processes can be digitalised. Treasury management systems have been around for more than a quarter of a century, electronic banking for more than 35 years. However, development never stopped and has gained a lot of momentum, especially with the introduction of secure internet technologies and the changeover to XML-based file formats.


The exchange of electronic account information and SEPA payments with some European banks via EBICS have established themselves as standards at least in Germany, Austria and Switzerland and France. In addition, other topics are being examined for their digitalisation capability, such as electronic bank account management (eBAM for short), automated bank fee control or digital guarantee application.


ISO 20022 Revolutionises Financial Management


With its standard 20022, the International Standardisation Organisation (ISO) has already provided far-reaching descriptions for the structure of suitable files for the exchange of financial information in different areas: e.g. "acmt.nnn" for account management or “camt.086” for bank fee reporting. The replacement of the SWIFT formats "MT7nn" by corresponding "trsv.nnn" files is in preparation.


CGI as a Global Standard for Payment Files?


Since 2010, representatives of banks and non-banks have been meeting in the "Common Global Implementation" (CGI) initiative trying to develop standards for global payment transactions that meet the requirements of all nations, accommodate sufficient information in different languages and can be sent around the globe cost-effectively in a fraction of a second.


Only Standardised APIs Can Be Used Efficiently


In addition to the renewal and internationalisation of file formats, the possibilities for their efficient exchange are also in motion: instead of file-based interfaces, information technology today likes to use web services and application programmable interfaces (APIs), where data flows from one system to the other without intermediate storage. In the future, banks will offer their customers a wider range of functions via APIs than before, e.g. in addition to payment transactions and account information, they will also be able to open, change and close accounts within the framework of eBAM. However, this will only benefit the corporate treasurer if the APIs follow a common standard and each bank does not cook its own soup.


The query of different data for the fulfilment of the Know-Your-Customer (KYC) regulations has shown that such individualised actions do not go down well with customers.


Whether New or Old: Data Protection is Crucial


With both new and old technology, care must of course be taken to ensure that security regulations are adhered to. In the future, effective encryption and authentication procedures will be supplemented by authorisations that allow conclusions to be drawn about the actual person making the payment, so that the actual originator of the payment can always be verified. Corporate seals, which, like a BIC, only contain company-related sender information, are a thing of the past.


Work Support Through Robotic Process Automation (RPA), virtual IBAN and AI


Routine processes such as the reconciliation of account statement items with planned turnover have long been automated by treasury management and electronic banking systems on the basis of predefined rules in the shortest possible time. Virtual accounts can increase the hit rate and artificial intelligence (AI) can effectively support the optimisation of processes.


Robotic Process Automation


Software robots (bots) can be used to collect similar information and take over these tasks in order to relieve humans of additional work. AI can be used to learn the tasks within the framework of Robotic Process Automation (RPA for short). In most cases, however, this is not necessary if sets of rules are cleverly defined. To introduce RPA, the processes under consideration must be analysed in detail. There is a risk of being tempted to completely redesign the process and create additional work instead of making it easier.


Virtual Accounts


"Virtual accounts" can be used for outgoing payments and "virtual IBAN" for incoming payments. In both cases, account numbers are used which correspond to the syntax of the IBAN, but behind which there are no chargeable bank accounts. To optimise the allocation of incoming payments, companies can buy virtual IBANs from their bank, all of which refer to a "real" bank account. Invoices to customers are sent with the virtual IBAN and through this unique feature the company recognises from whom the incoming payment originates or to which project or group company it is to be assigned. Through automation, errors are avoided and the allocation process is accelerated. Similarly, outgoing payments related to specific projects, objects or subsidiaries can also be made via virtual accounts, whereby only the one real payment account is actually debited. The number of further chargeable accounts can thus be reduced in some cases.


Big Data and Artificial Intelligence


More and more system providers are trying to accommodate AI in other areas of financial management, e.g. in monitoring outgoing payments or in liquidity planning. Finding out irregular payments is based on the principle of pattern recognition: outgoing payments that do not correspond to the usual course are resubmitted for closer examination. This assumes that one otherwise has a "usual course" that shows a roughly consistent pattern. To improve liquidity planning, past patterns are also assumed on the one hand and projected into the future. In addition, cost increases, currency relations or even weather forecasts can influence the development of cash flows in self-learning systems and make scenarios more realistic. Here, too, however, a lasting development without disruptions or shocks such as sudden pandemics or wars is a prerequisite for the informative quality of the planning.


Predictive Analytics


It is often sufficient not only to divide past plan data into the two categories "fully transferable into the future" and "not 100% transferable", but to analyse the last data basis a little more precisely. Uncertainty can be taken into account by multipliers so that, for example, plan values with 50-80% probability of occurrence can also be transferred to future views. In contrast to artificial intelligence, human experience values are incorporated here and built into a set of rules.


Instant Payments


New technologies, more powerful algorithms and processors are making the world faster. The further development of the split-second TARGET2 clearing into the real-time system TIPS for instant payments makes the banks' previous practice of delaying the settlement of transfers by value dates, sometimes lasting several days, appear unfounded. The new market infrastructure service allows money transfers even in real time, around the clock and on all days of the year. Most instant payments today are processed in less than three seconds, allowing last-minute payments. To make sure that the recipient knows that money has arrived in his account, the camt.05N was subsequently developed on the basis of the camt.054 - this important information was probably overlooked in the speed with which Instant Payments were planned at the beginning. The only reason for sticking to the "slow" SEPA credit transfer in addition to Instant Payments is probably the technical possibility of retrieval, which is not available with the split-second variant.


Real-Time Monitoring


Instant payments, but also the rapid changes in prices, rates and interest rates are driving more and more trading in real time. Internet trading systems for money market and foreign exchange transactions, the digital application for sureties and guarantees, intercompany clearing for the immediate settlement of receivables and liabilities in one's own company, incentive systems for faster invoice payment through dynamic discounting require real-time monitoring that people can hardly achieve without appropriate software support.




In this respect, the question is whether treasury management still needs people at all. After all, software works according to clear rules and if corporate guidelines clearly specify the scope for action, the conclusion (or dissolution) of foreign exchange hedging transactions can also be carried out by machines until the defined hedge ratio is fulfilled. In processes that require or allow certain freedoms, artificial intelligence can support and only when the system detects a conflict is human intervention necessary.


For this, of course, the financial managers need to know what led to the conflict and how to evaluate the causes and alternative courses of action. Neural networks whose decisions are not comprehensible to humans do not help here. Rule-based systems should be understood by at least two people in order to avoid the risk that the management will one day no longer be able to explain how a precarious financial situation came about, possibly leading to the filing for insolvency and the downfall of the entire company.


Being modern and up to date with one's software systems is the dream of many treasurers. Perhaps they also dream of a voice assistant like the one you already use at home or in your car. Of course it would be nice not to have to laboriously compile all the information from the system via evaluations and reports. But it is also known that voice assistants simply record snippets of conversations for learning purposes and that the learning on-board computer HAL 9000 in the movie "2001: A Space Odyssey" became a danger to people after it learned of its planned shutdown. Software that corrects human errors is desirable, but only as long as humans remain in control.


* For all those who do not know HAL: HAL 9000 is a fictional computer on board a spaceship in the novel "2001:A Space Odyssey" by Sir Arthur C. Clarke, which was masterfully filmed by Stanley Kubrick in 1968 as a defining science fiction epic for the cinema. The heuristic-algorithmically programmed computer takes care of the operation of the spaceship and can already conduct dialogues with the five scientists, like today's voice assistants. It also seems to learn human behaviour in the course of the journey to Jupiter. When he makes a mistake in his calculations, the humans are surprised at the result presented and decide to shut down the computer. To prevent its shutdown, HAL kills four of the five scientists. The last one finally succeeds with difficulty in successively switching off the system.


The cash and treasury management system Trinity TMS is developed and maintained with a lot of know-how in Germany. Enquiries from customers are answered by Trinity itself, so that a personal relationship develops between our employees and users. The German team is supported by a project office in the Netherlands, as many companies based in Eastern and Western Europe are now part of the Trinity clientele.


Trinity offers its customers individual "private cloud" installations in German data centres to relieve the burden on customers' IT resources while complying with EU data protection directives and the DSGVO.


Trinity has shown continuity for 25 years and will continue to do so in the future. Customised, experienced, faster and personal.




    *The data sent will only be processed for the purpose of handling your request. You will find further information in our Privacy Policy