AI x 2: Actual implementation of artificial intelligence

If you’ve briefly thumbed through any treasury or finance magazine of late or have cast a passing glance at agenda topics topping the major cash and treasury conferences over the last 12 months, it will have left you in no doubt: The robots are coming.

But what is it that they are coming to do exactly? Humiliate your feeble human mind, take your job and steal your spouse? Or are they rushing in to take away your dull, repetitive tasks and leave you with the time to focus on strategy and forward planning and shine in your organisation? And what will AI actually mean for areas such as cash management and trade finance?

“AI should be seen as an assistant for treasury that can monitor and execute a given set of tasks,” explains Christian Skøtt Maltesen, Head of Artificial Intelligence & Machine Learning at Nordea. “In short, treasurers can focus on the value-adding parts of the job and leave the repeatable and overly complex analyses to machines. Treasurers can focus on deciding policies, improvement opportunities and address the challenging issues where AI will bring additional insights.”

The field of Artificial Intelligence was established back in the 1950s but only now has it become a column-inch coloniser and crowd-puller at conference keynotes. And it’s not that AI algorithms are that new either – it’s just that explosive growth in online interaction has created a veritable embarrassment of analysable data, from which AIs can learn. This capability, built on sharp advances in computer performance, has also been fuelled by massive investment sums; in 2016 alone, USD 5bn was pumped into the industry.

But if treasury and finance departments are to take advantage of these advances, what do they need to do? Well, according to Thomas Jensen, Head of the Data Science Lab at Nordea “a lack of defined use cases, a dearth of suitable skills and no internal infrastructure and processes that can support AI applications” tend to be the early show-stoppers. “Without clearly defined use cases, that are grounded in the business, there is no reason to start implementing AI,” he expands. “Surprisingly many organisations are following a hype around AI, without having done their “due diligence” around whether AI can make a difference on their bottom-line. Even when this initial stage has been completed, challenges remain. The organisation then needs to put the right team in place, and because AI applications often have a very tight coupling between data and logic, this makes it hard to fit this new team in to traditional IT setups.”

Christian adds: “There can also be issues in terms of excessively high expectations from managers early on, resistance from staff, and access to capable staff to avoid paying expensive externals to do the work.”

Within treasury, some use cases have already been talked about while other areas seem to offer future opportunities for AI and RPA robotics (RPA robots are ‘dumb’ acting exactly along pre-set rules, whilst the AI is ‘self-learning’ and can update the rules based on new data). “With the possibility to automate many processes, it becomes important to make an inventory of said processes, and evaluate whether they make sense in the first place. To execute a bad process efficiently is not the goal,” chastens Thomas.

Christian underlines the importance of identifying the optimum applications of RPAs and AI: “It is crucial to immediately identify relevant use cases in order to become more familiar with the technology, the prerequisites and thus the opportunities for deployment. Based on this, a future road map can be developed. When we talk about which tasks in treasury or finance departments are suitable, we need to examine and understand what the business needs are, e.g., if the need is to automate rules-based processes where compliance and accuracy are critical, the RPAs are ideal. If there is any ambiguity, usually when the inputs into a process are unstructured or where there are very large amounts of data, then AI is the appropriate technology to use because it can manage that variability and, most importantly, get better at it over time through its own experiences.”

Nordea is already employing robotics and AI on a daily basis. Early AI technologies, i.e. “rule-based AIs” typically based on the “if-then” premise, have been around for quite some time and have boosted productivity in internal processes, such as compliance. The second wave of AI is now being rolled out: chat bots in customer service, dealing with routine customer queries. Thomas sees similar opportunities for treasuries, commenting: “Risk assessment and cash flow forecasting are certainly possible with today’s technology. Financial planning will probably be possible in the near future. Elsewhere, detection of fraudulent cases is an area where AI already is making a big impact, by filtering out the false positives, hence lowering the number of cases for human investigators.”

Whether KYC can be improved by AI is of particular interest to corporates. Thomas is tempered in his view saying “it depends… response rates can probably be driven up” but Christian is more optimistic, expanding: “Yes, if (and when) we move beyond a compliance-centred approach to one that focuses on determining the right questions, asked at the right time/frequency to detect undesirable patterns, then this can become a powerful alerting tool.”

So, yes, the robots are certainly coming, but they are coming to help us and not replace us. In fact, new jobs based on new skills will be created, most likely in a support unit to treasury and finance departments. Alongside an increased demand for machine learning and development resources, a new role of AI manager will emerge with responsibility of monitoring the AI applications to make sure they adhere to common standards and don’t begin to make biased decisions.

That does not mean treasuries and finance professionals should sit on their hands and wait. Skills within these areas can still be developed to ensure adoption and employment of AI is optimised, with Thomas advising: “One key skill, which is remarkably close to statistics, is machine learning. Knowledge of the most common algorithms, and how a system is build up around them will be important. This will guide people when considering the pros and cons of a given algorithm, and also what AI can do efficiently in general. This knowledge can be integrated into the departmental tool kit.” Christian concludes: “There should be a focus on appreciating processes and understanding data. This will allow these areas to deploy and follow-up on the output and quality of AI – and where they can improve/refine the work. If they can begin to understand their current high volume transactional processes and their performance, it will become clear what can be improved by AI or RPAs.”

The information provided within this website is intended for background information only. The views and other information provided herein are the current views of Nordea Bank Abp as of the date of publication and are subject to change without notice. The information provided within this website is not an exhaustive description of the described product or the risks related to it, and it should not be relied on as such, nor is it a substitute for the judgement of the recipient.

The information provided within this website is not intended to constitute and does not constitute investment advice nor is the information intended as an offer or solicitation for the purchase or sale of any financial instrument. The information provided within this website has no regard to the specific investment objectives, the financial situation or particular needs of any particular recipient. Relevant and specific professional advice should always be obtained before making any investment or credit decision. It is important to note that past performance is not indicative of future results.

Nordea Bank Abp is not and does not purport to be an adviser as to legal, taxation, accounting or regulatory matters in any jurisdiction.

The information provided within this website may not be reproduced, distributed or published for any purpose without the prior written consent from Nordea Bank Abp.

Related articles