BearingPoint IT Advisory supports clients with digital transformation

20 April 2018

The IT Advisory practice of BearingPoint helps clients with developing and implementing digital strategies. To bring digital transformation transitions to a successful close, the firm’s approach combines advice with governance and deep functional knowledge. “Providing direction and guidance is key” said John Septer, head of BearingPoint's IT Advisory service line in the Netherlands. 

How do I make strategy tangible and concrete? “That is one of the most frequently asked questions we receive from clients,” explained John Septer, who has more than ten years of experience in business IT. “There are multiple avenues that lead to the right solution, yet at the same time there are also many dependencies, which combined make the equation a challenging on.”

In their bid to successfully manage complex IT changes, organisations can according to Septer benefit from appointing a transition lead. He elaborated: “A transition lead helps shape, advise on and support decision-making in order to ultimately take the right steps, in a fact-based manner. A transition lead also ensures that the various stakeholders are aware, and that they have an understanding of options and decisions, plus know in which way the transition is heading. Control is the key word here.” 

BearingPoint IT Advisory supports clients with digital transformation

Besides ensuring ‘control’ is in place, it is key that organisations have clear insight in the capabilities that are necessary to generate the right value. Septer said: “Capabilities stand for technology, people, the process, content and data. These five aspects of a capability are all connected to each other and therefore serve as a precondition for success. It is a big puzzle that can be different in every situation – the maturity of an organisation in people and technology is hereby a crucial factor. That makes such transitions complex, but certainly not unsolvable.”

Septer indicated how BearingPoint’s IT Advisory practice can help with the actual translation of (digital) strategy into concrete plans. “Across the five aforementioned capability, we provide for the right translations into action, as well as bring in a proven methodology and implementation track record. On top of this, we ensure that there is a clear link with the set objectives within the organisation (strategy) and that change support is nurtured for the necessary activities.”

The IT Advisory arm of BearingPoint provides several offerings that help clients with their IT challenges, including: Setting up a digital business model; Agile ways of working; Software selection based on prototyping; IT related maturity scan on technology and people; (ERP) system implementation and change management; Roll out of an omnichannel commerce platform; and filling in the role of Digital Officer / Digital Solution Architect. 

To come to a successful result, BearingPoint follows an integral approach, combining several competencies that sit within its IT Advisory team. "We offer companies an independent view, guidance and advice on digital transformation processes. Direction and control are pivotal to our approach. Together with our customers we make the difference,” concluded Septer.

AI can improve operational risk management in banking

17 April 2019

Risk management is an integral part of banking. By taking financial risks, banks are able to generate the profits that are necessary to survive. Risk management aims to control this process by making potential losses more predictable. This makes the bank more robust to external fluctuations.

Whereas profits can be made by accepting certain financial risks, operational risk is intrinsically different. Operational risks only cause losses – financially in terms of bottom-line impact and non-financially in the form of for instance reputational damage. The consequences of operational risk events can have a large impact on an organisation and the financial system as a whole as experienced during the last financial crisis. It is not surprising therefore that operational risk is receiving more attention within the financial sector, with banks trying to minimise the operational risks they take, given the resources available while keeping in mind the strategic goals of the organisation.

In the past decade, there has been major progress in the development of artificial intelligence (AI). AI algorithms excel at data analysis and have evolved to the point where they surpass human performance for a wide variety of tasks. More and more businesses exploit these technological advances to optimise different kinds of processes such as marketing, sales and e-commerce, manufacturing and logistics. In today’s growing data-driven world, this trend is expected to continue on the back of widening opportunities for use cases, including in the area of operational risk management.AI can improve operational risk management in banking

Challenges in operational risk management

In June 2011 the Basel committee published the Principles for the Sound Management of Operational Risk (BCBS), which provides a framework for the development of proper operational risk management. Three years later, a survey was conducted to measure to what extent banks complied with these principles. One of these principles states that banks should write a risk appetite and tolerance statement. Banks reported that this is more challenging for operational risk than for other risk categories and attributed this to the nature and pervasiveness of operational risk. The banks that did comply with this principle often reported the use of backward-looking metrics of operational risk, such as operational losses as a percentage of gross revenue.

The above example underlines the challenges which banks face in the management and measurement of operational risk. Compared to financial risk, operational risk is a more qualitative field of study. Whereas financial risk management has been the main priority of banking for a longer time, operational risk management is much younger resulting in less extensive historical data. Predictive modeling becomes more of a challenge in this situation.

On top of that, the events in operational risk are much more diverse in scope. The Basel committee defines operational risk as the “risk of loss resulting from inadequate or failed internal processes, people and systems or from external events”. Internal fraud, data leakage and reputational damage are very different problems, yet can be very closely related as well. A lot of the data in operational risk consists of textual input which contain qualitative information. The qualitative nature of operational risk is reflected in the Basel framework, which encompasses guidelines for organisational structures, culture and awareness, and qualitative reporting.

The computer as a reader

Artificial intelligence could play a valuable role in upgrading operational risk practices, with in particular machine learning – the field of self-learning computer algorithms – showing promise. Machine learning algorithms can make predictions based on data fed to the algorithm. Recently, major progress was made in the field of natural language processing (NLP). NLP focuses on using textual data for predictions. As an example, an algorithm can learn to rate hotel reviews. By processing large amounts of reviews together with their given ratings, the algorithm can learn to give a rating to a new review it has never seen before.

“There are many opportunities for operational risk management to exploit AI and other related technological advances.”
– Lars de Ruiter and Matthias Geerse, Solid Professionals

In operational risk, many textual reports are written regarding specific risks or risk appetite of the organisation as a whole. These reports are generally written by risk managers – experts in their field. Assuming that risk reports hold information which is absent in historical loss data, this information could be extracted and used for predictive purposes. So why should computers perform this task instead of risk experts? The answer lies within the physical limitations of human being – as well as the natural biases they have to cope with. While humans may struggle to remember a piece of text, a computer algorithm easily processes thousands of books and finds structure within using statistical methods with near-perfect precision.

Towards prediction of operational risk

The performance of NLP algorithms using word embeddings has increased tremendously of late. Using the embedding as a starting point, these algorithms learn the meaning of full sentences and use this information for further predictions. This is more generally called sentiment analysis. A popular application in the financial setting is predicting stock price movement from news articles, or twitter feeds. A recent paper by Denmark’s national bank predicts corporate distress of firms by analysing their annual reports. Taking textual data under scrutiny can make a contribution to quantitative predictive modelling, even in the financial sector where the use of words is very different to day-to-day language.

In summary, there are many opportunities for operational risk management to exploit AI and other related technological advances. One possibility is the classification of risk events. As an example, a computer algorithm can read risk descriptions written by risk managers and classify them according to their impact and frequency. Combining loss data with risk reports, improved prediction of risk events might lead to more accurate prediction of future losses. A different application is the measurement of more abstract concepts such as the financial health or maybe the cultural aspects of a company.

An article by Lars de Ruiter and Matthias Geerse. Both are finance & risk consultants at Solid Professionals, a consultancy from the Netherlands.