AI will create a new frontier in banking

Since the mid-2000s, the entire spectrum of financial institutions – retail and wholesale banks, asset and wealth management firms, investment banks and clearing houses – have been under tremendous pressure to restructure to drive operational efficiencies and growth. The main drivers for this include the following:

• Regulations – A global European bank recently cited 80+ new regulations on their radar for implementation. Regulations are here to stay, across major markets globally, so every financial institution will have to continue to adapt.
• Revenue and Cost Pressure – Due to the low interest rate environment, margin erosion continues to impact the financial health of the industry at large. On top is the constant challenge to win new customers and markets, which makes achieving top-line growth an even bigger pressure point.
• Customer Experience – The ever-increasing expectations from customers, especially millennials, that rely almost exclusively on mobile channels versus traditional branch offices, has completely flipped business models upside down.  Customers today, of all generations, have vastly different expectations in mind.
• New Competitors – The increasing competition from other banks and non-banks, such as fintechs, is creating an entirely new landscape for traditional players.  Market-share is constantly being challenged, and this will only heighten in the years ahead.
• New Technologies – The introduction of new automation capabilities and digitally-advanced solutions is pushing the banks to consider new business models.

 

What is artificial intelligence?

AI as defined by our experts is as an area of computer science that emphasizes the creation of intelligent machines that work and react like humans. Some of the activities that AI technologies are designed for include image and speech recognition, learning, planning and problem-solving.

AI is not a one-to-one replacement for people. It’s not an all-powerful product capable of the same cognitive thought as we are. Instead, AI realistically consists of technology and autonomous or semi-autonomous machines that take on tasks or jobs that we either don’t want to do, or may be are unable to do. In short, AI is the science of making machines do those things that would be considered intelligent if they were done by people.

 

Start with a strong business case

Embarking on an AI journey can be a daunting task. Organizational bureaucracy, long decision cycles, stakeholder resistance and budget pressures will always remain a threat.  However, to mitigate some of these factors, we recommend firms to commence with a well-designed business case which is tailored to the specific function(s) in focus. For example:

• Front Office – Sales related business cases in this area are typically based on recommendation engines leading to revenue growth through increased customer satisfaction.
• Middle Office – Business cases are mainly built around risk topics, therefore cost reduction focused and based on pattern recognition algorithms to detect anomalies before risk materializes.
• Back Office – Here it’s all about process optimization, and therefore also cost reduction focused (e.g., based on prediction of future needs and proactive remediation).
• IT Operations – Business cases in the IT operations area are similar to back office use cases

 

How to translate planning into practice

Many organizations are currently experimenting with AI implementations within their innovation teams.  However, this means programs often remain far from the mainstream implementation teams where real scale can be achieved. An approach that can offer more measurable upside and return, we recommend the following:

1. Conduct a top-down exercise – This is a key step to determine the right roadmap based on prioritization of use cases and opportunities. Most importantly, this also allows you to deeply align with any program in the overall business strategy.
2. Data is your currency – The success of almost any implementation depends heavily on the availability and structure of data in high volumes. Effectiveness increases with share of structured data across its inputs, but one of the strengths of AI is the ability to integrate also unstructured data.
3. Quality assurance – At a macro level, three data sets need to be identified for any successful AI implementation: a PoC input data set, the main training data set and one (or several) validation data set(s).
4. Proof-of-value exercise – Independently from top-down (strategic) or bottom-up (random sample use case) approach, run a proof-of-value exercise (e.g., a proof-of-concept with a sub set of real production data to validate and refine the high-level business case).
5. Roadmap – Based on the refined business case, a roadmap and master plan for implementation across the enterprise can be created. We recommend to take other factors into account when prioritizing use cases for the roadmap. For example, the urgency or the increase in customer/employee experience that is desired.

Begin your journey.

Solving key business challenges and opportunities leveraging the power of AI can achieve much more than cost savings and operational efficiencies. It has the strength to impact every facet of the customer experience, which is in the end an over-arching goal almost every business is looking to achieve.

In addition, if the conversations are focused around how to better optimize and use the talent of your workforce – or how an organization can amplify the potential of their people through smart use of automation – an organization can plan and implement value-driving programs with the full support of their employees.

To read our entire Point-of-View on this topic, you can access our full AI in Banking POV paper here.