There’s a wealth of Artificial Intelligence (AI) projects underway across all areas of financial services as banks explore the potential for the technology to revolutionise their business.
According to estimates from Oliver Wyman, if properly deployed, AI has the potential to increase banks’ revenues by as much as 30% in the next five to seven years, and to reduce their costs by 25% or more.
Despite this, many banks are still struggling to realise the promised return on investment. To successfully achieve this, banks need to re-focus their efforts and consider three key areas before embarking on any AI project.
First, they must identify scenarios where AI will deliver real value to the business. Next, they need to understand what data is available and how it can be used. And finally, collaboration with others will be essential to leverage the best AI technology and the power of the cloud.
Build the use case
First and foremost, it’s essential to identify scenarios where AI will deliver real value to the business. These use cases should be grouped into three key areas:
- Customer engagement – using AI to serve up personalised recommendations to customers, and to enable consistent engagement through a range of channels, for example, integrating chatbots and voice-activated assistants;
- Business analytics – optimising the processes that support decision-making, for example, to deliver a faster response on mortgage or loan applications;
- Operational AI – using AI to digitise entire processes to deliver increased efficiency, reduced costs, and other savings, for example, in areas like fraud prevention.
Most banks already have initiatives underway to deliver on the benefits of operational AI, typically starting with robotic process automation (RPA) and from there moving on to the deployment of real AI and machine learning. Used correctly, machine learning can deliver huge efficiencies.
For example, machine learning is already being used to deliver huge value in the syndicated loan market, where it is being used to extract information from complex loan contracts. These complex contracts can run to 800 or 900 pages meaning there is a massive scope for efficiency savings, reducing processing time from two days to just two seconds.
Once a use case has been established, it’s imperative to be clear on the objectives for the AI project upfront. This includes setting goals for what success looks like, which can be measured and reported against. Targets could include a metric to reduce payment failures, to reduce the timeframe for a specific process, or to increase revenues by a certain percentage.
Being clear about the scope of the project is paramount. Beginning with one team in one geography is a good way to start before attempting projects on a larger scale. Proving the use case works in a particular scenario can allow initial success to be quickly demonstrated. The scope and scale of the project can then be gradually expanded – with the business value measured at every stage. This approach also allows for ‘fast failure’ so that if something isn’t working, resources can be re-directed and the team can start again.
Focus on the data
Banks are in a great position to use the vast amount of transactional data available to them, which has largely remained untapped for decades. The combination of AI, machine learning and cloud-based platforms, means data can now be used in new and creative ways to create increased insight for customers and financial institutions alike.
But crucial to the success of any AI project is understanding what data is available and how it can be used. There’s a trade-off to be recognised in gaining access to more data and the amount of data that can be legitimately used. Privacy is the number one issue – it’s essential to be aware of different jurisdictional requirements, and most importantly to build customer trust.
A report on Big data, AI, machine learning and data protection from the UK’s Information Commissioner’s Office outlines a number of key considerations before embarking on a big data project. These include:
- Determining whether the big data analytics to be undertaken actually requires the processing of personal data. In some circumstances it might be appropriate to anonymise the data;
- Being transparent about the processing of personal data, providing meaningful privacy notices to customers throughout the process;
- Embedding a privacy impact assessment framework into big data processing activities to help identify privacy risks;
- Following ethical practise to help reinforce key data protection principles;
- Developing auditable machine learning algorithms with a view to explaining the rationale behind algorithmic decisions and checking for bias, discrimination and errors.
Banks must be in a position to provide an “AI Explanation” to customers and regulators – showing what data is being used, how it’s being processed, and how particular conclusions are reached and validated, ensuring there’s no bias or room for false conclusions to be drawn. For example, if a firm is using AI to provide investment recommendations, it has to be able to explain the rationale for reaching those conclusions. It needs to be transparent and open with customers.
The use of data in AI is a continual ‘work in progress’. It’s important to start with a small, well-defined data set and prove the value of a particular AI application – before scaling up to big data proportions.
Collaborate to succeed
Technology is evolving so quickly that no bank can expect to do everything themselves in-house. Banks that persist with that mindset will fail.
Instead it’s essential to collaborate with partners, to leverage and incorporate the very latest AI thinking, tools and techniques, and to exploit the power of the cloud and open APIs in enabling open collaboration and innovation.
Today smaller banks are more likely to partner on AI projects; while larger banks are often keeping most aspects of their projects in-house – particularly because of the data challenges. But adopting such an approach is likely to slow down innovation.
Instead, if banks can get the balance right: defining the right use case; understanding what data they have and how it can be used and; leveraging the right people and technology for every part of the process, then they’re on the right road to achieving success with AI.
It’s essential to start small and be ready to scale, and to remember that AI and data analysis is not like a regular product that’s built once and deployed. Data is dynamic and constantly changing and requires a continuous deployment process. This is a where an agile, collaborative approach comes into its own: those banks that open up and collaborate to innovate are well positioned to deliver on the full potential of AI and to reap the rewards over the long term.
Verdict deals analysis methodology
This analysis considers only announced and completed cross border deals from the GlobalData financial deals database and excludes all terminated and rumoured deals. Country and industry are defined according to the headquarters and dominant industry of the target firm. The term ‘acquisition’ refers to both completed deals and those in the bidding stage.
GlobalData tracks real-time data concerning all merger and acquisition, private equity/venture capital and asset transaction activity around the world from thousands of company websites and other reliable sources.
More in-depth reports and analysis on all reported deals are available for subscribers to GlobalData’s deals database.