Insurance companies and banks are tantalized by the capabilities of artificial intelligence (AI) but have been slow to adopt a technology that could revolutionize their strategies in pricing, claim handling, fraud detection, and business planning. Generating  added in both traditional and new business lines. 

According to a June 2020 study commissioned by the Economist Intelligence Unit, “The Road Ahead: Artificial Intelligence and the Future of Financial Services,” which surveyed 200 executives at banks and insurance companies:

While there is a strong degree of confidence in the benefits of AI, the reality is that the technology is not largely in use: more than half of respondents say AI is not incorporated into their business’s processes and offerings, with only 15% saying the technology is used extensively across the organization. However, the benefits that have already emerged combined with respondents’ plans to double down on AI investment in the short-term show this technology is slated to drive a massive wave of future growth for the financial services industry. 

The benefits are crystal clear. 

What company wouldn’t want to:

1. Retain customers based on AI-powered churn prediction 
2. Improve business results using price optimization 
3. Shorten products’ “time to market” using AI and automation 
4. Reduce losses due to fraudulent claims both in terms of operations (save money on attorneys fees) and claim payouts

But as much as traditional financial institutions believe in the promise of AI, the road to adoption has been blocked by challenges, both real and imagined. Let’s take a look at some of the factors that have prevented financial institutions from incorporating AI into their operations. 

Professional knowhow

Companies operating in the tech sector today were born in the era of analytics and naturally understand the advantages of AI.

AI is an integral part of their day to day operations and they employ data scientists who use the technology in their research and development units to stress test scenarios and play out future probabilities. In contrast, banks and insurance companies are populated by managers who reached professional maturity in the pre-analytics era. These talented managers know that AI is critical for their competitive futures, but do not understand how best to apply the technology or incorporate it into their system infrastructure. 

Instead, they turn to third party vendors who build them expensive AI systems that ultimately fail, for the following reasons: 

• The technology manager (CTO) who is responsible for the project lacks sufficient experience in the AI field. They are thus susceptible to highfalutin promises and overspending for inappropriate equipment.  
• There is a weak link between the technology department and the business units that will ultimately utilize the AI systems. Hence, the expensive AI solution is underutilized.  
• Board members fail to understand that investment in data science (DS) is a long term process. After a few months without ROI, they abandon the project and question the technology department’s priorities. 
• It is challenging to recruit a quality DS team and retain them long term 
• Knowledge silos – wherein the DS unit holds all the institutional knowledge – are nearly unavoidable 
• In many cases the AI team doesn’t have any stakeholder presence on the management team, and they are assigned tasks that could be completed by other employees. 
• In many cases the definition of AI and DS are not clear to relevant stakeholders. Therefore, there is confusion surrounding its capabilities and functions. 

The global consulting firm McKinsey & Company reports that 9 out of 10 attempts at integrating AI fail, and only 6 percent yield any economic value. 

A plan for incorporating AI 

Despite the challenges, some financial institutions have attempted to build data science departments internally or to design a road map to incorporate AI capabilities. But establishing an entirely new division is a laborious process that takes careful planning and the delineation of clear goals. 

First among these, is establishing a team that will combine different stakeholders in the process. For example, the relevant business unit, a technology professional, and a data scientist. 

Second is the clarification of expectations, including setting a schedule for AI-implementation, expected return on investment, fit to the product ecosystem, and the ability to implement the solution throughout the organization. . In this step, the owner must  convince other C-suite executives that AI is a long term investment that will require the careful budgeting of resources and the expectation of unforeseen expenses. 

However, if a financial services company can overcome these challenges, there are numerous areas inside and outside the organization , like those listed above, that AI could improve. 

In addition, data scientists are now imagining the possibilities of other, more advanced capabilities of AI, including: 

• Augmented reality: This would allow insurance companies to virtually inspect homes after a file is claimed or confirm a vehicle’s safety features before providing a quote. 
• Frictionless business: AI could help businesses learn consumer habits, make recommendations, and simplify how products are promoted. 
• Data veracity: Insurance companies could use more accurate data to make better business decisions. 

Whatever the reasons, as long as traditional insurance companies and banks struggle to adapt to the new era, they will fail to deliver  higher quality services. Moreover, they  will expose themselves to stronger competition from current and future ecosystem  players who will seek to disrupt the market by creating higher standards and better products.

   Financial services companies would do well to quickly build AI capabilities and join other industries in bringing  better services and information to their customers.