It’s a good question. Virtually every AI expert foresees a large shift in financial services from humans to automated systems, and some experts even estimate this shift to replace half the nation’s banking jobs within the next decade.

The combination of cheap and powerful processing power, along with the millions of data points now captured through all our online activity and digital transactions have created an explosion in the potential for financial AI. The industry, whose services are prime for technological replacement due to their data-driven nature, could see a shift from manual labor that would save companies some $1 trillion in labor costs.

In short, artificial intelligence and machine learning are definitely here to stay, in a big way. However the degree to which automation and machine learning differs based on the sector in finance, and in many cases AI are actually augmenting existing human roles, rather than fully replacing them. To understand we take a look at a few of the most promising current and future use cases in the financial sector.

Portfolio Management and Algorithmic Trading

Known as “robo advisors”, and “automated trading systems” respectively, these traditional investment sectors have been some of the early adopters to machine learning.

For portfolio management, companies like Betterment and Wealthfront have gained popularity for providing personal portfolio management, for fraction of the fees associated with a physical advisor. Users set targets and risk preferences (like traditional portfolios), and the AI portfolio managers use data make suggestions to individuals, or automate the process. While the $283 billion managed by robo advisors is a fraction of the global $74 trillion in wealth management, it represents a 45% growth from last year, with 23% of investor believing the technology “will lead to rapid changes in the investment landscape”.

Algorithmic trading, while dealing with many of the same instruments as portfolio managers, is often used to augment humans. High volume data parsing is removing the importance of traditional news sources and sensational headlines, and allowing traders real time analysis of market movement on a particular stock. Algorithmic trading is also increasingly used in what’s known as high frequency trading, where an algorithm makes thousands of automated trades throughout a day based on set criteria. However due to the sophistication of these algorithms and the large amount of data needed, they are primarily used by hedge-funds, with little personal application.

Insurance Underwriting

One of the purest sensical uses for AI, algorithms that can quickly process literally millions of factors – including age, car accident history, loan repayment history, purchasing behavior, etc – present enormous potential to change the current insurance underwriting system.

All of the top four insurance companies – Liberty Mutual, Progressive, Allstate and State Farm – are not only currently using AI, but have held competitions to create algorithms that reduce risk as possible. The initial promising results point to increased efficiency and $400 billion in savings from replacing 865,000 insurance workers.

Customer Service/Financial Knowledge Assistance

This sector, more colloquially known as “chatbots” is actually one of the most targeted areas for improvement by AI and finance experts. It was recently ranked as the #1 potential use case by both top insurance executives and AI experts. Salesforce alone believes deploying AI for customer relationship management will increase global business revenues by $1.1 trillion by 2021.

There are two main problems with the current customer service model: the cost of workers, and time for available agents (people hanging up from being on hold) means reduced business opportunities. The latter problem is compounded by the learning time associated with ramping up associates to the new products offered in an evolving landscape.

Machine learning looks to solve for both of these. Automated chats, emails and search queries reduce the need for humans, while also learning new features instantly, and creating streamlined learnings across all bots to be able to produce better answers. However, human-like language processing has proven to be one of the most challenging parts of AI, and one of the largest current uses for AI in customer service via augmentation, where contextual information is instantly provided for human customer service agents.


Bottom line on automation in finance

While we’ve touched on a few of the current applications, machine learning is affecting every virtually every area of finance, from high end investment management and stock trading, to cost saving replacements for insurance underwriters and customer service. As a high volume, profitable, risk minimizing, data-driven industry, the financial sector is arguably the most perfect sector to be revolutionized by machine learning.

While no one can predict exactly how the human workforce will be replaced by or be supplemented by AI, financial and technological experts overwhelmingly agree that these machines are here to stay.

The downside is job retraining for those who are 20 years into a traditional finance career could be difficult. The good news is almost all of the technology we’ve seen has largely come in the form of augmentation to make humans more efficient, rather than replace them. Similar to how computers in general have increased efficiency, rather than lead to lower employment.

Ultimately, the changes are expected to accelerate between 2025 and 2030, so the next decade will likely see the largest phasing in of financial automation.

See it for yourself

Try our fully functioning demo account and see if Emburse is the right fit for you.