Looks like a number of movies these days are predicting the doom of the human race and the rise of the machines. Will machines take over the planet? Probably not. Will machines become your new colleagues at work? Maybe. Will Artificial Intelligence (AI)
become a core part of your IT systems and digital fabric? Absolutely.
During I/O 2017, the recent annual Google event, Google CEO Sundar Pichai talked about how the company is rethinking all of its services and products and considering a move from a “digital first” to an “AI first” world. Even at a time when we are seeing
significant strides across multiple facets of technology, it is still incredible to see major advancements over the last few years in data-driven technologies and algorithms.
At the same Google conference it was revealed how computer vision algorithms have progressed over the years. Just over a period of seven years from 2010,
algorithms today using machine learning in the Google Lens tool can ‘see’ better than what we humans perceive.
Fuelling this rapid advancement in machine learning are two key factors: a) an explosion and availability of data and b) the ability to get cost-effective compute power that can run powerful algorithms inexpensively and process massive amount of data.
Financial services are also at the cusp of AI-fuelled disruption as some key macro factors are coming into play : a) digital and cloud are becoming more accepted at financial institutions b) Financial firms have a massive amount of historical customer, market
and third-party data enabling them to gain greater insight c) With the possibility of a reduced regulatory burden, there is the potential to redirect dollars to focus more on building a competitive advantage through digital capabilities and d) AI use cases,
which provide benefits to customers and processes within banks, are showcasing the value of AI across the business.
What is AI?
There is no single definition of AI and people regularly use it as shorthand to talk about everything from building Robotic Process Automation (RPA) tools, to chatbots, to neural networks to deep learning. A commonly accepted definition comes from a professor
in 1959 who sums up AI simplistically: “Machine learning is the field of study which gives computers the ability to learn without being explicitly programmed.” This still hold true today.
A number of folks also refer to AI as simple business rule engines, which is incorrect. If you listen to Andrew Ng, the AI guru at Stanford, he shares that AI in itself is not a new concept. Neural networks have been around since the 70s, rose to prominence
in the 80s and then lost their shine. The resurgence in interest today is due to vast amounts of structured and unstructured data that, when processed together could provide new patterns and insights, in tandem with the availability of cheaper compute power
to cost-effectively process at vast scale.
AI in Financial Services
When we hear the term AI, we usually think of technology companies such as Google, Apple, Microsoft, IBM and start-ups working on deep learning problems or building tools and platforms. Yet, financial services companies have been working on creating real-life
AI use cases and are exploring opportunities to offer better cost service and save dollars.
For example, it is interesting to see how
UBS is using machine learning to develop new strategies for trading volatility on behalf of its clients. The technology scans a vast amount of trading data and creates a strategy based on learning from market patterns. This has to be then approved by human
There is also
JP Morgan’s COIN program, which parses financial deals. Traditionally, this process took legal teams thousands of hours to complete. This automation program looks at interpreting commercial loan agreements and has helped JPMorgan cut loan-servicing mistakes,
most of which stemmed from human errors in interpreting 12,000 new wholesale contracts per year.
AI at Scale
The challenge financial services firms face as they scale AI is more how to manage the legacy overhead and multitude of data sources within the enterprise with no common threads to tie them together. As companies build their data and digital organizations
and get a better handle on their data, AI is poised to deliver deep analytical insights.
A use case we see in customer engagement involves going from the unification of data to the personalization of customer experiences and ultimately to generating insights from customer data.
Three things are required for enterprises to leverage AI at scale:
- Confluence of expertise – How do you collectively assemble teams across IT, business, AI, data inside and outside the company to build the data set?
- Platform thinking – How do you create a toolset or product model with the customer at the centre of the platform for delivery?
- Conception pipeline – How do you operationalize the use cases, build the ROI and measure your value leveraging AI?
The real challenge of moving to AI at scale still seems to be talent scarcity. Talking to a client recently, they referred to a demand and supply issue. As a firm, they were ready to commit to developing AI solutions but the lack of available talent was
The concept might be decades old but the raw expertise required, combined with industry expertise needed, is a very small subset of talent available today. This is why we see fintech companies of different shapes and sizes collaborating together and incumbent
banks building bridges with the academic world to establish that talent pipeline.
Moving toward an AI first model within enterprises will require building the right framework to leverage the technology and securing the right talent to bring the vision to life.
As a fan of 2001: A Space Odyssey, I might not have a HAL in my life, but I do see a multitude of Amazon Alexa’s, IBM Watson’s and IPSoft Amelia’s learning from us humans every day. For us, this is just the beginning. We are seeing AI at scale becoming
an integral part of our personal lives, and increasingly we will see it as part of the day-to-day financial services operations.