AI machine learning has far-reaching benefits for the finserv industry, from operational efficiency to the fight against financial crime. A recent Twitter chat hosted by Refinitiv answered some of the biggest questions about using AI in financial services.
- AI in financial services is chiefly being used to improve efficiency by performing mundane repetitive, manual tasks, as well as to assist with fraud detection.
- Experts in a recent Twitter chat hosted by Refinitiv agree that machine learning is gaining traction, but many believe there is still a big gap between the reality and the hype.
- To find out more about using AI in financial services, download the Refinitiv report: Smarter Humans. Smarter Machines.
How is artificial intelligence (AI) impacting the financial services industry? The responses from our recent Twitter chat reveal how AI machine learning is being used for both simple, repetitive tasks as well as in more sophisticated — sometimes unexpected — ways.
For some, the largest benefit of AI machine learning is that it forces financial services companies to reinvent themselves and focus more on customer experience. “They either disrupt with AI or are disrupted by AI”, said Helen Yu, founder and CEO of Tigon Advisory.
Other participants in the Twitter chat included Geoff Horrell, Refinitiv Director of Applied Innovation, London Lab, and Alex Jiménez, a U.S.-based chief strategy officer focused on digital transformation, innovation and digital banking.
Find out what they had to say and then join the conversation.
Q1: How is AI being used in financial services?
The discussion found there are currently two main use cases. Firstly, AI is being used for mundane, repetitive and manual tasks, which can help to improve a firm’s efficiency. Yet, there are questions about the effectiveness of this.
The big question is whether #machinelearning with its probability based results can fit into traditional rules based RPA processes.
— Geoffrey Horrell (@GeoffHorrell) August 27, 2019
The other trend noted by many of our experts was fraud detection.
Fraud prevention is by far the most widely adopted use. It only makes sense as simple algorithms have been used in this area for a long time.#finserv #fintech #banking #MLReadyData
— Alex Jiménez (@RAlexJimenez) August 27, 2019
Fraud detection, for sure. But #AI/#ML is growing in other areas: #investment management, #workflow optimization, #risk and #credit assessment. #CloudReadyData #MLReadyData
— Jeff Marsden (@Jeff_Marsden) August 27, 2019
While many financial institutions would be pleased to hear this, we were reminded that AI is also already used by criminals as well!
At the moment only a small fraction of financial crime is detected. I expect the volume of fraud and crime detected will increase as techniques improve. At the same time, criminals will also become more sophisticated – they are already using #AI!
— Geoffrey Horrell (@GeoffHorrell) August 27, 2019
Q2: What is the largest benefit of AI/Machine Learning?
The use of AI for monotonous tasks was mentioned again.
There will be a huge wave of outsourcing of monotonous, tedious and highly manual efforts from over-worked humans to some kind of AI/ML algorithms. Insurance claims processing, trading compliance and client sentiment analysis are all high on the list.
— Craig Iskowitz (@craigiskowitz) August 27, 2019
But largely the focus was on using AI to help finserv companies to complete tasks quickly, usually those requiring a lot of time due to the amount of data available to businesses.
There has been a growth and democratisation of the software tools to handle data, which has allowed FIs to capture more data. Now the promise of ML is to automate the creation of insight from that data.
— Geoffrey Horrell (@GeoffHorrell) August 27, 2019
@yuhelenyu, however, took a different stance.
Q3: What is the biggest drawback of having bad data in machine learning in financial services?
Our experts suggested that while bad data is a problem for machine learning, there are other issues like black-box AI and accessible data, which are also of concern.
Regardless if it is #MachineLearning #AI or reporting, getting to the data and having it accessible and easily assessable is probably a bigger issue. Been my biggest challenge through the years. @xceptor is a good data management tool to help with bad data
— Brad Nelson (@bradnelsonops) August 27, 2019
The problem isn’t “bad data,” it’s black box #AI #ML. #ExplainableAI is a must for #finserv. I believe regulators will have to make this a requirement. We already see examples in industry of unintended results of biased ML#finserv #fintech #banking #MLReadyData
— Alex Jiménez (@RAlexJimenez) August 27, 2019
Q4: What impact has AI/Machine Learning had on talent and technology in your business?
The answer here was clear, with all experts agreeing that the use of AI has forced the talent market to become more competitive, with upskilling of employees being key.
Forcing employees to upskill. Some are proactive, others are in denial. #futureofwork
— Brad Nelson (@bradnelsonops) August 27, 2019
Q5: How far is reality from the hype in machine learning?
While all of our experts agreed that machine learning is gaining more traction each day, many believed we weren’t there yet when it comes to real AI.
The question is how well those of working in this area can communicate the concepts clearly without being overtaken by the hyperbole.
— Geoffrey Horrell (@GeoffHorrell) 27 August 2019
The gap is vast. I’ve seen FIs tout capabilities they haven’t yet tested never mind actually roll out. I’m very disappointed by the real marketing uses out in the wild. #finserv #fintech #banking #MLReadyData #AI #ML
— Alex Jiménez (@RAlexJimenez) 27 August 2019
Q6: How is your business keeping pace with new AI developments?
Jay Palter said it was the same for any fast-paced tech sector, the key is to attend events, read as much as you can and follow key influencers in the area.
Keeping pace with developments in #AI & #ML is the same as other fast-paced #tech sectors – attend key events, read as much as you can, follow key influencers and use them as filters. #fintech #insurtech #wealthtech
— Jay Palter (@jaypalter) August 27, 2019
Geoff Horrell also shared how Refinitiv is monitoring and publishing at tech conferences, but more importantly listening to our customers with an open mind, proposing that this may lead to you finding new ways to use existing technology and data.
Q7: Is organizational transformation essential for the adoption of AI across the digital enterprise?
This question got a 50:50 response.
Across the enterprise? Yes. But we are early in the point solution phase currently. Organizations will be transformed by #AI/#ML capabilities and begin to look very different in the future. #CloudReadyData #MLReadyData
— Jeff Marsden (@Jeff_Marsden) August 27, 2019
I’m going to negative on this one. I don’t believe there needs to be “adoption” of #AI. It will be slowly integrated into new software all across an enterprise’s existing systems. We will all be using #AI-driven applications before we realize it.
— Craig Iskowitz (@craigiskowitz) August 27, 2019
Q8: How are your customers implementing and using AI/Machine Learning?
My customers, #banking customers, aren’t out there thinking about #AI #ML. They are looking for convenience, value, better experiences, and sometimes connections. #finserv #fintech #MLReadyData
— Alex Jiménez (@RAlexJimenez) August 27, 2019
Working for a big bank, it is all about controls, operational efficiency and customer experience. Big drivers are reducing manual repetitive tasks (cost/controls) as well as driving the proactive aspect for customers to drive revenue but 80/20 is still cost/controls aspect
— Brad Nelson (@bradnelsonops) 27 August 2019
To learn more about the use of AI in financial services and how it is evolving, download our report: Smarter Humans. Smarter Machines. and let us know your thoughts.
