I was talking with a senior banker, who told me that he was in charge of the Artificial Intelligence program in the bank. I was impressed as he is part of the executive leadership team of the bank and not the CIO. We have invested widely across the board in AI, he told me, and achieved great results. Sure, I hear that a lot. How many people have you laid off as a result? I asked. I always like numbers, and most bank AI developments appear focused upon cost savings. Less than you would think, he replied, in fact no one.
I was surprised and wondered what the point of AI is, if it’s not to automate jobs. He replied that this is not the intent of AI at all. The aim is more efficiency, more processing, more capabilities and generally more intelligence. It is to augment the human process, not replace it.
I asked him what happened with people whose jobs are automated then, and he said they get retrained. After all, the bank has invested heavily in educating these people in the bank’s ways and culture, in giving these people real-world experience in front-, middle- and back- office banking and that they have a great deal of knowledge that the bank did not want to lose. That is why the bank focused heavily on retaining these people and their skills and experience.
Ha! So, you’ve retrained a branch teller to code in Python? I retorted skeptically. Funnily enough, we have, he pushed back.
As the conversation continued, it became quite clear that the company had invested heavily in moving people to other roles as software automated their jobs. As branches shut, customer service support increased; as accounting people were removed, developers increased; as compliance people reduced, trainers and designers of AI code increased.
This is a move that I’ve heard of in a few banks (not usually in American ones) and is all about the shift of human focus to using machines to do jobs better rather than focusing upon machines to replace jobs. My favourite example is in wealth management where UBS has automated client instructions.
The system scans for emails sent by clients detailing how they want to divide large block trades up between funds. It then processes these and executes the transfers. It saves time by doing a task that would normally take a person about 45 minutes in only about two minutes, while freeing investment bankers up for other tasks, such as calling clients. Financial Times, July 2017
This means that wealth managers, who are highly client oriented intelligent people, can now spend many more hours dealing with their high net worth client’s real needs – the psychological and emotional – rather than just administering their needs – the mathematical accounting and administration.
This is what augmented banking is all about: getting humans to do far more work on the heart and less with the head. It’s not something that comes naturally, but if banks are to redeploy people from jobs that are automated to jobs that are augmented, then they need to focus upon what those future jobs will be.
According to MIT, those jobs will be trainers, explainers and maintainers of machines. Trainers train the machines to do their jobs; explainers explain what jobs the machines are doing internally and externally; and maintainers sort out the machines when they go wrong.
These roles are clearly demonstrated as being necessary when you look at the Amazon AI engine that rejected all female applicants, or the extreme hate fed into the Microsoft twitbot. You need to fine tune machines on a continuous basis if what they are learning is wrong.
Overall, I can see AI being something that all banks will buy into in some form or other, but their objectives are often very different:
- some are doing it to reduce costs, which means getting rid of people;
- whilst others are doing it reduce risk, which means applying AI to AML and cybersecurity;
- whilst others are doing it to increase the depth of their customer relationships, which means customer-intelligent marketing and servicing;
- whilst others still are trying to do all of these things, and maybe that’s too much to take onboard in one go, especially as customer-intelligent services demand a rationalisation of the data architecture.
Generally, there are things that can be done with AI that are really useful, however. As Deloitte found, when they surveyed 200 financial professionals, the key areas are:
Embed AI in strategic plans: Integrating artificial intelligence (AI) into an organization’s strategic objectives has helped many frontrunners develop an enterprise-wide strategy for AI that various business segments can follow. The greater strategic importance accorded to AI is also leading to a higher level of investment by these leaders.
Apply AI to revenue and customer engagement opportunities: Most frontrunners have started exploring the use of AI for various revenue enhancements and client experience initiatives and have applied metrics to track their progress.
Utilize multiple options for acquiring AI: Frontrunners seem open to employing multiple approaches for acquiring and developing AI applications. This strategy is helping them accelerate the adoption of AI initiatives via access to a wider pool of talent and technology solutions.
In conclusion, the time is right for banks to get into AI, as the technology structures are right to leverage AI. The availability of superfast processors and huge memory core in today’s system allow for near unlimited bandwidth and processing power for a fraction of the cost of legacy platforms, even those from the last decade, let alone those from the last century. These changes should force banks to re-architect, re-engineer and re-think their application architectures.
This is the second of a series of blogs I’ll be posting before this year’s SIBOS in London, September 22-26, where I will be working with HPE.