When The Bank’s Customers Are Replaced By Custobots

I was interested to see that Gartner’s
IT
top ten strategic technology trends for 2024 include the arrival of machine customers (what they call ‘custobots’ but what I call “economic avatars“, a precise and descriptive label that the virtual reality visionary Jaron Lanier introduced some years ago). These are the smart, non-human actors that can autonomously negotiate and purchase goods and services in exchange for payment. Gartner think that in five years there will be 15 billion connected products with the potential to behave as customers, with billions more to follow in the coming years.

APIs Not Brands

This all sounds a bit hyperbolic, but the fact is I think that Gartner might well be right. They suggest that organisations pay strategic attention to the switch from human to machine customers, and I have long thought that this should be a priority for financial services organisations in particular.

Since I don’t know anything about marketing, I am fascinated to see how banks will adjust to acquiring robot customers who do not care about the bank’s logo or TV ads or which sports team it sponsors. So when my smart wallet uses open banking data and decides that I need to open a savings account or get a loan or refinance my mortgage, how will my finance bot decide which provider to use? After all, I don’t really want to be in the loop because I’ve got better things to do.

As I am sure this is true for most people, most of the time, in the not-too-distant future, our financial decisions, transactions and analysis will be performed by bots operating under relevant duty of care legislation with the co-ordinated goal of delivering financial health. I don’t think this is a bad thing, as I sure the even a rudimentary finance bot can do better than I can when it comes to managing money.

Given that I intend to hand over responsibility to my finance bot. then, how will that bot go about choosing which accounts to open, which services to use and which oracles to listen to? I imagine that it will use a combination of reputation and relevant other data (eg, economic forecasts) to work out which account is right now and then I’ll just click OK and hey it’s all done. The reputational calculus will of course involve fees and rates but instead of using the Victorian substitute of brand for actual data, my bot will look at API functionality, open finance interface availability, service uptime and so on.

(I say “Victorian” substitute because the the first U.S. registered trademark, filed for paints, was issued on 23rd October 1870 and the first U.K. registered trademark was issued on 1st January 1876 for the red triangle of the Bass Brewery.)

This means that banks, financial organisations in general and, of course, fintechs will be selling their products to machines, not to people. Well, their machines will be selling things to customers’ machines. Now, people have tried having AI make financial decisions in the past and truth be told it has;’t worked out too well. ETF Managers Group’s AIEQ
AIEQ
, launched in 2017, uses IBM’s
IBM
Watson AI platform to analyze millions of data points from news, social media, industry, and analyst reports, plus financial statements on over 6,000 U.S. companies, and technical, macro, and market data, among others so it provides a useful case study.

How did it do? Well, ovver the last five years, it returned 4.9%—trailing the 11.78% five-year return of Vanguard’s benchmark S&P 500 index fund and for another comparison, two large actively-managed funds (the American Funds Growth Fund of America, at 9.81%, and Fidelity’s Contrafund, at 11.04%). That doesn’t sound particulary successful to an amateur observer such myself.

Yet many people are billish because they think that historical robo-advising was essentially jazzed-up machine learning. The custobots that Gartner is talking about will use AI and deep-learning algorithms to deliver something very different and they will require very different services from financial institutions. As an obvious example, companies may have to provide specific APIs to support the needs of bots rather than people since bots can search through more data, access more sources and process more transactions than any person might do. Levels of service acceptable to a customer may be competely unacceptable to a custobot.

Those custobots may not be so far away. Commonwealth Bank of Australia (CBA) is already examining how it can use generative artificial intelligence to create faux consumers who can test new products. Dan Jermyn their chief decision scientist, said the technology can enable machines to perform experiments on products to see how popular the products may be. They are drawing on simulated experiences of daily life to emulate behaviors to improve qualitative and quantitative understanding of how customers might respond to changing contexts, everyday financial challenges and new products.

It’s a pretty fun way of making a SimBank, but surely it is not much of a step to turn those customer bots into customers’ bots. if you see what I mean.

Machine Money

Bot-to-bot transactions might take us into some new territory when it comes to money, by the way. The German Banking Industry Committee (GBIC) is the voice of the leading German banking-sector associations. I had always assumed it to be a very conservative organisation, so I was fascinated to see that they called for some form of tokenised private-sector money be developed to meet corporate demand arising from Industry 4.0 and the Internet of Things. They envisage that such a money would facilitate transactions based on “smart” (ie, automated) contracts and thus increase process efficiency.

(The idea of giving the machines the cash they need to spend may sound pretty radical but Commerzbank was already trialling blockchain-based machine-to-machine payments between electric charging points and Daimler Trucks back in 2019.)

Gartner are surely right to point towards this bot-on-bot action as a focus for financial services moving forward and those transactions will undoubtedly lead to amazing changes in financial services. But even Gartner may be underestimating the changes they will bring to money itself.