In the labyrinth of news about AI, payments transformation and regulation, much of it can feel generic.
Payment providers and AI investment?
When it comes to payments, where are payment providers and banks directing their investment in AI, how is it supporting developments in payments specifically?
In short – can automation, machine learning help payments firms navigate large payments modernisation projects, and quite apart from this, should we be navigating a more serious discussion on energy resource and compute power?
A recent paper issued by RedCompass Labs asked some of these questions of 200 ‘senior’ payments bods at European and American banks.
Apparently, more than half of respondents (54%) are planning to use generative AI for payments modernisation, with 42% “actively considering it a possibility”. And only 4% currently having no plans to incorporate it altogether.
It would also seem that the more expertise an organisation tends to have, or perceives themselves to have, the more ready their enthusiasm and conviction around the implementation of generative AI and its potential benefits in payments modernisation.
Top AI Payments Priorities
Technical tasks (36%), impact analysis (35%) and strategic tasks (31%) are the top categories for the use of AI in payments systems. And as you can see these are fairly equal in their priority.
Technical tasks include things like software testing and code development; Impact analysis includes regulatory updates and system upgrades, and Strategic tasks include advising on new payment strategies and planning for regulatory compliance.
Certainly, there is a lot happening in terms of upcoming payments mandates and changes to be implemented. In terms of instant payments alone there is huge and unavoidable swell, alongside which there is the global migration to ISO 20022.
Cost Benefit Analysis
Another key challenge in building an AI strategy is how much it’s going to cost.
Particularly pertinent when a strategy evolves around cost-cutting, which most rudimentary AI applications are about.
Cost is usually assessed in terms of designing and implementing the system, or buying and integrating software.
However, with technology, and specifically AI, the cost of running can actually become too high due to the sheer energy required to power the models.
Payment technology providers are wise to this and some, such as Synechron, are promoting ‘edge computing’ to address this issue and create a more efficient approach in every way.
Edge computing, as it has come to be known, is where workloads are powered and processed at the user location as opposed to ‘being sent’ to a large data centre.
“The decentralised compute approach reduces the reliance on data centres, which are often energy-intensive. For instance, Google and Microsoft now consume more electricity than 100+ countries,” says Ryan Cox, co-Head of AI at the firm.
Besides the optimisation of resource allocation and reduced energy consumption, an added bonus is a reduced need for costly data transmission to centralised cloud systems, thereby lowering latency and bandwidth.
This allows real-time applications to operate more reliably, he says.
On trust
Being able to trust any model is a further great leap and one which impacts cost reduction.
The extent to which AI applications can self-administer or require human checks and balances is perhaps always going to be a question.
One of the RedCompass survey questions was the ratio of human involvement: AI deemed to be required to trust generative AI in a big project such as ISO 20022 or instant payments.
The answer was 49:51 in favour of AI. So, by no means independent and fully automated.
This is perhaps heartening for many out there, and by the same token offers perspective on the mammoth undertaking payments firms are straddling, and straddling with due care and caution.
Payments Cards & Mobile Opinion
To what extent edge computing could apply, and how, to different sizes of payments firms with huge modernisation prospects would have to be discovered but it begins to apply a broader, consequential lens to the establishment of new technologies, to consider the legacy and the footprint left behind.
The post Where are payment providers directing AI investment? appeared first on Payments Cards & Mobile.