Established banks, startups and disruptors met in London this week to discuss the implications for introducing AI into finance
Data spectrums, bots, culture changes, regtech, shared data ontology… There’s jargon aplenty in the world of fintech (financial technology). This week the industry met at London’s Hurlingham Club where the second edition of the Intelligent Automation in Finance Conference sought to clarify exactly what automated technology, and its buzzwords, mean for banking. “I chaired this event last year,” says ISG’s head of UK, Ireland and Netherlands Barry Matthews. “We talked about the robots are coming – I don’t think that’s the case anymore. They’re already here and we’re now seeing the robots evolving and the robots expanding.”
The key technologies in this space are robot process automation (RPA) and artificial intelligence (AI). They allow faster capture and processing of data, be it from the market, from a customer or elsewhere. For example: such automation combined with the sharing of data via open banking, means, once you consent, your details can be accessed quickly to transfer accounts within days, without the need for laborious paper work or face to face meetings.
However, that doesn’t mean all technology is good technology.
“AI is bollocks,” UiPath’s chief evangelist [sic] Guy Kirkwood says he told a conference last year. “[But] that’s mainly because when most organisations go for AI, the C-suite will say ‘we want AI! We want AI!’ and they really have no idea about what that actually means…”. “They also want blockchain,” adds BNP Paribas’ chief digital officer Thierry Derungs.
The sentiment has resonance: Last year Jon Theuerkauf, former managing director & group head of performance excellence at BNY Mellon, told the world: “Forget AI, I don’t even know what it means – why are we jumping on it if we haven’t done the basics. Currently BNY is ‘consciously incompetent’ – we are really proud of that. We know what we don’t know.”
The same could be said of broad sections of society and the media, who, not without reason, have equated the rise of the robots to a rise in unemployment. But while RPA and AI are often sold to large organisations with the tag line ‘reduce your HR overheads’ (bots can work 24/7 at much lower costs) the rise of the machines is not in conjunction with job losses says Kirkwood: “Headcount is unaffected, despite that being the initial raison d’etre for RPA.”
The idea of destructive fintechs using robots to shake up the sector without providing synergies is also erroneous. Derungs credits the “rise of the ecosystems”, saying that as CDO of one of the world’s largest banks he has no desire to become an agile digital platform overnight. Rather, he wants to use the best of the new fintech providers to get the most efficient and tailored service for clients, which will also hopefully help him negate the 5,000 monthly rule changes and the data legacy issues that he currently has to deal with. In fact, because startups don’t have those legacy issues, they have the liberty to provide an array of products, says Megan Caywood, chief platform officer of Starling Bank: “it’s like Borough Market!”
The quality of the robo-work
However, if the bots are to free us from financial monotony, rather than condemn us all to perennial problem fixing, the issue of appropriate application remains. This is most prominently shown by the deficiencies of data quality and data processing. One speaker from a very prominent banking body, highlighted how banks will ask for large data sets but then not be able to use the information properly, getting bogged down by erroneous entries and overlong lists and thereby creating further frictions within their internal departments and with organisations and scrutinising regulatory bodies worldwide.
The solution should be in standardisation. Much in the same way shipping containers and barcodes have defined the global transaction of goods, so the codes known as legal entity identifiers (LEIs) could be used to facilitate the transaction of services and monies – creating an informed process that only seeks relevant data. The UK should seize this opportunity to roll out LEI numbers worldwide they said: “If the UK were to do that it would put us ahead, but the government needs to back it more in my view.”
Standardisation will no doubt follow on from familiarisation. Away from the main stage, a number of workshops are underway detailing the companies bringing automation to the masses. One automation product allows clients to build bots: bits of software that follow simple commands. Those trialling it put together 21 steps that take the live data from the FTSE 100 and scrubs it to leave just those companies which were up 2% or more. Can you build bots to create other bots? Not quite it seems, but you can build in a degree of knowledge-based selection once they are given enough variables, “but it has no brain power – it must be rules based,” says the lead instructor.
The company showcasing this, Automation Anywhere, claims to have reached a $1.8 billion valuation in ten years and to have created a digital ‘workforce’ of 600,000 working 24/7. Their product even has a ‘ROI’ (return on investment) button you can click at any point to reveal how much you are ‘saving’ and ‘earning’ via bot automation (how this is calculated is not clear). In a similar vein Kirkwood’s UiPath says it grew by 850% last year and that it now has 70,000 users worldwide.
Such numbers are unsurprising to one AI and robotics specialist, who points to the concept of a fourth industrial revolution and the possibility of a singularity event which would “redesign the world”. But the challenges remain: automated use of data only works if it is practical and is only socially palatable if it is ethical. The Royal Society of Arts recently released a report on perceptions of AI and its impact on work and society more broadly. “The point is that we need to make deliberate and informed choices in how we design, develop and deploy this technology,” said Benedict Dellot, head of the RSA’s Future Work Centre. “Nothing can be left to chance.”
Sign up to get the latest Glint news
Receive the GLINT newsletter with the most popular content, platform updates and software guides.