Stronger together

CitiGroup’s new Payment Outlier Detection solution is just one example of how a more inclusive approach to innovation can produce positive results says Gulru Atak Gundem from the bank’s Dublin lab

Artificial intelligence (AI) and its close cousin machine learning (ML) have myriad potential applications, perhaps the most prominent of which is the chatbots we’re all getting more and more used to interacting with daily, in lieu of real humans.

Many of the use cases have thus far been driven by an increasingly demanding generation of customers looking for a more seamless service experience. Global bank Citigroup, though, is also mining the technology for corporate functions, including treasury risk management.

Its Dublin-based Innovation Lab was behind the launch of Citi’s Payment Outlier Detection solution this summer, for accounts payable in 90 countries worldwide. Created by the bank’s treasury and trade solutions business with the support of AI and real-time risk management company Feedzai, the new offering uses advanced analytics,
AI and ML to help proactively identify anomalous payments that don’t reflect clients’ previous patterns. When it spots something that doesn’t seem to fit, it flags it in real time to a human employee, who can then make the decision whether or not to let it through. It could be an irregular payment to a periodic supplier, or a new type of payment which is part of a fresh initiative, in which case staff can authorise it with minimal delay.

The solution is an analytics tool that helps clients identify risks in payments, but it is not meant to replace client controls, and is not guaranteed to identify all risky payments. The new solution is timely, given the dramatic lift in global payment volumes, fuelled by factors including increasing digitisation and automation, as open banking takes hold. All of this has increased the stakes for sophisticated cyber attackers, the scale of whose relentless efforts mean banks must be more vigilant than ever. This, coupled with the rapid growth of instant payments and increasing client expectations around processing speeds, is also challenging corporate treasury

functions to look at fresh approaches.The new service’s launch followed an extensive global pilot with 20 clients. The Payment Outlier Detection system is designed to require minimal integration effort and it uses advanced statistical machine learning algorithms instead of legacy rules-based logic to analyse payment patterns. This enables the system to automatically adjust in response to changing payment patterns as businesses evolve, expand and globalise, says Citi.

Gulru Atak Gundem, who runs CitiGroup’s Innovation Lab’s treasury solutions team, explained that this is just the latest example of the bank’s historic commitment to innovation.

“It’s something we’re really proud of, because we opened our first innovation lab in Dublin in 2009, immediately after the global credit crisis. I think we owe a debt to the senior management at that time who were thinking ahead of the game,” says Gundem. “The industry, including Citi, lost some of our innovation muscles when we went through our own survival times and I think having the innovation lab actually helped us to create a content environment where people could continue innovation practice and share its practice and principles with the rest of the organisation, without risking the bank’s reputation.”

With that in mind, how does the bank now go about bringing hothouse ideas like this latest project, to its day-to-day activities?

“We’ve gone through lots of versions of our innovation labs. We’re probably now on version 4.0. Because what we noticed, over time, was that when you have an idea and the labs work on it, then try to give it away, it becomes almost like adopting someone else’s baby. The receiving party doesn’t really want to adopt it and the innovation lab doesn’t want to give it up. So we’ve found, over time, that it’s helpful to start working together on the ideas very early on, with all the relevant constituencies, both in the bank and outside of it.

“And that’s how, two years ago, we conceived Citi Payment Outlier Detection. We are now enhancing it with one of our partners, Feedzai, and it’s leveraging machine learning and looking into the payments behaviour of corporate clients. But we started out working with our payments product partners very early on, our data scientists, the product managers and risk and control functions, because, ultimately, we needed  their approval as well. That turned it into a success, rather than the innovation lab ideating and trying to sell its own ideas to others.”

To mark the 10th anniversary of the bank’s Innovation Lab in Dublin, Citi has launched a FinTech Challenge for transaction banking. It is ‘looking for fintech innovators with
ideas that have the power to transform how we work and drive client value’. It is inviting fintech innovators to find answers to problem statements it has set out, including how it can reimagine the client experience and make data work smarter.

Shortlisted fintechs will be invited to showcase their solutions to executive sponsors at the FinTech Challenge Day on 5 November 2019, at the Dublin Innovation Lab. They will also have the chance to pilot their solutions with leading Citi product teams in 2020.

Citi operates in more than 160 countries and jurisdictions, with around 200 million customer accounts. Its Citi Treasury and Trade Solutions (TTS) arm provides integrated cash management and trade finance services to multinational businesses, financial institutions and public sector organisations. It boasts the ‘largest proprietary network with banking licenses in over 90 countries and globally integrated technology platforms’ and focusses on digitally-enabled treasury, trade and liquidity management solutions.

Now that it’s brought its payment outliers technology to market, what’s coming next, in terms of innovations that are going to have similar impact on its corporate audience?

“I think we’ll see more and more solutions which leverage machine learning and natural language processing. In some cases, this is making things better for us internally, which will eventually turn into a better client experience. There are also more things that we plan to give to our clients, like leveraging cognitive contracting, using natural language processing to better negotiate contracts,” she continues.

“We also have a centre of excellence for blockchain in Dublin, as many of our clients, and especially their treasury departments, are interested in leveraging blockchain or distributed ledger technology (DLT) for certain pain points they’re experiencing, like inter-company settlements. We are also hearing from big tech, around things like issuing their own stablecoins, and are looking into those areas as well.

“But I think our main difference is our focus on connected everything. If you look at the connected economy, connected mobility, the on-demand economy and so on, the business models of our clients are being disrupted and transformed heavily, probably more than ever. That’s why we are speaking to lots of our clients, our corporate treasurers, to see how we can join forces and build our own ecosystems.

“Because I really think that it will not only be companies in the same industry that come together, but there will be more and more partnerships and different business models where companies from different industries come together and build their ecosystem. That’s how we are looking at the future.

“My personal wish is that if the world is more inclusive, from a financial services point of view, it can achieve more. I think we will see it, because consumer behaviour and expectations are changing and will ultimately influence corporate behaviour. People who are asking for things in their personal life are asking more from their corporate experience as well, which will eventually democratise things.

“Look at open source software. In the old days, we used to negotiate heavily when we were buying software, as most companies would want the IP to remain theirs. But now, most of what the big technology companies develop is open source, and look at how that has transformed the pace of innovation. Most of Generation Z, and I’m not even talking about Millennials, won’t be afraid to share their data.”