FinTechAutomationSuccessful tech adoption requires give and take

Successful tech adoption requires give and take

Gert Sylvest, Head of Frontiers, the R&D arm of supply chain and payments company Tradeshift, discusses the evolving payment technology landscape and why successful use of technology requires plenty of give and take from treasurers.

How has the payment technology landscape evolved since you launched Tradeshift back in 2010?

“In 2010, when Tradeshift was founded, it was very clear that the market was dominated by several different business models, most of which revolved around closed networks of suppliers and buyers, allied closely to logistics providers. Typically, the focus was on digitizing high volumes of documents between, say, the five biggest suppliers. That meant, overall, the digitization of payments by volume was low. Perhaps only 3% of transactions were digital. That was despite plenty of effort being put in by various tech companies and banks to digitize supply chain relationships and build investment in technology, so that we could develop certain standards.

“That meant many businesses we’re missing out on some big benefits. Master Data is a good example. When we started go into large enterprises, selling AP automation and digital invoicing, we were actually quite shocked at how poor the data quality was. Typically, 80% of suppliers would have serious question marks around their data. Companies were also running up to 100 different control processes.

“Since then, we’ve seen the movement away from the typical isolated management of relationships. Instead, people are actually asking what happens when you build networks. Rather than working in isolation people are increasingly managing information for the benefit of multiple customers in a case of recording setup information once, rather than on multiple occasions. This is supported by notifications whenever people upload, amend or add information, which can then be subjected to whatever quality control is in place. Onboarding becomes much easier, removing one of the barriers to better buying, including from SMEs.”

What incentive is there for suppliers to join a digital payments revolution?

“It all comes down to the value proposition. Instead of asking suppliers to supply information, and the relevant information that you need, the conversation shifts to accessing early payments, which is clearly attractive for the supplier and builds the relationship between the two organizations. Ultimately, it’s necessary to incentivize people to serve up the relevant information.

“For me, the big change that we’re seeing today is the combination of technology and business models, which is really moving the payment environment forward. When we started in 2010, most business models were buyer-centric. Typically, suppliers were holding the cost on a cost per transaction basis. At the same time, the cost of digitization was high, so that was clearly a barrier for people to participate. They had to go through the digitization process, which involves making changes, which comes with a cost.

“As organisations have embraced technology and the payment environment has evolved, we’ve seen the development of two-sided market models that are much more collaborative. If I was a supplier I would constantly be asking ‘what’s in it for me?’ Whether it was access to faster payments or financing insurance, and so on. That leads buyers to open up too.”

In 2019, what are the technologies driving developments in payment technology?

“From a payments perspective, AI and machine learning will continue to disrupt the space. We’ll see more and more machine translation of invoices and other documents into digital format, which is basically taking pictures, images, and PDFs of invoices, and so on, and converting them into a structured format. The real benefit here is in that it allows automatic processing. Indeed, you can run two or three matching processes at once, and/or two or three automatic payment processes at once.

“The marriage between technology such as machine learning and AI, and the collaborative, network-based approach I mentioned earlier, is key here. When you have suppliers connected, you can use machine learning to pick up the invoice and automatically process it. This is also where the two-way, give and take relationship comes into play. At the start, while you’re helping the machine to learn, it might not be perfect. So, you need to ask suppliers questions and ask them to improve the information that they provide, either in the back office on the invoice. That really is the lifeblood of machine learning technology. The more you work with it, the more it will improve. The bigger the base of customer feedback you have, the more you improve.

“Once the level of information builds, you can add machine learning to other payment processes to automate more. Scalable solutions also allow you to provide tools to every employee in the organization so that they can improve their processes to be smarter when it comes to buying.”

Is this start slow and accelerate technology usage approach a good one for treasurers in your opinion?

“When you look at some of the examples of machine learning it can be scary. It can also seem like it’s a world away from treasury. However, if you look at who’s actually making money on machine learning today, and who are making the big investments, it is a little bit boring. That’s good news. For treasury it’s all about automating processes to make business more efficient, to improve speed of payments, to reduce errors, etc. It’s all about streamlining business. In many ways, machine learning is at its most successful when it’s invisible.

“The challenge for treasury comes in that it may not be good enough to take your old processes and your good old data silos and apply some machine learning to it. Sometimes you must alter your processes to fit technology. A give and take approach is needed if you are to reap the success and benefits that are available. You need to think in a different way about how you get the information. Where do you source information from?

“A good example of this is how you drive spend analytics. If you’re looking at the invoice structure, for example, the way that that is processed in the paper world is at every stage you throw away a lot of information. By the time it ends up in the system, only a few key pieces of information may be held. But, by actually fully understanding expenses you can really make a big difference to both the success of machine learning and the business. There’s something about capturing information at its very source that is hugely empowering.”

Blockchain is a much talked about technology in the supply chain. How do you see treasury departments using Blockchain going forward?

“In addition to the much talked about cryptocurrencies, which are facilitated by Blockchain, I think the technology fits with some more general financial services trends such as the move towards digitization of assets, the way you look at decision making, rather than just pure automation. The move towards digital cash is now being embraced by banks, including central banks, while regulation such as PSD in trying to open up banking. This will all help the development of Blockchain solutions.”

Do you think treasurers really want to embrace new technology?

“I think there’s great interest from treasurers for technology, although it does maybe depend a little bit on industry. Across the financial sector, we’re seeing a lot of demand and interest, to the point where some of the largest players are creating competing products that target the treasuries of today. Most are competing for access to this digital transaction volume.

“I think there are many aspects of treasury that will benefit from technology and there is interest in everything from reverse factoring to receivables financing to dynamic discounting. What is taking some time is cross-disciplinary solutions where treasury and supply chain actually work together to drive the digitalization agenda while helping on the working capital side. The message is now really resonating with organizations because if you can actually connect the processes from a source into payments, and categorize data at a purchase order level, you can drive spend classification. This can be fed back to users in the organizations so they can make more independent spending decisions. You can also improve working capital by deploying reverse factoring, for example.

“This conversation involves multiple people in the room, which we all know can be difficult, but it’s hugely powerful and the potential is huge. It’s a very exciting future for technology in the payments and treasury space.”

Related Articles

Is treasury missing the artificial intelligence opportunity?

Automation Is treasury missing the artificial intelligence opportunity?

1m Austin Clark
Treasury departments 'slow to adopt fintech from non-traditional vendors'

Automation Treasury departments 'slow to adopt fintech from non-traditional vendors'

1m Austin Clark
From moonshot to reality: AI's advances in corporate banking

Automation From moonshot to reality: AI's advances in corporate banking

12m Alenka Grealish
The future of banking: it’s all about sharing

Automation The future of banking: it’s all about sharing

2y Edward Berks
Cash application: more information + more automation = unparalleled efficiency

Automation Cash application: more information + more automation = unparalleled efficiency

2y Nils Strachanowski
How algorithms will set your bookkeeping to autopilot

Automation How algorithms will set your bookkeeping to autopilot

2y Lee Murphy
Towards the bank of tomorrow

Automation Towards the bank of tomorrow

2y James Kipling