Supply chains are becoming increasingly extended, complex and global, meaning costs spiral out of control and payments can easily run late.
To solve this issue, Previse and Virtualstock, have partnered to enable procurement professionals to ensure their suppliers are paid instantly with Previse’s advanced machine learning technology while at the same time managing procurement across their entire supply chain in real time.
PaymentEye spoke to Paul Christensen, CEO, Previse, on how AI is bringing supply chain payment solutions into the future.
Why are supply chains becoming more complex, and what problems is this creating for organisations and their suppliers?
Globalisation has led to supply chains becoming increasingly extended and complex. Companies operating in the retail sector for example, can work with tens of thousands of suppliers from all over the world to get the products or materials they need. In addition, there is a greater concentration of smaller suppliers, driven by a series of factors including technology making it easier for SMEs to launch and distribute products. For the largest corporates, spend on suppliers can represent hundreds of millions if not billions of expenses. Given the long time lags between ordering, delivery and payment, there is huge cash flow pressure on the suppliers. Over the last decade, the largest corporates at the start of those chains have used their negotiating position to extend payment terms, introducing additional pressure that ripples all the way down the supply chain.
While most buyers would like to be able to pay their suppliers faster, the complexity of their internal processes often leads to slow or late payments for suppliers. Our research shows that late payments negatively affect 77% of small suppliers. This has catastrophic effects as often they have to access expensive short-term finance, and frequently face going out of business. Suppliers also have to increase the prices of their products or services meaning buyers are also punished. Slow payment is a lose-lose situation.
How can these problems be remedied?
Previously, solutions to this problem such as supply chain finance and dynamic discounting were only available to the largest suppliers due to onboarding costs and compliance checks. This meant the smallest suppliers were left by the wayside and still suffering from slow payments. Furthermore, those solutions solve problems for the large buyers and are not intended to help suppliers.
However, with new technologies such as artificial intelligence (AI), there are now solutions which enable buyers to pay suppliers of all sizes earlier or instantly, solving this problem for the entire supply chain. Our recent partnership with Virtualstock combines our AI technology with its eProcurement platform, The Edge. This enables procurement professionals to ensure suppliers are paid instantly while at the same time managing procurement across their entire supply chain in real time, creating a seamless ‘purchase to pay’ supply chain model with cash-on-delivery for suppliers.
How else could AI be adopted in providing payments solutions in the future? What do you think is the next technological advancement required in B2B payments?
I think the focus should be on the adoption of recent technological advancements in payments – we have a long way to go! If more companies, both buyers and suppliers, continue to accept and adopt these advancements, we can solve the slow payments problem globally.
After the scourge of slow payment is eradicated, we could start looking at how AI can be applied to other parts of business. AI holds huge potential for automating cumbersome back-end processes which take a lot of time and cost a lot of money. For example, machine learning can take over sorting through invoices from companies that currently need it done by hand. This will not only speed up this process but also increase its accuracy, saving companies time and resources.
Is B2B is the new focus point for innovation in the payments space?
B2C payment innovations have progressed to the point where customers can buy a coffee in Starbucks with the tap of a card and the cash transfers instantly. Now, imagine if the customer walked out of Starbucks with a coffee and said, “Send me an invoice and I’ll pay you in three months.” That’s what currently happens in B2B. That is a huge, global problem for $124 trillion of B2B commerce. With machine learning technology solutions, we will get to the stage where B2B will be the same as B2C and everyone can get paid instantly. The technology is there, we just need buyers and suppliers to use it.
Solving problems in B2B would make a huge difference to the overall economy. Late payment for example, causes the death of 50,000 businesses every year, putting people out of work and holding back the UK economy. The Federation of Small Businesses estimates that if all businesses were paid on time it would boost the UK economy by £2.5bn annually. With technology now able to ensure businesses are paid within 24 hours, widespread adoption of instant and early pay solutions will have a major positive impact on the economy.
Whitepapers
Related reading
Central banks best suited to issue digital currencies
By Aaran Fronda A recent report by the Official Monetary and Financial Institutions Forum (OMFIF) said that central banks rather than private ... read more
Instant payments: innovations inbound for corporates
In 2020, instant payments look set to continue their current trajectory to become the biggest trend in payments. While these schemes already offer numerous benefits to corporates, leveraging innovations such as APIs and request to pay will go some way to unlocking their full potential, argues Michael Knetsch
Obstacles exist for banks to meet ECB’s instant payments goal
The cost of joining instant payment platforms will be one of many hurdles banks and payment services providers must overcome to meet ... read more
Banks must be aware of “biases” in data used to train ML models
Financial institutions need to be conscious of biases in the historical data that is being used to train machine learning (ML) models, ... read more
