Our clients are experiencing first-hand the power of AI within Accounts Payables and Digital Transformation. Over the past three years, we have trained our AI model on millions of supplier invoices. As a result, our AI-assisted Optical Character Recognition (OCR) software is able to accurately identify over 98% of data points on supplier invoices. This means supplier invoice data can then automatically be matched against purchase orders, contracts and receipts data within our clients’ internal ERP systems. This means Accounts Payables teams can be more efficient in reducing their invoice cycle times from days to hours. This means better supplier relationships, more control and enhanced data-driven decision-making with Accounts Payables. While our AI is transforming Accounts Payables, AI as a whole is at the very beginning of a very powerful transformation curve for Financial Services as a whole.
The Fintech Summit here held in late 2023 in Sydney is in its 10th great year. I was delighted to be asked by Glen Frost the event producer to be on a panel exploring the subject of AI and Financial Services. Rebecca Cope the digital economy partner at Ashursts, the International Law firm and summit sponsors, moderated the discussions, which included Gordon Campbell from Rich Data Co and Yao Tang. Below, we examine the key takeaways from the panel so that you can see how your business may need to deal with AI within Financial Services.
Table Of Contents
- The Importance Of Bona Fide Data
- Low-Level Tasks Will Be The First To Be Replaced
- The Insourcing Of Outsourcing
- Analyse Existing Internal Data First
- Beware The Bias In Your Data
The Importance Of Bona Fide Data
If the data models used to train your AI are faulty, the results will be challenged as well, reinforcing the adage “garbage in, garbage out.” This has been witnessed all too often with the infamous “hallucinations” that ChaptGPT seems to generate from time to time (considering that it has primarily been trained on data directly from the internet up until late 2021). Big tip: check any data and information derived from Chat GPT content that you may have written yourself – don’t just assume it’s correct.
Low-Level Tasks Will Be The First To Be Replaced
Low-level jobs will be prioritised in the early phases of AI implementation in Financial Services and beyond. This is especially true for low-level interactions with web-based AI chatbots, which will improve further. These tasks are seen as lower risk and would have less impact on an organisation in the (hopefully) unlikely situation that the technology will fail to perform.
The Insourcing of Outsourcing
We will witness the “Insourcing of Outsourcing”. If AI can replace low-level customer service and low-risk activities it makes sense to let the technology deal with these low-level tasks. At SpendConsole, we are already seeing the early beginnings of this trend with some of our clients bringing offshore activities back onshore through the use of AI. This gives more control back to local offices
Analyse Your Existing Internal Data First
Analysing current internal data can be the lowest-hanging fruit in finding improved outcomes while decreasing the need for manual intervention. For example, during the last three years, Spend Console has been utilising AI and Machine Learning to extract data from our clients’ supplier invoices received in a variety of formats, including excel, pdf, email, and Peppol.
Beware the Bias in Your Data
Bias in any data, whether statistical or ethical can skew the outcomes for any business. We must establish regulations to ensure that decisions are not adversely biased, which can lead to decisions that may inadvertently promote disadvantage.
AI is here to stay and needs to be embraced by all. It has the capacity to drive significant digital transformation across organisations. This digital transformation will, however, require appropriate regulatory frameworks, which will take time to implement.