A Different Approach to Managing Spend – Analytics and Machine Learning

At first, a novelty, Artificial Intelligence (AI) has been evolving rapidly over the last few years, and we’re starting to see how it can deliver real business value. There has however been both a tangible sense of excitement and concern around AI and machine learning. It seems, just like flying cars, we over-estimate how quickly humans and businesses can implement a new technology. Although still in their relative growth phases, both are making leaps and bounds in innovation. Don’t worry though, we still have a long way to go before a robot takes our jobs.

The big data explosion has accelerated the growth of intelligent technologies. Data is the lifeblood of AI. Intelligent computers that can think for themselves have been around for a while, but before the emergence of big data we didn’t have the computing power to drive them, so we are only now starting to reap the benefits of computing, machine, and deep learning.

It took a small development at Google however for AI to go mainstream. Word2Vec is a neural network (based on the brain) that changes words into vectors with meaning and context, allowing machines to analyse and learn about context from phrases. For example, take an invoice with a line of text that says “plastic, 500 ml, Crystal Geyser,” the technology will infer that it’s bottled water even though it doesn’t say so because the machine has some prior knowledge to establish context.

In recent years, we’ve seen a growth in innovation across the finance industry. The growth of technologies such as crypto-currencies, contactless and mobile payments, and the introduction of intelligent algorithms demonstrate it is an industry willing to innovate and embrace new ideas.

It is in managing spend for businesses though where I believe the next major opportunity exists – particularly for the development of AI and machine learning. It is something we at Coupa are already doing. Our technology gathers data from invoices, purchase orders, card information, travel, and expenses; in short, data about anything that businesses have spent money on. All this data is then joined so that it is consistent before being run through three processes; two machine and one deep learning. The machine learning is primarily for normalisation and deep learning for classification. It’s a blend because machine learning is quick, but neural networks can take hours.

[clickToTweet tweet=”Once this #data is processed it becomes ‘knowledge’ within the #system and allows us to classify 85-90% of spending. The final step, #DeepLearning, can then take classification up to 95% #AI #MachineLearning #Spending #BigData #Data #Tech” quote=”Once this data is processed it becomes ‘knowledge’ within the system and allows us to classify 85-90% of spending. The final step, deep learning, can then take classification up to 95%.”]

Once this data is processed it becomes ‘knowledge’ within the system and allows us to classify 85-90% of spending. The final step, deep learning, can then take classification up to 95%. With the information this data produces, procurement teams can help to answer business-critical questions such as, who is buying what from whom, when, in what quantity, where are they shipping it to and how are they paying for it. These are all classic spend optimisation questions that procurement teams have been looking to answer for a long time.

Insights sourced through AI, machine learning and deep learning can be applied to business spend management by normalising, enriching and classifying spend data, eliminating the need for costly, inefficient rules-based integration and transformation of raw data to understand spend footprints. Businesses taking advantage of AI within a spend management system can also gain visibility on total spend including tail-end, which offers more control of the buying cycle.  It can also provide prescriptive recommendations, such as who the most reliable supplier for a service is, or warnings about risk.

AI’s ability to be ‘all seeing’ across entire organisations gives it the power to support security and theft prevention. In spend management it can be used to identify fraudulent and suspicious behaviour by creating a fraud profile based on analysing aggregated data for expenses, purchase orders and invoices. This profile score and related spend transactions can then be used to alert a company’s internal auditors or finance personnel for further review and action.

As AI and machine learning technologies require large amounts of historical and ongoing operational data, to truly access and critically benefit from insights, organisations are best served to work with an expert partner. The historical knowledge required extends to multiple languages and character sets, as well as industry-specific or otherwise esoteric commodities. The value of AI to the market is as a result why it is so prominent in Coupa’s latest update – R20.

Up until this point spend analytics has always been a niche deep in the back office. Total visibility via analytics is the keystone to addressing the issues that businesses across the globe face, particularly as disruption forces efficiency drives. Utilising these technologies will help companies save money and act more effectively and responsibly. I find it all very exciting, particularly because I know it can make a massive difference in transforming a company for the better.

+ posts

CIF Presents TWF - Miguel Clarke

Newsletter

Related articles

Generative AI and the copyright conundrum

In the last days of 2023, The New York...

Cloud ERP shouldn’t be a challenge or a chore

More integrated applications and a streamlined approach mean that...

Top 7 Cloud FinOps Strategies for Optimising Cloud Costs

According to a survey by Everest Group, 67% of...

Eco-friendly Data Centres Demand Hybrid Cloud Sustainability

With COP28’s talking points echoing globally, sustainability commitments and...

The Path to Cloud Adoption Success

As digital transformation continues to be a priority for...

Subscribe to our Newsletter