All Blogs

AI for AR and the CFO Office- the Good, the Bad and the Ugly

It’s no secret that the dubbed CFO Stack, new systems built for the CFO, have been getting much deserved attention lately. In the past decade, we lived through the big flood of enterprise solutions. They were mainly built for Sales, Marketing and IT. Arguably, the CFO Office is one of the last digital frontiers in the organization. New entrants have emerged in categories like Spend Management, Treasury, AP and Procurement, FP&A, and A/R & Billing. 

In parallel, and let’s be honest, more importantly, OpenAI has done wonders with Dall-E and ChatGPT. AI has gone through hype cycles before, but this feels different. People want to see AI run our business in the day, and keep us company at night. That is not realistic, at least not yet. Big technology improvements come in increments, but human beings are very impatient creatures with a heightened sense of imagination. 

So, we thought it's a good time to share how we at Received think about AI in the CFO Office.

So, here it is, the Good, the Bad and Ugly, not in that order.

The Bad - thinking AI can do anything

When you think AI can do anything, you come up with bad results. We are not in a state where AI can replace humans quite yet, not with judgment nor precision.

Today's AI algorithms have an accuracy of about 80% (IGI global), and while that's highly impressive, there are some areas where you just must have a 100%. In other words, human judgment allows precision and accuracy where a machine fails.

Customer Support in finance is a good example. Billing disputes and collection efforts are delicate and fragile. These are moments where you need to chase, nudge and have difficult conversations with your customers.

Enterprise Customer Support is about relationships, trust and expertise. Customers will always prefer to speak to a real human to explain the context of their situation. A statistical model can compose coherent sentences, but is easily manipulated, inconsistent, and provides false data.

Indeed, ChatGPT feels remarkably human, which is a wonder on its own. However, we need to remember that there’s no-one there, no matter how badly we want to believe. Attempting to resolve disputes through an automated system can quickly lead to increased frustration and won't be of much assistance when it comes to getting customers to pay on time.

To sum up, we’re a far distance from replacing human judgment. Racing to replace it is a bad idea.  

The Ugly - being couch-potato lazy

There's no doubt that AI can have great benefits for finance teams. One example is analytics, where AI is perfectly designed to identify patterns, trends, and insights that can help teams make smarter decisions. The problem lies with lazy use cases, where instead of using AI capabilities to provide actual value, it saves you a mere 3 clicks.

Don’t get us wrong, we all want a personal assistant, but do we want a personal assistant that does the bare minimum? A free text query that gives a graph of your collections in the last 12 months has novelty to it, but doesn’t provide much value. We need more than that.

Companies need to use AI where it actually makes an impact, rather than simply adding it for the sake of claiming to be "AI-driven". AI's greatest advantage over humans is its ability to analyze large amounts of data. Instead of overlooking those capabilities, we should use AI models that can anticipate our needs and provide actionable insights that assist in decision-making while optimizing processes and bringing immediate value.

When you save 3 clicks exactly, the appeal will wear off very fast, and the feature will be forsaken. We think that incorporating AI into systems should solve big problems, and not just aim for the lowest hanging fruit. 

The Good - save time, money, rinse and repeat. 

AI experiences need to meet the same 'new product threshold' - they must be twice as good at the same price. Boring, we know, but it’s about saving time and money.

There are plenty of these opportunities in the CFO Office, and in A/R and Billing specifically. Take for instance invoice automation. It is obviously not the sexiest task, but automating it is a huge improvement for the finance team, and besides saving precious time it also ensures an error-free and timely process.

Billing is such a complex process, much more than people tend to think. AI can truly transform how finance teams manage these processes. 

A great example is data classification. There are many systems that are integral to the A/R and Billing process - CRM, ERP, Usage data, Banking, PSPs, to name a few. However, there is one major problem inherent in all of these system integrations - all these systems ‘speak different languages’ meaning different object naming and behavior, different logic, and inconsistent data formats.

Creating data integrity and consistency is an ideal use case for AI. Having a unified data classification layer that maps the different objects, logic, and data formats between all systems is crucial for creating a smooth end-to-end process, and ensuring accurate and unified data for analysis and decision making.

The cherry on top, is data prone to human errors; for example, invoice reconciliation can be a nightmare for some companies. Companies lose a lot of time and money due to these problems, and trust us, it hurts. This is another great scenario where AI is the perfect solution and is capable of handling the heavy lifting for us.

We have big exciting plans for using AI where it will truly make a difference, to make Finance teams lives better, and make Companies smarter.

Explore our billing solution.