FinFloh’s Post

Transcript

How have you seen AI impacting finance operations at a broader level? There is a phenomenal work I can do. I'll start with your company. I'm impressed with decision intelligence, what it does right, basically predicting which particular customer will pay late and this is that you are trying to get a kind of a risk score associated with it and then trying to put it up in a perspective that, OK, what should be the dynamic pricing. That's very interesting. You need to produce a lot of summary when you are producing your financial reports. Generally, I can take care of that. I think another use case could be having a chat bot where you ask relevant questions and it goes through the entire platform and gives you the answer rather than you. Well, this is a very interesting one because these financial controllers do not have the time to pan or specify their requirements as well. So I want to understand the reasoning behind this number. It should be able to respond back. It saves a lot. Look at it from an auditor standpoint. Give me the entire audit trail for this particular invoice number. Somewhere down the line, you are moving towards an area where with the help of AI you can make the entire order to cache predictive and you can have dynamic contracts as well. I'm a collector, which is the first customer I need to call. Do I pick a customer who's hardly $100 but has been waiting for 120 days, has not paid me? Or do I pick up a customer with just 30 days overdue, but it's $50,000 customer? How do I prioritize? So collections will be moving into an agentic AI way into the future with a triage 2, obviously a human in the loop that if in case somebody is completely paranoid, emotionally charged, OK, transfer it to human. In 2019, Microsoft announced cognitive APIs. This was a very big disruption from an AI standpoint. So they were using machine learning algorithms to let's say read 300. Different types of handwritings and accurately capture it with a 98% rate. I think with the cache application, right when you receive the creation statements, right? And then today if you look at a standard cache application model, the money arrives in your bank. If there is the payment reference or indication of a particular invoice for what it has come over into any comments or any reference field unique identifier, you can automatically match it. That is automation. But let's say there is nothing mentioned in there. This is the metadata I can determine which is the best fit match for this particular invoice keeping human in the loop that I believe this is what it is. Instead of you finding it, I have gotten it for you. Can you please confirm if this is the right one? Correct the operations. It's just a decision that has to be taken with AI. I have to do feature engineering, make it simple, maybe define new attributes, OK, and then use AI because the metric that you are after or that is the key important input for you, it might not be there into the existing data set. It might be a combination or some sort of formula that has to be implemented to get to that. It has been always difficult to convince somebody to invest into technology because even if they buy it, they might not be understanding it fully. The same issue is there with AI. People have a kind of this despondency whether it will work or it will not work, but you should know where you're putting the tools, not just writing poems for you. It can deliver much more business impact. Yeah, it shouldn't be like this, buying AI for the sake of it and not knowing how exactly to use it. Any real life example of where you feel AI has improved a company's overall finance process in general? There was a very large multi billion dollar technology conglomerate. They had a lot of travel and expense reports. OK. So they put AI for audit. So they built an AI, which is, which are the right candidates to qualify based on different parameters. And they originally anticipated, OK, this was a technology services company, managed services, we'll give you 30% productivity. Productivity was 75% once they implemented because there is no need, it has eliminated false positives, correct. It will not replace jobs. Do not see AI versus humans, see humans with AI versus humans without AI. So it's not about you not adopting it, it's about your competition who has adopted it. They are commercial numbers. Their productivity, their customer experience, their accuracy, their business value has taken a shot. And every time it will go high, it will be at your cost every time somebody else gets a customer. Your opportunity to get the same customer is gone. So you to be best in class, AI is what you have to go for right? Forget a if it is not A, it will be something else. Can you match up the speed and efficiency of AI? You cannot.