Today, we're gonna talk about, what finance leaders must get right about AI.And this talk was very it's very similar to a talk I delivered to about a 100 finance leaders, recently in person in San Francisco. So it has a little bit more emphasis on the adoption in of AI in the finance suite, which surprisingly is, taking the lead this time, versus maybe the previous technology transformations. Tech finance department has always been known as conservative and a little bit more, slow to the to the adopting the new technologies. But it's very different this time, and I'll tell you why. And a little bit about myself. I am a strategic adviser in the intersection of AI finance and governance. I'm very happy to connect, on LinkedIn, by the way. So feel feel free to, send me a connection request later on.

I had a investment background for twenty years, but since then, switch lanes to the vice CFOs and board directors, especially on the audit committee, when they talk about AI strategy oversight and definitely see a lot of change of attitude over the last year or so. What we'll cover today is, let's talk about current state of AI in finance, and then I'm gonna, share some success stories and, try to in try to distill the successful, ingredients. Try to distill some of the ingredients that led to the success. And then, of course, we want to, share some failure stories. We do have a lot of them and lessons learned from there. And finally, just touch upon quickly on effective communications with sports. When you are a finance leader, understanding what boards are looking for and how to better communicate to the board and other stakeholders, obviously, is very, very important. And lastly, just summarize the actionable takeaways for you. And the but let me take a pause and check the chat.

Please let me know if you have any questions along the way. So, I'll stop and answer them. Right now, we're good to go. current state of AI in finance. AI adoption is definitely widespread and rising, which is not surprise to anyone at this conference. I think what's really surprising is already 71% of the companies surveyed by KPMG already use AI in finance division. And interestingly, it's a widespread so if you look at the details, maybe tax is a little bit slower, but other than tax, a lot of accounting, planning, budgeting, forecasting, all the all the, all the functions within finance departments are actively embracing AI. And this pose a question because you may or may not know that the CFO suite has already dealt with this crowded software stack problem. You see there's a huge map, and by the way, it's doubling and tripling every couple months because the AI, software solutions offer to these CFOs and finance leaders.

How do we really think about what tools to choose and what tools to stick with? I strongly believe that we are going through a, recoupling, which means that AI tools may start with a function or a specific pinpoint to gain trust. But pretty quickly, once that trust established, especially the data is trusted, then, then over time, these AI solutions will gain share within that tech stack and perhaps creates opportunities for CFOs and finance leaders to deal with not so many vendors, but maybe one or two or a few vendors that would provide almost like a jack of all trades, and re allow these leaders to rethink what really is finance for in supporting and enabling business growth.

It's a very exciting time. So now let's look at some of the, success stories. The one is, specifically on the pinpoint that experience with, by a lot of finance department, which is when they scale, they they find their month end or quarter end closing period becomes longer and longer because there are just so much to be, identified as, you know, the abnormal errors or abnormal numbers, and they need to find out what really is happening behind the scene.

It creates a significant block in scaling. And in this example, this company is a growing b to b business facing the same problem. Interestingly, they are between the stage where they can manage everything with spreadsheets to the stage that they probably need to buy some, really, good solutions Instead of going into the, you know, the the typical, incumbent solutions, the CFO decided to take this as a as a opportunity to really think about what, AI can do.

And one of the big thing AI can do, they find out, is to bring in some of the unstructured data, very quickly into the contacts of monthly month's end closing process, which led to a reduced close cycle to four days with a lot with a significant less, manual effort during the time.

No one wants to burn that late night oil, to do manual manual stuff in month's end, and the c CFO definitely motivated people to get rid of those with AI solutions. The key ingredients here are they really start with this specific pinpoint, which is months close, various, the flux analysis, figure out the contacts of the flux. And, definitely, with that one flow, with a lot of people focusing on getting rid of that, you know, the headache, motivated the entire team to work together with vendors about data requirements, about validation. And interestingly, this is also a, area where outcome is easily verifiable. So it it's either correct, on the correct. When you do a manual process and compare the result with the AI, output, it's easy to to, verify and build that trust. Success story number two is dealing with the overload of data.

So when, when you are in the planning aspect or forecasting aspect of the finance function, You're constantly dealing with inputs from different departments, sales, marketing, even, you know, tariff these days. Right? So how can you reflect those data into forecast? Used to be the workflow takes a couple weeks. A lot of discussion back and verifying data, and then verifying assumptions. Now with AI at this company, they implemented a, analytics platform basically connecting some of the different data sources with AI's ability to decipher, these structures, with business, with finance, with marketing. The result is definitely improving that forecast confidence. And, also, because the data is relatively fresh, everyone in the business, department maybe used to think the finance, you know, data is a little bit above their head when it comes to, business, translation.

Now they can use the AI tools to translate, like, what does, for example, the COGS, cost of goods sold in this situation means. And that really helps more discussion and further improve that, effectiveness of go to market, campaign forecast, for example. Ingredients here, interestingly, is also about, of all, data layer. Really, it's important to make sure that's that's reliable and correct and also establish that data governance Because when you are connecting all sorts of data from all sorts of, departments, who can see what data when becomes paramount? And how do you, create a policy that really doesn't, make data, you know, data leakage a problem in your business, gross forecast process is very important. And starting from that data, being very firm on that data layer, build AI capability above it is the most important success ingredients in this case. And, interestingly, both examples and many more share some common traits, and I listed here. Number one is start with specific business province. Either it's small or very, complex.

It doesn't matter. As long as it it, aligns with your, business strategic goals as a company and really pay huge attention early to data and governance above them and have that measurement to to measure the success and measure the accuracy, of those outputs AI outputs really help the team to align and build trust with AI systems.

Once the trust is in place, the imagination, creativity will blossom. And then, lastly, just have that phased approach with quick wins so that, you know, you you invite more stakeholders into this, this AI journey. We'll quickly go through the failures. One is this company is actually a large financial services company. They have all I would say most, one of the most, abundant resources to allocate to AI investment and talent recruitment. And that's actually, in this case, hurt them because they rush into deploying next gen AI analytics across their finance function because they can afford to. Right? And a few million dollars are invested in hiring expensive cost AI talents to build, predictive AI models in house.

And interestingly, afterwards, the finance team realized no one used them. Why is that? Because the AI models, while maybe technology wise, is frontier, But, really, it's hard to integrate into workflows, and the project is abandoned after nine months with no measurable, business impact. Maybe the only positive impact is the executive got to talk about their AI leadership in conferences, during that stage. But, nothing really, have the success to show to the board and to the stakeholders for the company. Lesson learned is no matter how many how much money you have, if you don't begin with business objectives, you are gonna fall into that problem with you have a hammer, you just look for nails, and we all know how that ends up. And, also, just dive into technology and then build that frontier, performance type of tier performance type of, applications may not really solve the pinpoint.

And, really, a lot of companies are not suitable to build in house for many solutions. Why is that? Because AI, unlike other technology, maybe in the past, is not once one once and done type of, investments. You need to monitor the application usage and model, draft and then really adjust to and adapt to business need changes. You not only need someone who's capable to build the version, but you need a really great product development or product management people talent in house to continue to refine and improve and enhance over time. It's a huge commitment, a lot of hidden cost and hidden risk. Are you ready? That's the problem. And the story, I'll be quicker, is the health care company. This is one of those, finance leadership, failed to engage with, business leaders in other departments about why, and how this workflow is becoming.

And this is a rare case where, finance lead innovation in this case. A lot of things are becoming super efficient, but the problem is operational, side of things are feeling left behind because they don't know what to do, what to provide to utilize this workflow. And, therefore, they go out and buy their own stuff, own AI solutions. There are repetitive AI investments across different divisions creating a huge waste. And this story is definitely need to balance the engagement throughout the project. And across these two failures, you can see there's there's a this common of this common mindset of technology and then figure out what can we use in business. That's completely the number one, pitfall in AI adoption. And lastly, I just wanna quickly touch upon effective communication with sports.

Interestingly, you will find out, boards want to know the same stuff, AI or not. They wanna know whether your strategic investments are aligned with goals, your risk management is in place, your talent, are, supporting these initiatives, and your ROI framework and measurements are in place to measure AI investments. And you need to bring in a lot of context on why and how you're doing what you do with AI. The good news is this is nothing new. It's all about business goals and outcomes. AI is just almost a superficial, layer of these discussions. Always, focus on what boards really care about. The takeaways from this, talk, I hope, is, this is convincing you that this is AI journey worth taking. But there are a couple of, you know, lessons learned from successes and failures that you can take away today to start your AI journey on the right foot or to continue your AI journey with confidence, which are start small, prioritize, data governance, and invest in people to support that and definitely measure, along the way and lead collaboratively across functions.

The future ready finance leader are, the leaders who can fully embrace AI but not being caught up in this whole technology mindset. And that's it for today.