Latest / Extropy on AIR: Finance Transformation, AP Automation & AI in Finance / Halfway Through 2026: What Actually Happened to AI in Finance
Transcript
- Ravit Gutman: welcome back to our show. I'm Raveet, and as always, I'm joined by Mike. Mike, we are officially past the halfway mark of 2026. And I don't think either of us have had predicted the six months we just watched. So I'm very excited to have you with me here. And let's start unfolding half one of twenty twenty-six. Mike Jasper: Yeah, my first question is where did the time go? I blinked and it was snowing here in New Hampshire. now it's full blown summer. but second, I would say if I looked back at my predictions from January, I wasn't even close. if you fell asleep on New Year's Day and you woke up today, you'd have missed so much in the AI market and what's going on in in finance with AI. we've seen Ravit Gutman: Just. Mike Jasper: AI agents go from things we saw at, trade shows or demonstrations to everyday life. there's been a full blown showdown between the US government and AI labs. We've got China, opening models and undercutting the frontier lab space. and the last headline we previewed was AI Labs coming to Wall Street. Within 72 hours, Anthropic and OpenAI both launched enterprise deployment arms with major financial service partners. One targeting mid-market firms like community banks and regional health systems, the other going after large enterprises with billions in backing and armies of forward deploying engineers. Ravit Gutman: And for those of you not trying to understand this, but in plain English, it means that the AI companies aren't just selling you the software anymore. They're selling you the implementation. And that's a big deal for anyone deciding who to partner with on their transformation. it validated the work firms like ours have been doing for years. So this is really what nailed for us also choosing our own model as well internally. Mike Jasper: Yeah. So now, Rabit, now that we've kinda covered some of the headlines, let's jump into what this means for finance teams. Ravit Gutman: we did the homework for you. we've been through the big H1 reports coming from KPAG FinTech Poles, from JP Morgan Outlooks, from the Cambridge and Judge and Business School Global Surveys and all the different AI reports And today we're going to do a first half review of what happened and what type and what you should do about it in the next six months. Mike Jasper: Yeah, and we look at this in four segments, Rabit. There's the big headlines. We covered some of those as is the opening teaser. we have to look at what this means for finance teams. how does everything that we're seeing in the market impact them? There's the regulation clock that's ticking away right now. And then, another lightning round of predictions on what's going to happen the back half of the year, because we were so accurate on the first half. So let's let's jump into those. Ravit Gutman: I'm interested to know if you can give me your number one story of the first half of the year. Mike Jasper: Yeah, that's pretty easy. for me it was the fact that we saw agents go mainstream. AI was a concept. We saw some proof of concept agents in the market. I saw a lot around, contract revision and contract review. that's where I initially saw things starting. but now they're continuously running in the background. we've seen all the different finance solutions, putting out agents within their platform that run autonomously. They run in the background. it's not something separate that people have to go put in and monitor, but it's helping to automate different aspects of the workflow of those manual tasks that they're dealing with today. Ravit Gutman: If you remember when we started at the beginning of the year, we talked about how less than six percent of actual organizations saw true value coming from AI in in 2025 when we went through that recap. And it's amazing to see some of the numbers that are coming back right now. one of the most amazing numbers that I've seen lately coming from the Stanford AI index was that We are at the point that 78% of global organization at least deployed one AI business functions that is really really showing that true value. And that is huge change from where we were in the beginning of the year And all of that is specifically around those type of agents. that are coming in. And what's been amazing the impact on the financial firms and seeing how much the agentic AI is coming in the finance department and showing the true value there. one of the mind blowing numbers that came recently was fifty two percent of financial firms or financial teams are actively piloting and deploying different types of agentic tools today, which is huge compared to other departments where we see that. Mike Jasper: Yeah, a interesting footnote on that same survey the most mature AI application in finance right now isn't a finance task at all. It's software engineering. 42% fully deployed AI within their their software engineering departments. even the banks, learned to trust AI with writing codes and they did that before that's sort of a a precursor to Trusting AI to manage money. Ravit Gutman: that is what happened also with the KPMG fintech report that came right after. that shows those venture investment reaccelerating in the first half and seeing how they're putting their money on a lot of those AI enabled startups that are right now making half of the fintech deals that are out there and for those venturists. And looking at Series B, Series C, and any type of other non-AI peers that we had before. one of the things that has been interesting, looking at all of the conferences we were in, we were trying to figure out how does AI fit in terms of security, in terms of trust, as you just mentioned, Where are we going to see more and more AI native firms or different type of AI gentic tools that are coming in? from what we've seen was Those two types of things are changing even for us in Extropy. how you're transforming your own business and how you're transforming the client work for this year as well. and that is what has been interesting also for us internally in Xtropy, as we share that we had our own exploration back in 2025 that we needed to revamp as well. the changes that are happening so fast within the modelings. if I look at the first half of the year just on the models and how much have changed in terms of the AI capabilities and the type of covering that we're getting from the different types of tools, and all the guardrails that are coming with this right now. We've seen the government stepping in and saying, okay, this model is too risky. you can deploy that and stopping an Tropic from deploying that. we see more and more companies are trying to build their own tools, vibe coding themselves. I'm saying this as a CEO myself, finding myself as well doing this and building those different types of digital assets that would make that impact internally for your team, for your customers, changing how you. built for the business and became this theme of real world asset tokenization that we're watching and it's it's been remarkable. I found ourselves working with different teams in our customers to how we can approach AI as well, which didn't even think that we're gonna be in that space what do you think have changed in terms of buyers when they're looking at at figuring out partners and how we are approaching deals differently as well from our side. Mike Jasper: one thing I've noticed is with AI, like any other technology, there's always a build versus buy mentality. and when AI was first hitting the scene, everybody was looking at how do we build AI? And that's where we saw a lot of struggle with organizations, figuring out how they get value, how they take this new technology and make it work for their individual organizations. What I saw was a lot of organizations buying instead of building initially, getting some of those agents that are built into the tools that they use today. And that gave them the experience and that value proof case that they could go back and say, okay, here's a very tangible area that we were able to use AI to improve our process, improve our auditability, improve our visibility. And that gave them that confidence to start looking at How do we actually go out and build the areas that aren't in going to be native in these applications? And I've seen a huge transformation in our clients and and people I talked to about that strategy. It went from very cautious to adoption through existing technology. And now they're opening their eyes we've seen so much benefit from these tools that that have been embedded in our ERP system and in our order to cash system and in our AP system. now how do we take some of that experience we've had with those and leverage it in some of the more unique areas of our business that somebody's not gonna put out an agent for us? and that's one of a whole bunch of conversations that that we have every day. Ravit Gutman: We had CFOs, they just wanna hear the AI that the tool has AI without the understanding of what does how can AI support. definitely we see people more educated, more conservative about AI, understanding what it means for them. But let's talk about what does it truly, truly mean for our finance folks on the team. let's bring it home to our world right now. And when we're looking at that CFO office for AP accounts receivable, where our finance team at with essentially AI at mid-2026? what are some of the things that you're seeing? Mike Jasper: So I'm seeing a lot of the areas where they're evaluating their process, they're evaluating the gaps that they found and putting together priorities of where do we start first? and they're starting. That's that's probably one of the bigger changes I've seen. You know, we've anecdotally been able to point out a thousand different places and organizations where AI could be a benefit. But being able to put together a priority list and start working through that. has been something new. I've seen a lot more organizations start taking the leap. And that starts with the most manual tasks where they're high repetition, looking at, matching processes, looking at things where existing technology has been good but not great. And they're looking at how do we how do we optimize a lot of those areas and putting together their wave plan for the next six months. Ravit Gutman: And I love that you said the wave plan because that's one of the biggest lessons learned that came from 2025 with a lot of companies with how we consult today and advise our clients. So a lot of them have gone through that pilot first program or do that proof of concept. what we have been trying to explain for many years with digital transformation, that it's garbage in, garbage out. You don't wanna automate that inefficient process and you don't wanna underestimate complexities of things. And you do need to focus on what does it mean in terms of change and impact to your team. And people are starting to understand how to approach project differently. And they even have some of the scars of on trying to put AI and things that are messy and how important it is to do the work and make the changes And do what needs to happen in order for AI. it's no longer that magic pill that everyone thought about it in the beginning. And there is a strategy behind digitizing and building that true data foundation. and this allows us also to build a lot of the programs and the actual journey and seeing this is a journey to get where you're looking to get to. Mike Jasper: Absolutely. some of that exploration was definitely needed to be able to get to the strategy. even look at how we approached it internally. we challenged everybody on our team to go explore AI and see how it can impact their day-to-day job and how they can use it to optimize different areas within what they do. And now we brought it back internally to, okay, now that we've kind of got a good feeling for these things, how do we prioritize these? How do we work together? So we don't have three different groups working on the same project. And project one and project two are going to work together to produce a greater output. that strategy is something we're seeing. And the organizations that kind of dove into AI and they may not have seen the value initially. I think they're following a similar path to what we're doing at this point, where they're now starting to see, okay, there's a lot of benefit here, and we have to run this as a strategic initiative. not just as a magic pill, as you said, so that we can see what we're doing. it's becoming more of a real software, a tangible product versus this idea that was in the market initially. Ravit Gutman: True. But we still live with that execution gap being that number one problem on many of these. and one of the stats that came for one of the reports was that they look at those financial departments, as we said, it looks like eighty one percent have tried to adopt financial department has tried to adopt AI in one way or the other, but Mike Jasper: Yes. Ravit Gutman: 14% call it a transformation that was part of their strategy and they were successful in it. So definitely we see the numbers are climbing, which is great and I love it. but we still see the execution gap still holding one of the biggest elements here. So I wanna say that one of the more close to home items that I've seen in terms of teams an impact I would say with AI is FPN. and that is something that I would say we have seen growing tremendously in everything to do with forecasting, analyzing, that is where AI has shifted things dramatically for a financial department. And now it's going back and pulling things back to more of the core foundation department, as we said, like the accounts payable and then the accounts receivable, and making sure that all the data for FPA is accurate and and can pull those data. this reverse engineering of trying to figure out what is the best use case. But we're definitely going to see. And we are seeing a lot of the partners that we're working with and the huge changes they're having with their softwares and with AI as well and what it contributes to our customers. Mike Jasper: Yep, absolutely. Ravit Gutman: one of the things that Is still sitting is regulations. so now that this is the segment that no one loves, and I have at least went to four different conferences this year to try to figure out how our partners are handling regulations, try to understand how the big four are handling regulation, and just try to get the answers. How do we approach regulation in this time. and We got to that point that we know that we need to have this piece already in terms of a compliance already there. and we even received some dates that things need to be in place by a certain time, especially coming from the EU. But I'm curious, like what are you predicting in terms of that regulation clock that is coming for the second half of the year? Mike Jasper: the regulation clock, we've got August second, I think is one of the dates that's about three weeks from from where we're recording right now. and that's for the EU's AI acts high risk obligations. But I think that that is something that some organizations are focused on. A lot of them aren't. and that brings risk into these projects. thinking about how How we looked at e-invoicing, invoicing used to be a wild, wild west. you could send whatever you wanted. It's slowly become more and more regulated. You know, France has put out their e-invoice mandate for September. and there's a lot of fines and penalties and things that organizations have to be aware of. That same regulation is starting to move into AI. And that's to me almost even more impactful because it's going to sit at the foundation of your process. It's not the output that you have to manage, it's how you get to the output. And that's gonna take a lot of planning and a lot of strategy and a lot of advisory on how do we build something that is going to meet the regulations as stated, but also how do we make sure everything's documented? How do we make sure everything is is in a way that if a new regulation comes out next year, how are we able to go back and make sure that we're we're compliant with that without Scrapping and rewriting everything we did. Ravit Gutman: And we've seen that happen with the E invoicing you more than anyone. and I think before anyone says we're in the US, this is not our problem. if you have an EU operation, this holds true also to you. And this is becoming everyone's problem. And we see where that feel fear mongling is happening right now. So if your AI touches either employment or credit decision in Europe, or it's very likely that it is your problem. And I think that's one of the things that firms don't understand that they're missing that earlier checkpoints and are technically already somewhat out of compliance. And we're seeing that a lot with our customers as well. Like everyone are running after like I said on the e-invoicing mandates and trying to get ahead of this. And This is where we're gonna see a big shift as well in the consulting world to help a lot of customers doing their day to day while making sure they're hitting those compliance. Mike Jasper: Absolutely. there's a lot of uncertainty and there's gonna be a lot of change in the next one to three years around regulatory actions. and to your point, organizations becoming more and more global. And if you have a tax ID in a European country, you're obligated to review and apply that law internally. I've talked to some organizations where they say, well, we're just gonna do a different process in the EU than we do in the US. That's not optimal. That's not gonna get you that that consolidated optimization that you're looking for. that even happened back with some of the privacy laws. originally, EU had its own privacy laws. we were just gonna do that and do something completely different in the US. organizations found that ultimately it was too complex to try to manage every regulation individually. So they had to merge those together and have one corporate strategy for global regulation. That exact same thing's gonna happen in AI. And the organizations who plan for it and do it from the gate from the the start, they're gonna be a lot more nimble than the organizations who ignore this and wait and then realize later they have to come back and revamp everything for compliance. Ravit Gutman: That's true. And and not even to speak about all of the the pendants that is happening right now where that concentration risk is. from the beginning of the year so many companies implementing essentially open AI foundation models that exist within the organization and As we've seen already how things have shifted between Entropic and OpenAI this year and even Gemini and we want vendor outage policy or outage or a policy change and that is going to become a systematic event. and I would say that one of the things that it's interesting, even seeing looking at some of those data protection acts, that You need to understand the models that you're using. You need to understand what is the security for these models and you need to report on those models. and this is becoming a thing, like I said, not just in the EU, but it's also here in the US. And that concentration is also a risk. Okay, so I wanna start touching into stuff that I'm more excited about, which is some of the have to predictions that we're seeing. So one prediction each of us can talk about in the second half and then a rapid yes or no if we think that will happen. Mike, you're first. Mike Jasper: All right, I think by end of year, the the question of what model do you use, that becomes the the boring question. we're gonna move more towards orchestration. I think the way we work with AI is gonna change again. Think about how we we started with our prompts and just the way we interact with AI has completely changed already in the last six months. I think that's gonna continue to change and we're gonna have to continually adapt to work with that. Ravit Gutman: I would agree on that. I think that's we've seen that and internally as well. mine would be that pilot to production gap that finally will start closing in finance. And not because the tech is improved, but because CFO will start putting that AI line item as that twenty twenty seven budget with that demanding hour high behind it. Nothing matures a pilot faster than that budget review against it. And I feel like we're starting to have conversations about it for projects that we are preparing for twenty twenty seven. So I'm very all in on that. Mike Jasper: Yeah. I would fully agree. And I'd say the demanding ROI is gonna be the new piece. everybody's been exploring. Now when we have to have to achieve what we say we're gonna achieve, that's gonna bring a whole nother level of work to AI. okay, Rabbi, rapid fire. The EU AI high risk deadline for twenty does that slip to twenty twenty seven? Yes or no? Ravit Gutman: With everything that happened with the e invoicing mandate, I would say slips to twenty twenty seven. What would you say? Yeah, I think there's just too many mandates right now and it's too much to cover. And knowing the EU and everything they're handling with right now, yeah. Mike Jasper: I would agree. I would agree. Ravit Gutman: Okay, so now my turn. Mike Jasper: Yep. Ravit Gutman: I think this one we already have but let's see how where you're at on your news. But at least one major enterprise publicly blames an AI agent for a material financial error before the year ends. Mike Jasper: that's one hundred percent gonna happen. Ravit Gutman: I we definitely had a couple of blames to I yet on the financial side, but we've definitely seen I think one of the big fours being blamed for some of these and we're we're definitely seeing some news there. Mike Jasper: Yeah, that that's not even a question. That's that's going to happen. Ravit Gutman: Sure. let's hold our thoughts and I think we'll grade ourselves on that December episode of where we are in those type of predictions. But I'm definitely excited to what's coming up ahead of this, Okay, let's bring this one home. three takeaways from the first half of twenty twenty six. So one is that agents are the infrastructure now. And the question is which workflow, not whether two for finance team or where the money and the production will go, not pilots, pick that process from end to end, build that plan. And number three is that governance is no longer optional. make sure you inventory your AI before August, whether Brussels decides or not on the mandate, but we'll see from there. Mike Jasper: Yeah. I'll say one thing the first half of this year taught us is that six months is a lifetime in the AI market. So whatever we just said in this episode, check back in December. Let's see how accurate we are. it'll be a interesting to see. Ravit Gutman: Yep. And this is episode was useful, so please share this and with one finance leader who still is in that pilot prerogatory. we'll see you on the next time and keep transforming forward. Mike Jasper: Thanks, guys.