Why AI in CPM Still Fails Without Trusted Data - with Kelley Lynn Kassa, BARC

00:00:00: The biggest challenge facing end users today in CPM is data.

00:00:06: And I would argue that every CFO and FPNA leader should have on their whiteboard, it's the data stupid!

00:00:14: Our studies get into more granularity.

00:00:19: I'll give you a little preview of our upcoming trend report.

00:00:22: We actually break out data management with Data Security, Privacy and Transparency but it all comes back to.

00:00:29: the biggest challenge is solving the data riddle until You get that solved?

00:00:53: Hello and welcome to the Data & AI Culture podcast.

00:00:56: My name is Karsten Bange, founder and CEO of BARC.

00:00:59: And my guest today it's my esteemed colleague Kelly Casa.

00:01:03: she's part of our North American Analyst team and a true expert on Corporate Performance Management.

00:01:11: We will talk about AI usage in CPM The biggest trend talk about what companies are actually doing with AI, where they see benefits and opportunities but also challenges.

00:01:29: I think Kelly has a great view from her perspective being in the North American market but also working with global companies but also on the Global Barg Analyst Network meaning that she can pull from real-world examples but also from the Barg research, which gives I think a very good overview on how companies are adopting this big trend of AI and CPM.

00:01:58: And where they see opportunities, but also real benefits when using AI for CPM?

00:02:04: Enjoy the episode!

00:02:06: Hello Kelly welcome to The Data & AI Culture podcast.

00:02:15: I am great.

00:02:16: I had my morning rowing practice on the Charles River today outside of Boston, and i'm raring to go!

00:02:23: That sounds marvelous.

00:02:25: so AI in CPM In The Office Of Finance is our topic today because you're an expert at that And it's a big topic in the market as well.

00:02:35: Everyone talking about it.

00:02:37: companies are trying to figure out how that works.

00:02:40: Why do you think is there such a big interest around AI or using AI in CPM, Incorporate Performance Management?

00:02:49: That's a big question with lots of different answers.

00:02:53: I think on the very high level they're so much promised with AI not just specifically for the Office of Finance but all sorts of data applications.

00:03:06: But then when you get down to the specifics They're still faced with a lot of manual challenges.

00:03:14: you would be surprised at how many companies.

00:03:17: Still use Excel for planning and forecasting in budgeting even though they might also have a dedicated CPM solution.

00:03:25: when your looking at excel- Our data coming out of the planning survey, twenty-six talks about The opportunities to gain efficiencies To flag Anomalies to reduce manual errors.

00:03:55: You know that's the.

00:03:57: That's the hardcore promise if you will aside from the really aspirational prospects for AI and finance?

00:04:07: Yeah, so you mentioned like reducing manual work basically raising efficiency as one of the main drivers.

00:04:15: What about quality?

00:04:17: what about effectiveness?

00:04:18: is there also a driver that companies expect different outcomes or better outcomes?

00:04:23: Yes but and I'll get to the button in minute with uh In particular with North America's moving a little more quickly than then Europe um north america At this point, more likely to target advanced planning use cases.

00:04:40: So they're looking at improving forecasting accuracy in particular with or by capturing complex drivers.

00:04:50: They are also looking faster creation and evaluation of scenario planning Which is a very interesting aspect.

00:04:59: when companies buy CPM solutions You know, they're looking at how quickly can I get into scenario planning and all the good fun stuff?

00:05:10: And reality is as our colleague Christian Fuchs mentioned recently on LinkedIn that even companies with dedicated CPM solutions They want to do scenario planning but they're not there yet.

00:05:25: So AI brings that promise of some faster creation an evaluation of the scenario planning aspect that I think is very interesting, and this gets into the butt.

00:05:39: Companies are looking at AI to enable more data.

00:05:45: Um, you know sort of the hesitate to say legacy because it's only about ten years ago But a lot of those systems were very siloed and finance was in its own little silo.

00:06:00: And now the promise of AI and where companies can go.

00:06:04: Can be more High-level enterprise wide strategic planning but they need to get all that data In the right place and in the right ways.

00:06:16: Okay, and so you mentioned advanced planning of something that companies in North America are moving faster into compared to Europe.

00:06:26: Do you see other differences?

00:06:28: Different types of behavior maybe in adopting CPM solutions for example...

00:06:32: Yeah!

00:06:33: So well I should say yes or no.

00:06:37: the data says the North American market is moving more quickly.

00:06:45: is that whether you're in North America or Europe, or the rest of world there's a significant issue with lack of data availability.

00:06:58: Reliable data quality and ultimately a lack of trust in AI results.

00:07:05: because if we don't have all the right data that's governed then how do believe the results AI gives you, whether it's scenario planning forecasting consolidations.

00:07:19: You know?

00:07:20: It is a lot of hard work to deal with data foundation and that's commonplace all companies are looking at right now.

00:07:30: Yeah, absolutely.

00:07:32: Isn't that a place where AI can also help?

00:07:35: I mean actually now.

00:07:37: quite a few guests talked about how AI improves data management in general.

00:07:44: do you see this for CPM use cases or in CPM installations?

00:07:50: Absolutely it is great use of AI.

00:07:53: and when we think what the CPM vendors are talking about AIs, they're talking about agentic workflows and this agent talks to that agent.

00:08:05: The reality is we're still in the plumbing phase of AI.

00:08:09: In one use case I've heard recently as Unilever Global CPG brand They're using AI for their data quality And data prep activities that then funnel into the plans and the budgets in the forecast.

00:08:28: so.

00:08:28: That is a great use case an it's a great place to start with A I. you know you don't want us start with- the idea and we hear this a lot.

00:08:38: yeah why do I even need CPM software?

00:08:41: can I just throw insert your AI tool at my spreadsheets but start to use AI in those pilots, cleaning up your data and fixing your data.

00:08:59: Where do you see the limitations of using general AI platforms or especially gen-AI platforms for CPM?

00:09:09: Especially for planning, we just mentioned no!

00:09:11: We cannot do that... But why can't

00:09:13: we?!

00:09:14: What's the problem?

00:09:17: I would turn that around a little bit.

00:09:18: You know, we all use the phrase or most of us now are using the phrase human in the loop and as you might recall from our did an analytics retreat in greater Denver back in May Ben Shine of Domo actually said Human In The Lead okay?

00:09:37: And so one of the challenges is...you need to understand your data and what's going on in order to then trust your results.

00:09:51: So an example, which is not a CPM example but a real life example.

00:09:56: I took my favorite Gen AI tool And i uploaded a rowing video To it and said that im at this seat Im doing This!

00:10:05: I want longer stroke.

00:10:07: How do I fix the problem?

00:10:09: After six or seven prompts It kept giving me The wrong answer But I only knew it was wrong because.

00:10:15: I know enough about what I was doing and enough about rowing to say no these seats are aren't behind me they're ahead of may- and i can go on around but I won't, the thing is that you need a human in the lead too.

00:10:32: understand if those outputs right or wrong.

00:10:37: You can just trust that when gives you is correct.

00:10:43: And do you see the role of CPM software changing with now a GenTig, or it's first-of-all GenAI being in the mix and finance departments do use it?

00:10:58: Or also now with the move to more GenTIG AI.

00:11:01: How do CPM vendors react to those trends?

00:11:07: I think... It is going help users get into that promised goal, if you will of having more accurate forecasts.

00:11:17: Of forecasting more easily...of scenario planning where you can understand if X happens the result would be Y. I think that eventually we'll get to really sexy part.

00:11:32: but it's right now a lot of promise But still is hard work.

00:11:40: What do the vendors do?

00:11:41: Do they build new functionality in AI-supported.

00:11:46: Let's stick to example scenario planning, I mean that has been covered for the last twenty years.

00:11:52: so what have changed?

00:11:54: So...what is changed obviously is AI.

00:11:59: but when i'm thinking about what has changed if im an end user there are some Very different approaches to CPM.

00:12:07: So you're gonna have some legacy tools that's, That are kind of bolting on AI features because all the sudden it's The hot sexy exciting thing.

00:12:17: then You've got a lot of emerging players that Are claiming they're AI native.

00:12:22: Some aren't and I think that makes the software selection process challenging.

00:12:31: And especially in this age AI demos in CPM can look very, very exciting and lots of sizzle.

00:12:42: And the challenge for an end user is to figure out what fits their specific need-and use case?

00:12:52: It's not a matter which vendor is best.

00:12:56: or hey this vendors got bolt on AI but maybe You need it's sort of like.

00:13:06: I use the example of them hamburgers.

00:13:09: A lot you know, The best hamburger might be the thirty two dollar hamburger.

00:13:17: If you look at statistics, you know the certain golden arches is probably best if you look a sales volume.

00:13:26: But you may just want to twelve dollar hamburger and that doesn't mean all those other hamburgs are bad.

00:13:32: it's what best for you.

00:13:34: And so when your looking at the CPM software vendor landscape, ya know?

00:13:41: You need to figure out.

00:13:43: is AI native best-for

00:13:45: me?,

00:13:45: Is more of a legacy comprehensive platform approach Best For

00:13:50: Me?,

00:13:51: Is there specific things that I wanna solve in my planning That leads one vendor over another?

00:13:59: You can't just believe the demos, and frankly you really want to look at how those tools play against your own data.

00:14:08: And that's how AI is changing the landscape in relation to selecting your software.

00:14:15: Yeah!

00:14:16: That makes a lot of sense.

00:14:17: I was thinking about... Is the AI adoption in let say leading CPM vendors rather similar across-the-board?

00:14:29: Or is it a way for vendors to differentiate greatly?

00:14:33: So also, the customers or prospects and people that are interested in it.

00:14:39: Is this something where they will see big differences?

00:14:42: how vendors use AR?

00:14:44: I think absolutely!

00:14:47: If you look at what the vendor's saying.

00:14:50: as i mentioned agents running off all over.

00:14:54: One of the things I learned, having been to a number of user conferences this spring is you then on the other end of the spectrum have a lot of end users that are just trying to figure out where to start with AI.

00:15:06: And so I could see a lot differentiation in not just how vendors appear but also they guide and sort-of hand held their customers' prospects in the AI adoption journey.

00:15:26: You know, right now everything's in the messy middle of okay.

00:15:29: we've got to start.

00:15:31: I think maybe not a lot of companies.

00:15:33: Um, you know one of the things that we're seeing is companies in North America are moving more quickly on pilots and part of That Is?

00:15:44: We're also seeing that there's a lot Of peer pressure for north american cfos And fpna leaders to adopt ai quickly um.

00:15:54: So I think it's A very interesting time To be In The CPM Space because There Are All These Different Factors driving how users change, whether using CPM software?

00:16:08: Absolutely.

00:16:08: You mentioned the peer pressure.

00:16:10: so there seems to be a pressure to apply AI but we have increasing discussions that companies are also increasingly question their value and return of investments in AI.

00:16:22: maybe bit of appointed questions Are CFOs doing better job than using AI or do they run into same issues?

00:16:33: Yeah, maybe because of peer pressure.

00:16:35: They're investing in their eye but then see okay many use cases actually do not yield the returns they were hoping for.

00:16:45: Do you see it as same thing like every other department or a company?

00:16:50: Or is that different from the office of the CFO?

00:16:52: Well I think if asked one our data management colleagues That question would argue that Data Management has much better use case.

00:17:02: I have, i've written about the CFO and FPNA leaders as the AI entrepreneurs.

00:17:09: And I believe there are really strong use cases.

00:17:13: partially because The CFOs in the FPNA group they got access to more data across the enterprise than say other groups would.

00:17:24: So if I'm, you know we'll take sales planning If i am doing or well it will take supply chain.

00:17:31: You Know?

00:17:31: I might have limited data and just have supply chain data whereas The Office of Finance can get access to supply chain Data Sales Planning Data Inventory Data The hard finance numbers And more complete picture with your data the more likely you are to succeed with AI.

00:17:53: And so I think they're seeing a lot of good results in applying it, and a lot if also comes down again back to what we talked about is eliminating some of these manual processes.

00:18:05: So those were quick wins when it comes to AI.

00:18:08: That makes sense.

00:18:10: You mentioned advanced planning often application area off AI where companies do maybe new things or can do things better.

00:18:18: I wonder about enterprise planning in the term of that.

00:18:21: For quite a while, we had to trend.

00:18:23: that said companies often are pretty sophisticated and financial planning.

00:18:28: but when it came to operational planning there was a lot of room for improvement?

00:18:32: And also the integration of both as big driver.

00:18:37: first question is do you still see that or our company's now pretty advanced with Enterprise Planning?

00:18:43: second question would be how does AI appear?

00:18:47: So they're still not there with the strategic planning and I believe Christian had a late in post recently on it that.

00:18:55: That their getting better with the core finance- aspects of planning.

00:19:02: but Where they can go with AI is that AI will enable planning to become more continuous, and not just in the office finance but across enterprise.

00:19:24: Scenario planning gets faster and richer.

00:19:29: In some cases there are technology solutions which enable almost real-time access of data.

00:19:37: so if you're thinking about maybe a coffee company That's looking at I want to do some scenario planning around.

00:19:49: These X number bags of coffee are going to stale out soon and how do i move them- when you have more real time deeper richer data AI is then gonna enable you know, product loss.

00:20:13: Yeah I mean that's a good example but it also shows maybe AI helps companies in general to use data and AI more operationally meaning with the direct influence on operational processes not just for controlling purposes or reporting purposes.

00:20:40: Do you agree?

00:20:41: Absolutely.

00:20:42: The challenges, that's been the promise for twenty-some odd years was they say the proof is in the pudding.

00:20:49: it's actually getting to the execution.

00:20:53: You have some companies who are so advanced.

00:20:56: For example there were a gas company I'm going my dates wrong but in early two thousands that was using operational data against one of our favorite planning tools to figure out when to bring their oil rigs back in the Gulf of

00:21:18: U.S.,

00:21:19: back online after Hurricanes Katrina and Rita, of where companies can go, but then on the other end at the spectrum you've got people that are saying I finally hit a wall with spreadsheets.

00:21:36: I need a CPM solution?

00:21:38: Where do i

00:21:39: start?".

00:21:40: That's a big spectrum in between bringing oil rigs back online to...I need something and-and I would posit that AI is going adoption of CPM so that companies can start to narrow the gap between leaders and laggards.

00:22:00: Kelly, you have now been on board with Bach for quite a while!

00:22:03: You've taken part in lot our research studies... I would like hear from you about challenges that are running into.

00:22:13: what does data say or do latest study says?

00:22:18: What challenge is companies reporting And I think that's always interesting to talk about, because it is something we can tell people this what you should look out for.

00:22:29: These are problems others have run into so probably want make sure they address these and aware of them Can give us an overview?

00:22:41: The biggest challenge or the data.

00:22:45: whether looking at North America a lack of trust in the data and AI results, and data quality.

00:22:55: And our surveys are saying fifty percent of users just don't yet trust the data to go where they want to go.

00:23:03: that's the biggest challenge.

00:23:06: I'm gonna show how old i am.

00:23:08: I'm dating myself now but back in um two when Bill Clinton was running for president one of his chief strategists James Carville had written on it's the economy stupid.

00:23:22: And I think, The biggest challenge facing end users today in CPM is data and i would argue that every cfo an fpna leader should have on their whiteboard.

00:23:35: It's the data stupid um...and we can our studies get into more granularity.

00:23:43: I'll give you a little preview of our upcoming trend report.

00:23:46: We actually break out data management with Data Security, Privacy and Transparency but it all comes back to the biggest challenge is solving the data riddle.

00:24:00: until you get that solved You're not going be as successful with AI As you would want.

00:24:08: So really focus on that.

00:24:09: That's your recommendation idea?

00:24:11: Absolutely!

00:24:11: That makes a lot of sense, so first fix the foundation or make sure you can trust your data.

00:24:18: Let us look at other side of coin when there are challenges and also opportunities.

00:24:23: What do we see here?

00:24:24: what is the opportunity companies maybe want to pursue Or they would like to see When they use AI in CPM?

00:24:36: You know, there's lots of opportunities again for agents running off and doing all your hard work.

00:24:41: And then having a four day work week.

00:24:44: but I think what the big opportunity is that you're data foundation becomes a strategic differentiator?

00:25:05: if your data foundation is buttoned up and where it should be, that then you look much better.

00:25:14: And you're in a better shape to be successful.

00:25:16: I think the other thing we are going see governance as a differentiator.

00:25:22: As AI takes on more planning tasks The governance of the data will be important.

00:25:31: That's big opportunity.

00:25:35: And lastly, I think that as much there's an opportunity.

00:25:38: I think companies should show themselves a little grace.

00:25:42: you know North America.

00:25:44: at the moment we're leading in experimentation and CPM Global maturity is gonna value going to vary and frankly The maturity of how your using AI it's Gonna vary company by company industry-to-industry.

00:26:00: You know one of the things were seeing which?

00:26:02: Obvious when you think about it is SaaS-based businesses are moving more quickly to adopt and deploy AI than other companies.

00:26:11: I mean, that makes sense But there's a lot of opportunity And i think It's a great time To be in the space because things Are changing so quickly.

00:26:21: I completely agree.

00:26:22: We touched on the topic of agents several times now during this conversation, i think if we want to take a look into the future.

00:26:30: so what will be next?

00:26:31: What do you expect as a market analyst to happen in the next years In that market?

00:26:38: for sure Agents are very high on the list.

00:26:41: but before we talk about That I'd like You To give maybe us an overview.

00:26:46: Where Do You See Agents at Work Today?

00:26:50: already in CPM software.

00:26:52: What tasks or topics are there practically solving?

00:26:57: So that's an interesting question, and when I think about the CPM market frankly i tend to focus more on the planning side than the consolidation side.

00:27:08: but the truth is we're seeing a lot more.

00:27:10: um agentic successes.

00:27:17: monthly quarterly annual close, those are repeatable manual arduous not fun tasks.

00:27:26: And again with like the example of Unilever using AI and agentic AI to clean up their data.

00:27:35: this is where we're seeing real wins that they can put agents create workflows in the closing consolidation process, take care of all that manual check.

00:27:47: You know?

00:27:47: The same checklist you do every month.

00:27:49: your agent can go off and do it And then that gives more time to actually analyze the data and understand It!

00:27:57: This second area we're seeing is related to that which is anomaly detection.

00:28:03: Agents can much more quickly figure out this doesn't look right?

00:28:08: And then you have the time to figure out, why does it not look right versus... You spending all that time trying to figure where the needle is in a haystack.

00:28:22: Your agent can easily say oh there's that needle!

00:28:27: What are going do about?

00:28:28: But if your busy try and find the needle-you don't have any time for anything

00:28:32: else.

00:28:32: Yeah makes lots of sense.

00:28:34: That sounds like bright future.

00:28:35: So what's your take?

00:28:37: Where are we heading with this?

00:28:38: AI is getting more powerful by the day, it seems.

00:28:42: Um so obviously where have even more capabilities in the future?

00:28:47: how would that affect CPM or other future challenges and opportunities?

00:28:53: especially do you see?

00:28:54: I

00:28:54: think presents a great opportunity for people within the finance function As long as few is.

00:29:04: ten years ago Finance CFOs of P&A leaders were still seen.

00:29:09: It's just being counters those people in the corner office that just care about The numbers and the money.

00:29:15: Now their role has changed Incredibly quickly, you know it's no longer Just about spreadsheet skills But about the ability to understand AI.

00:29:27: And how do use AI within the finance function?

00:29:32: how to get AI to make you, the finance executives smarter more agile quicker.

00:29:41: And I think that is actually a tremendous career opportunity, career evolution.

00:29:53: I now have the right data, the trusted data, trusted outcomes to be a strategic advisor on the enterprise level.

00:30:03: That's great outlook!

00:30:05: Thanks so much Kelly.

00:30:07: i really enjoyed our conversation.

00:30:09: it was great to hear what is going in North America and the trends you are seeing And with that, I can only say thank you for supporting BARC in North America.

00:30:20: Obviously i hope to see and speak to you soon again!

00:30:24: Bye bye.

00:30:25: Thank

00:30:25: You so much Carsten.

00:30:26: if you get to Boston ,I'll get ya out on the river.

00:30:29: That sounds great.

00:30:30: let's do it bye-bye.