Latest / CXOTalk / AI Agents in Finance with HPE's Chief Financial Officer (CFO) | CXOTalk 914
Transcript
- 0:00Cracking the code on determinism was actually the hardest part of
- 0:05the journey that we embarked on. We actually Co engineered with
- 0:08NVIDIA on a NYM to help crack determinism because you just
- 0:13can't get finance data wrong. Every company says AI creates
- 0:17value, but you can prove it in the numbers.
- 0:21Marie Myers is Chief Financial Officer of HPE, where she's
- 0:25deploying AI agents across a 3600 person finance
- 0:30organization. Marie, how do you assess the
- 0:33value that generative AI brings to a business?
- 0:37As a CFO, you know, I spent my life and my career really
- 0:41assessing value. When you look at AI, it is a
- 0:44little bit different, Michael, I've got to say, I think you,
- 0:46you have to look at it through a much more broader aperture and a
- 0:49much broader lens. So I tend to look at a set of
- 0:52discrete and sort of more subjective variables.
- 0:55I tend to break it up into what are the direct impacts that I
- 0:58expect AI to have, but then what are those indirect ones as well?
- 1:01And sometimes what you find, Michael, is the indirect may
- 1:04sort of actually be much greater than some of those initial
- 1:07direct ROI benefits. So you have to step back and
- 1:12take a much broader sort of thought process around how you
- 1:15think about ROI on AI. Can you elaborate on that?
- 1:20Because the the value of AI doesn't necessarily show up in
- 1:27the hard numbers. First of all, I'd say is
- 1:29ultimately the objective is that you will see a return on your
- 1:33investment. So I don't want to start there.
- 1:34You know, you do want to see, you know, what you spend, you're
- 1:37going to get a discernible, you know, sort of outcome from that
- 1:41investment that you make. But I, I think what I've tended
- 1:45to do, and I've, I've done this over the years with a lot of
- 1:47automation and sort of early projects I was involved in.
- 1:50You know, we looked very discreetly at the direct, which
- 1:53is relatively easy to put hard numbers around.
- 1:57Some of the the areas that are a little bit more difficult to
- 2:00really put the hard numbers around are things like speed,
- 2:03like AI has a huge impact on the speed and just the efficiency of
- 2:08how you actually get work done. So you have to find a way to
- 2:11sort of say, OK, if I'm going to introduce AI or a genetic AI
- 2:16into my organization, what benefit am I going to get in
- 2:19terms of that sort of overall process?
- 2:21Can I do my work faster? So I would use that as one
- 2:24example of where, you know, you'd look at speed and try to
- 2:28assess what impact would that have quantitatively or
- 2:32qualitatively on your organization.
- 2:35Other areas that, you know, look, I think about particularly
- 2:37as a finance professional are areas just like accuracy and
- 2:42errors, right? You know, over the course of
- 2:44many, many years, a lot of finance work has been really
- 2:46transactional. Humans have done a ton of fat
- 2:48fingering on invoicing. And let's face it, we've always
- 2:51had error rates. So can AI help you reduce those
- 2:55error rates? Can AI help you reduce fraud?
- 2:58You know, all of that? I put a little bit more in the
- 3:00indirect bucket, but you can put numbers around that stuff,
- 3:02Michael. So you're looking both at the
- 3:06hard hard direct numbers as well as a set a a relatively broad
- 3:14set of indirect as you just to use your term.
- 3:18Absolutely. And you know, what we do is we
- 3:21actually sort of define that indirect bucket so that we can
- 3:24help to, as we see initiatives and projects come up and we're
- 3:28going to have to take some place, some bets, we can use
- 3:30that framework to assess how that project really plays out.
- 3:34And sometimes look, you know, sometimes some projects are
- 3:36going to be, they're going to have a better scorecard and, and
- 3:40others you may get into it and find out, hey, what I thought I
- 3:42was going to achieve. I'm not going to.
- 3:44So you want to know that too at certain stage gates along the
- 3:47way so that you can decide whether to keep going or
- 3:51actually to stop the investment because I mean, there are going
- 3:53to be some false starts in in this type of work as well.
- 3:57To what extent is your evaluation of the AI value the
- 4:04same or different from traditional more well known
- 4:08technologies, whether it's ERP or CRM or or other areas that
- 4:12are just more well established? I spent about a decade working
- 4:16in in areas like RPA, which is robotics, process automation and
- 4:21sort of ML and then evolved into AI.
- 4:24This framework really came from when I first sort of started out
- 4:28my career as a global controller and then as a CFO, where I
- 4:32started working on implementing bots into a lot of the
- 4:36transactional side of finance. And what we needed was a way to
- 4:40assess how, you know, the expense of the licenses,
- 4:43frankly, and the impact that they were having on the
- 4:45organization. So we came up with this
- 4:46framework and then I went on actually to be the CFO of a
- 4:49Philly large RPA startup and it became clear that the industry
- 4:54needed a framework to really help, you know, CF OS like
- 4:58myself get comfortable with these investments.
- 5:00So use that framework over the course and evolved it over the
- 5:03course of about of a decade. So it wasn't something I picked
- 5:06up in a textbook and sort of ran with Michael.
- 5:08It was really something that evolved through our PA.
- 5:12Now, I've, I've been on a number of ER PS and I would say that
- 5:15when we've done, you know, ERP investments, it's not that much,
- 5:20it's not that dissimilar. I mean, you know, what we do
- 5:22with these gigantic capital investments that you make on the
- 5:25IT front, you'll always have, you know, a very prescriptive
- 5:29sort of ROI bet that you make right out front, which may
- 5:32include, you know, I'm going to have obviously a much higher
- 5:36productivity rate. I'll see these processes get
- 5:38better. But sometimes on those big
- 5:40initiatives, it's really hard to see the benefits because they're
- 5:43over, they're multi year journeys.
- 5:44They're not necessarily journeys that are very short and agile.
- 5:48So I think that's the difference.
- 5:49Some of these IT programs take years to see these productivity
- 5:54benefits sort of, you know, really play out.
- 5:56Whereas with AI, you can see those productivity benefits and
- 6:00those indirect benefits that I mentioned much faster and much
- 6:03sooner. So that's why I've always been a
- 6:05fan of RPA. And as it evolved now into
- 6:09agentic AI, you can see benefits on a much more rapid, rapid
- 6:14pace. That's really interesting.
- 6:16So it also has implications for the IT organization and the
- 6:24nature of the value that IT delivers back to the
- 6:29organization. Even though there is more, can
- 6:33we say uncertainty with generative AI?
- 6:38Because it's not, it's probabilistic, it's not
- 6:40deterministic. So at the end, you don't really
- 6:43have the certainty of where it's going to end up in the way that
- 6:46you would say implementing a CRM system.
- 6:49This is a team sport. You need everybody at the table
- 6:52here. You need to have the business
- 6:54leaders with the business that put the acumen around the
- 6:57process. You need your IT team.
- 6:59They are the folks who are typically the custodians of a
- 7:02lot of the data. You need your compliance teams.
- 7:05You need everybody sort of playing their role and being
- 7:08part of a successful, you know, AI sort of driven initiative in
- 7:13an organization. So that that's sort of how I
- 7:16thought about it, which is much more of a collaborative
- 7:18engagement model. That's where this works best.
- 7:20And then to the second part of your question, which is about
- 7:23determinism, it's a subject near and dear to my heart because
- 7:27myself and a lot of my team have worked on a platform that we
- 7:30collaborated with Deloitte on called Zora, we call it Alfred
- 7:34at HPE. And one of the problems we had
- 7:37to tackle and solve right up front was exactly the issue you
- 7:41raised around determinism. Because as a finance
- 7:45professional, you know, if I was to embark on a journey where we
- 7:48would use AI in a broad enterprise capacity, it's simply
- 7:52not good enough to have a deterministic answer, which as
- 7:55you know, means I can't, you know, sort of use a model, ask
- 7:59it a question. And there's no guarantees that
- 8:01every time I ask that question, which is, hey, what was the
- 8:04revenue that the company shipped yesterday, that number is
- 8:07different. That number needs to be correct
- 8:10irrespective of who asked the model where in the world it it
- 8:14needs to be, you know, it cannot be probabilistic, it needs to be
- 8:18deterministic. So for us going through that
- 8:20Germany of really calibrating, you know, the right answer from
- 8:24the model every single time was really important.
- 8:27We had to, to work with a video on their names to really move
- 8:30from probabilism into determinism, which was a very
- 8:34important journey for us. And we, we sort of found that as
- 8:37being literally foundational to the work that we've done on AI.
- 8:40We have some questions that are coming in, and if you're
- 8:45watching on LinkedIn, pop your question into the LinkedIn chat.
- 8:50If you're watching on Twitter X, use the hashtag CXO talk.
- 8:56And I urge you to ask questions because Jen, truly, when else
- 9:03will you have the opportunity to ask Marie Myers, the Chief
- 9:07Financial Officer of HPE, you know, pretty much whatever you
- 9:10want. So ask your questions, take
- 9:12advantage of this. Marie, you mentioned Deloitte,
- 9:16just coincidentally, our last episode was with Bill Briggs,
- 9:20Deloitte's Chief Technology Officer, and he was telling us
- 9:25about your AI deployment. But give us a sense of the, the
- 9:31scope of that deployment. Give us some context.
- 9:34And then we have questions that are coming in and we'll we'll
- 9:36jump over to those. We formed a great partnership, I
- 9:40would say, with Deloitte on this journey.
- 9:41We decided to use the Zora platform, which Deloitte has
- 9:46really been at the sort of forefront of developing to help
- 9:49us, you know, start from what I would call the transactional
- 9:53side of finance. So we looked at how we could
- 9:55implement Zora's capabilities and really Dr. efficiency in
- 9:59speed areas like accounts payable.
- 10:02And then we sort of took out a challenge to take that platform
- 10:07and really see how we could scale it into a core enterprise
- 10:12process, which was our operations called.
- 10:14And through that journey, we've called our platform inside a HPE
- 10:18Alfred, which is actually Batman's wingman because you
- 10:21know, we sort of see Alfred as our wing person frankly.
- 10:25So the OPS call, which we have every Monday at about midday
- 10:28around this time is probably one of the most important processes
- 10:31that finance in the company, you know, actually drives on a
- 10:34weekly basis and cracking the code on what was a very
- 10:37laborious, very sort of old school PowerPoint, you know,
- 10:40rear vision process and really taking a genetic AI capabilities
- 10:44and using that platform was one of the the key objectives we had
- 10:47over the last 12 months. And what was different about
- 10:50that journey and, and why we collaborated with Deloitte is 1.
- 10:53I wanted to work with somebody that I knew could be a trusted
- 10:56partner that took compliance and governance really seriously.
- 11:00And secondly, we partnered with Deloitte on an on Prem solution.
- 11:04I'm sure we'll talk about this. We actually decided in order to
- 11:07really retain the security and the compliance we wanted, we use
- 11:10what we call private cloud AI architecture, which is basically
- 11:13on Prem infrastructure to house the data.
- 11:16We have an interesting question from Twitter.
- 11:21This is from Chris Peterson who says how do you deal with the
- 11:25non determinism and lack of explainability when you apply
- 11:30Gen. AI in a financial context.
- 11:33I think this is the question that we we all want to know.
- 11:36Cracking the code on determinism was actually the hardest part of
- 11:42the journey that we embarked on in really calibrating Alfred the
- 11:46platform. And it took time and in fact, I
- 11:49would say we actually Co engineered with NVIDIA on a
- 11:53nymph to help crack determinism because you just can't get
- 11:57finance data wrong. It like as a CFOI need to be
- 12:00before we could actually even trust Alfred.
- 12:04We needed to be sure that if we engaged in the surgenic AI
- 12:08journey that the prompts that we were going to use, we're going
- 12:12to generate the right outcome and the same outcome and every
- 12:15time. So we have about half a million
- 12:18data elements in Alfred and it is just amazing to me.
- 12:23I'm in Alfred almost every day now as we're getting ready to go
- 12:26live across our company testing it to see, do I get the same
- 12:31answer every time. And I'm pleased to say so far,
- 12:34the answer is yes, Michael. You have moved from using AI as
- 12:41an advisor into executing real financial work.
- 12:47As you expand into areas like accounts payable, credit
- 12:50collections, where do you allow the AI to act independently and
- 12:56where must humans people stay in control?
- 13:00This is such a fundamental agentic AI issue.
- 13:04I think you're kind of hit at the core of the the evolution of
- 13:08AI and specifically agentic AI. I'd I'd start out by saying
- 13:12we're still in the really early innings of the maturation of
- 13:17Gen. AI and a genetic AI in inside
- 13:20enterprises. You know, I'm an ex auditor.
- 13:23So I have a fairly, how do you say strong risk appetite.
- 13:27And I'm perhaps a willing to be an early adopter on
- 13:32technologies, but to do so in a very thoughtful, risk averse
- 13:36way. So for me the mantra has always
- 13:38been human in the loop. So we use AI to help us, you
- 13:43know, really drive faster outcomes, but there is a
- 13:47critical sort of well understood role for the human that we
- 13:50document as part of the process. So there's always a human in the
- 13:54loop in how we're deploying AI. So that's, like I said to you, I
- 13:59think we're at the early innings and the early stage of this
- 14:01evolution. I'm not to say how things are
- 14:03going to look, you know, maybe 2-3 years from now, but for now,
- 14:08you know, having that human engagement and the human
- 14:10involvement is, is super important, Michael.
- 14:13We have A and really another interesting question, this time
- 14:17from Nick Qureshi on LinkedIn. And Nick says, what are the top
- 14:25three use cases with the highest ROI through AI in your expert
- 14:30opinion? At an enterprise level, I think
- 14:32look the the top use cases that are clear out there today and I
- 14:35don't think there's any sort of dispute that these areas are
- 14:38widely acknowledged as having fairly extensive AI applied to
- 14:44them today. The first one is a lot of the
- 14:45software engineering, I've talked to a lot of board members
- 14:48who see pretty impressive productivity gains across
- 14:54different types of software engineering.
- 14:56So I think software engineering, code development is fast
- 15:00becoming probably the hottest area for the deployment of
- 15:04generate, you know, generative AI.
- 15:05The other one is marketing, So a lot of image creation, a lot of
- 15:10campaigns these days, you know, you're really seeing this sort
- 15:13of rapid expansion of generate AI in that space.
- 15:17And then the last one I'd say is IT itself, I mean just to help
- 15:20IT teams process a lot of the sort of work flows in IT.
- 15:24So those are sort of the three top enterprise use cases.
- 15:27And when it comes to finance, I would say it's still from an ROI
- 15:30perspective at the transactional layers.
- 15:32So wherever it's like the early days RPA, you know, a lot of it
- 15:36was all those manual invoices. We used to get a lot of that
- 15:40stuff. We had the bots processing.
- 15:42You know, a lot of that area today is where still you can see
- 15:44AI is having, you know, greater impacts and perhaps, you know, a
- 15:48lot of the RPA. So transactional finance,
- 15:50wherever there's a lot of routine sort of repetitive tasks
- 15:54is where you're seeing AI really have an impact in finance.
- 15:58And then it's going to scale, Michael, as you sort of nail a
- 16:00lot of those workflows and you get into more complex areas like
- 16:04forecasting you, you'll start to see those applications broaden
- 16:08out. Can you dive a little bit into
- 16:11the types of applications because I know these are AI
- 16:16agents and so maybe give us some insight into what kinds of AI
- 16:21alications these are? The areas that, you know, for
- 16:24finance that are critically important for AI, you know, I
- 16:28would say starting in the transactional world.
- 16:31And honestly, when I was a global controller, I had spent a
- 16:34lot of time with deploying RPA in these areas over a decade
- 16:40ago. And they're exactly the same
- 16:41areas coincidentally where you're seeing generative AI
- 16:44really lean in and take a very sort of, we have a high, high
- 16:49impact. And I would say it was accounts
- 16:50payable, a lot of the credit and collection areas and then and a
- 16:54lot of the traditional accounting areas, because this
- 16:56is a lot of sort of contractual analysis that needs to go on.
- 16:59That becomes frankly, you know, as a ex analyst, I used to do
- 17:03some of this work myself. It was taking a contract, going
- 17:06through the terms and just trying to apply it to an
- 17:08accounting standard. You can have the AI do a lot of
- 17:10that, that reconciliation today all the way into internal audit,
- 17:14whether it's controls checking. So a lot of that repetitive work
- 17:18that you saw financial professionals having to do in
- 17:20the past can, can frankly be done pretty easily by AI today.
- 17:24And what it does do, it allows, you know, a lot of people to
- 17:28free up time to actually do the analysis that you just didn't
- 17:31have time to do. If I look back on my first job,
- 17:33Michael, I mean, I love to tell the story.
- 17:35I was a, you know, I was an analyst in a tech company.
- 17:39It was my first job out of Graduate School.
- 17:41And I consider myself, I was back then a bit of a pivot table
- 17:44queen. I used to be able to take tons
- 17:46of data and build, you know, amazing sort of analytical mind,
- 17:51you know, analytical pivot tables around that data today.
- 17:55And that used to take me hours and hours.
- 17:56I can remember as a young girl, I'd be after 10:00 at night sort
- 17:59of building these pivot tables. I didn't have time to spend on
- 18:03the analysis to really determine what was happening with all this
- 18:06data that would take me well past midnight into the early
- 18:09hours of the morning to do. I like to think that today,
- 18:12because of AI, the AI can run a lot of preliminary analytics and
- 18:16all that data for you today. So you don't need to do what I
- 18:18did back 30 plus years ago. And I think as a professional,
- 18:23that's a better use of my time to spend my time analyzing those
- 18:26results and really interpreting the information as opposed to
- 18:29sort of pulling all that data together to even sort of start
- 18:33on the analysis. And then as a follow up, we have
- 18:37a an excellent question from LinkedIn and this is from
- 18:41Katarina Collin, Sarah, who is herself a global CFO and she
- 18:49asks this, have you seen any degradation in decision quality
- 18:54when teams rely too heavily on AI outputs and how are you
- 18:59mitigating that? There's been some sort of
- 19:01commentary out there in the market where folks call it
- 19:03colic, call it AI slop, which is where people get kind of lazy.
- 19:08But let's face it, let's put away the AI slop, but let's just
- 19:10get into work slop. I mean, it happens in work.
- 19:14People get, you know, they get comfortable when they do the
- 19:18same thing over and over again. And this look, with AI, people
- 19:21can use it as a sort of easy way to get work done.
- 19:24So your standards can't drop cater, you know, that's what I'd
- 19:27say. I see it with my own teenage
- 19:28kids, You know, they try to use the AI to get the homework done
- 19:31and they come and show me the homework.
- 19:32I'm like, come on, guys. You don't even answering the
- 19:34question that the essay was supposed to be written on.
- 19:38So you have to, you have to keep the standards irrespective of
- 19:41whether you're using AI, You're not using AI.
- 19:43You can't let you know, AI become an excuse for low quality
- 19:48work. And I think in the early initial
- 19:51days, there certainly was a fair bit of that going on.
- 19:54People are like, Hey, I'm just going to use the AI to write a
- 19:56memo. And if the memo is not really
- 19:58like, you know, we have to prepare memos for a lot on on to
- 20:01sort of summarize a lot of our analytical packs.
- 20:03And, and my staff were using them and I'd read the memo.
- 20:06I'm like, guys, this memo really, it doesn't even sort of
- 20:10summarize what's in the deck. You need to really read the, the
- 20:13quality of the AI memo to see if it's, it's any good.
- 20:16So that's just part of, look, forcing your own folks not to
- 20:20become too lazy with AI. You know, I think it's very
- 20:23powerful on data. I think on the memo writing and,
- 20:26and some of that where I see folks, you know, they, they kind
- 20:29of can get a little lazy sometimes.
- 20:31She has a follow up question and I, I, I love the questions from
- 20:39the audience. You guys in the audience are so
- 20:41awesome. Let me mention this would be an
- 20:44excellent time to subscribe to the CXO Talk newsletter.
- 20:48So go to cxotalk.com and subscribe so we can notify you
- 20:53about upcoming shows just like this because you guys in the
- 20:56audience, we love your questions and you should attend and join
- 21:00us. OK, so the follow up from
- 21:04Katarina is also how do you balance speed from AI with the
- 21:09need for human judgement and skepticism in financial decision
- 21:13making? You need to have that balance.
- 21:16Let's go back to what I said earlier about there has to be a
- 21:19role. I call it, you know what I've
- 21:21called it is human in the loop that there needs to be a piece
- 21:25of the equation where the human is involved in evaluating some
- 21:30of the AI output and applying some of that judgement.
- 21:33Because I, what I would say the AI today, I see it much more in
- 21:39an augmentation stage as opposed to sort of replacing a lot of
- 21:43the human judgement. I'll give you an example with,
- 21:45with Alfred where we're, you know, preparing to use Alfred
- 21:48for all of our operational calls.
- 21:51What I do like is that it will analyze all the data on the
- 21:55performance of the business over the last week and then it will
- 21:57be very prescriptive and giving me targeted areas of insight
- 22:02that require the next level of, of judgement.
- 22:05So I deliberately and consciously use the AI to help
- 22:11narrow the lens on issues in the company.
- 22:13And then I go back to my controllers, I go back to my CF
- 22:16OS and ask them to go and dig in and validate what really is
- 22:18going on. So in the past, I would have
- 22:20felt like I would have had to go looking for a needle in a
- 22:23haystack to figure out what the issue was.
- 22:25Today I can be much more surgical and pointed and say,
- 22:27look, guys, I see these three things coming through the data.
- 22:31Now please go and give me the insight or give me the plans of
- 22:35what you're going to do to go fix those issues.
- 22:36So I'm not relying today on the AI to replace and solving the
- 22:41issue. I'm relying on the AI to provide
- 22:43much deeper focus on the issues we need to go solve and then
- 22:47bringing the human, the benefit of human judgment to help solve
- 22:51the issues. And I think that's the piece
- 22:52that I'm still somewhat skeptical about AI to actually
- 22:57take all those years of experience and all that judgment
- 22:59and sort of, you know, replace me, frankly.
- 23:02Would it be accurate to say that you're using the AI to gather
- 23:09information, data and analysis to present to you, which you
- 23:16then or someone on your on your team then analysis or applies
- 23:21human judgement to? Is that?
- 23:24Yeah, I would say takes you synthesize and really provide,
- 23:28you know, clarity on what the issues are and then the decision
- 23:32on what to do is being driven by the human today.
- 23:35I mean, like I said, Michael, I think over time, you know, that
- 23:38will continue to evolve and we're talking about a much more
- 23:40complex workflow than just sort of invoice processing, you know,
- 23:45contract analysis, which I call sort of transactional workflows.
- 23:50I'm I'm moving up the stack in terms of more complex workflows.
- 23:54I use various AI tools everyday. So for example preparing for our
- 24:00discussion and it goes out and can research hundreds and
- 24:05hundreds and hundreds of sources.
- 24:08But at the end, when it comes down to our conversation, I have
- 24:15to check every fact, every every stat or fact it comes back with
- 24:21before I include it in a conversation like this.
- 24:24I verify it myself. I think that's a great example,
- 24:27Michael. You're doing exactly what I
- 24:29would think what we do, which is you're the human in the loop.
- 24:32Like you're using it to gather, you're using it to synthesize.
- 24:36But when it comes down to that final judgement call, you're
- 24:40making that. And that's what I meant when I
- 24:42go back to my first job, I mean, all that work, just gathering
- 24:46the data, synthesizing the data. My question for you is how long
- 24:49did it take you to do that without AI?
- 24:51For me the AI is drives time, efficiency and also breadth and
- 24:57therefore quality. Yeah, exactly.
- 25:00So I think it's another great example of, you know, where the
- 25:04power is at today. Now, I think what we've got to
- 25:06see is how it scales over time and that last mile of the
- 25:10judgement call that you mentioned that you're doing.
- 25:12I'm doing the same thing today in my office.
- 25:14This question is from AJ Sharma on LinkedIn.
- 25:17And he says, Marie, how do you ensure the agent is acting on
- 25:23trusted, reconciled financial data, not fragmented or or stale
- 25:29sources across systems? The most important part of any
- 25:35AI journey is data, data quality.
- 25:39Data cleanliness is the number one issue that I think holds
- 25:43back a lot of companies and organizations and really
- 25:46embarking on AI journeys without clean good data, you're really,
- 25:52it's impossible, frankly, almost Michael, to move forward.
- 25:55So what we did at HPE, for example, is that, you know, we
- 25:58worked, we spent, you know, countless amount of time really
- 26:03ensuring that we had the right data layer in place, but the
- 26:06right quality of data. So we were fortunate enough that
- 26:09we did have inside the company a, you know, a sort of data
- 26:12platform that had been largely architected and, you know,
- 26:17credibly built over a period of time.
- 26:19And that was the foundational layer that we actually used for
- 26:23the work that we did around Zorro and Alfred.
- 26:25So without that though, you you couldn't even get, you know, on
- 26:28the Monopoly board. You couldn't even begin to pass.
- 26:30Go, Michael. We have a question from
- 26:34Elizabeth Shaw on Twitter who says what are the uses and
- 26:39potential uses specifically of AI agents and agentic AI for
- 26:45FPNA financial planning and analysis.
- 26:47And she's also interested in the kind of autonomy that you allow
- 26:54the agents to maintain. For FPNA, you certainly my team,
- 27:00we, we made a conscious effort to actually first embark on a
- 27:04workflow transformation inside of FPNA to allow that team and
- 27:08to start to build and use agents.
- 27:10I think in FPNA there is just an enormous amount of opportunity
- 27:13on a lot of that basic block and tackling on data that we talked
- 27:18about earlier. So whether it's being able to
- 27:21build analytical views, you've got so many sort of easily
- 27:27usable agents inside products even like Microsoft Copilot
- 27:30today for example, they have a genetic workflows that are
- 27:33already there that FPNA teams can easily use with confidence.
- 27:38They actually have an Excel bot, except as an example, which
- 27:41helps you, you know, really rapidly enable and use Excel
- 27:45and, and helps FP and a teams frankly get their analytical
- 27:48work done. So a lot of progress made in
- 27:51whether you're a Microsoft customer and there are other
- 27:54sort of LMS out there as well that have all that sort of
- 27:56really pre configured agents out there that you can use.
- 27:59So I'd say FPNA is a great customer for Agentic AI.
- 28:04Marie, there was an article in Fortune that said you are
- 28:10positioning the Chief Financial Officer role as a steward of AI
- 28:15across the enterprise. When it comes to AI, how should
- 28:21ACFO work with other leaders such as the Chief AI Officer or
- 28:26CIO? It's a team sport.
- 28:28I think like we spoke about earlier, Michael, there is a
- 28:31really important role for, you know, a lot of key positions in
- 28:36the company, whether it's the IT partnership that's essential,
- 28:40the link back with the business that's essential.
- 28:42I would actually say the link with the controls and the audit,
- 28:46absolutely essential as a company embarks on really the
- 28:49early days of AI and governance and controls.
- 28:52But where I think the CFO has a unique sort of perspective is on
- 28:59the ROI conversation, which is actually where we started the
- 29:01call. Because you know, if you think
- 29:02about it, the CFO is really the sort of the steward of capital
- 29:07of the organization. So you can help the enterprise
- 29:10and the company really navigate the decisions and the
- 29:14investments in AI and then help them to evaluate those outcomes
- 29:19as well. Because I think this is an area
- 29:21that if you look at a lot of the literature that's been written
- 29:24about AI, it's perhaps the one area that most folks have
- 29:27struggled to really build a clear framework around.
- 29:32So I think there is a great role for the CFO to be able to be the
- 29:35steward of capital and help to really navigate ROI, you know,
- 29:40projects for AI. Is your relationship with other
- 29:45team members, other leaders at HPE, different when it comes to
- 29:51generative AI and agentic AI than with other areas of
- 29:56technology or or other areas in general?
- 30:00In some ways it probably is, Michael, because you know, in my
- 30:04team today, I have a group that's actually in my strategy
- 30:07organization, my true strategy officer helps to sort of drive a
- 30:11lot of the transformation initiatives in the company.
- 30:14And we obviously have transformation initiatives that
- 30:17the backbone is actually the deployment of generative AI.
- 30:20And these are examples of where the team is interfacing and
- 30:24working, you know, collaboratively with these
- 30:26organizations to help Dr. transformations where AI is the
- 30:30cornerstone of success. And that's a little bit
- 30:34different. And in the past, you know, I
- 30:36think in the past, we would have, you know, perhaps helped
- 30:40those organizations, you know, through ERP implementations,
- 30:43etcetera. And you would have stepped back
- 30:45and sort of said, look, here is the here is the plan.
- 30:47You know, like the typical finance person, you know, I'll
- 30:49check off every quarter to see whether you've made progress
- 30:52against your plan and whether or not you've hit your milestones.
- 30:55But what's different about AI is it's much more agile.
- 30:58So one of the initiatives were working with our supply chain
- 31:00team, for example, involves really deploying generative AI
- 31:04in their customer organization, helping them really manage your
- 31:09customer service all around the world.
- 31:10And so as the team has been deploying generative AI, they've
- 31:13become very successful and they've found new and different
- 31:16ideas that emerged even sort of six months into the initiative.
- 31:19So in order for my team to really help them evaluate
- 31:24whether or not these investments are paying off, you have to be
- 31:26really close to that organization.
- 31:28So you have to understand the work they're doing, the success
- 31:31they're having, some of the obstacles they're hitting.
- 31:34Whereas in traditional sort of IT projects, you just want to
- 31:37have that level of agile engagement.
- 31:39So I think the agility of of deploying AI means that it's a
- 31:44different, different engagement model than what we've seen in
- 31:47the past. So we're typical IT projects.
- 31:50We do have quite a lot of CIOs who watch CXO talk.
- 31:57And I'm wondering how does Jen AI and everything we're talking
- 32:04about change the role of the CIO?
- 32:10And, and obviously you're looking at it through the lens
- 32:12of being a Chief financial Officer.
- 32:14But any thoughts on that? And also any advice for CIOs who
- 32:20are trying to navigate this difficult transition period from
- 32:25before AI to we don't really know where it's going exactly?
- 32:29I think this is a great opportunity for CIOs to, you
- 32:33know, really become really sort of have a pivotal role in the
- 32:38company. I think CII always look at it as
- 32:41a super important role for the company because technology is
- 32:46the backbone and sort of is the foundational ingredient for a
- 32:51lot of companies. And so the role of the CIO to me
- 32:54has always been one that's essential and important.
- 32:57I just think today for CIOs, you know, they need to be, you know,
- 33:01right at the table and have a really important seed around AI
- 33:05because they are typically the custodians of a lot of the data.
- 33:08They understand the financial architecture.
- 33:11They are driving that financial architecture in the company.
- 33:14So they have to play a really important role in controls and
- 33:17governance. So the relationship between the
- 33:19CFO and the CIO, you know, kind of close out there and say it's
- 33:22a really important one. Katarina Collins, Sarah has yet
- 33:27another question. She's asking great questions.
- 33:30She's ACFO and I, I love she's bringing that insider knowledge.
- 33:35And she says, as AI reduces the need for data collection and
- 33:39consolidation, how are you redesigning the finance
- 33:44organization and redefining the core capabilities you expect
- 33:48from your team? You know, as we sort of prepared
- 33:51for some of the, the generative AI and the genetic AI work we've
- 33:55been doing, we focused very intentionally on actually
- 33:59redesigning some of the critical processes and in the finance
- 34:03organization. And I'll, I'll use FPNA because
- 34:05we talked about it a moment ago as one of those examples.
- 34:08So we actually decided to pull all the FPNA team together under
- 34:151 central organization and then look end to end across their
- 34:19work and try to drive a much more standardized approach to
- 34:24how we get the work done so that then we could actually layer on
- 34:27top a lot of the agentic AI or generative AI capabilities.
- 34:32So we were deliberately, let's just say deliberately went about
- 34:38the workflow 1st. We, we didn't just apply the AI
- 34:43without sort of redesigning the operating model and then looking
- 34:48at the workflows themselves and then figuring out where in that
- 34:52workflow could we then leverage the AI so that we would have a
- 34:56much more sort of one. One thing I really think is
- 34:59super important as the CFO is you want to have, you know, data
- 35:04one time. You want one time data, you want
- 35:06it to be the voice of the truth. You don't want to have the
- 35:08multiple. A lot of companies struggle
- 35:10with, I've got a version of a, you know, I've got a report that
- 35:13comes from one team that contradicts another team.
- 35:15So you want this one source of the truth.
- 35:16So for us, the sort of guiding mantle was one source of the
- 35:20truth and that helped us drive this whole transformation that
- 35:23we're trying to drive with with the FP and a organization.
- 35:26One of the things I find most interesting about this is you've
- 35:31got this technology, generative AI, that is driving change
- 35:38through many different parts of the organization.
- 35:44The tentacles are extending in in so many different directions.
- 35:48Absolutely. You know, it is certainly areas
- 35:51like software marketing you, you're starting to see agents,
- 35:55you know, working in all different sort of areas of
- 35:58enterprises. I think the real question is how
- 36:01is this all going to work together?
- 36:02I I don't quite have a good clear answer for you yet.
- 36:05That is something I personally need to spend time on, which is
- 36:09how does this sort of system work together in the future?
- 36:14Let's talk about capital. AI demands enormous capital
- 36:20commitments in an environment where the technology evolves
- 36:24faster than the investment cycle.
- 36:28Models become obsolete in months, infrastructure costs are
- 36:31unpredictable, and returns often appear as speed and agility
- 36:37rather than hard savings. How do you navigate this AI
- 36:42investment minefield and place large consequence consequential
- 36:48bets on a target that keeps shifting?
- 36:52You really have to experiment, Michael.
- 36:53I mean, you cannot sit here and wait for the ultimate Nirvana to
- 36:57emerge because as I mentioned earlier, AI is iterative.
- 37:02So you kind of have to be prepared to learn as you go.
- 37:05It's a much more agile way of working.
- 37:07I, you know, worked in a company, you know, worked over
- 37:09the decades, many of us were using sort of very waterfall
- 37:12approaches to how we got work done, how we did, you know, IT
- 37:17sort of, you know, big ITR plea implementations today with, with
- 37:21AI is agile. You're going to get better as
- 37:24you go. I'd say on Alfred, for example,
- 37:25I think we're on version 65 plus.
- 37:28So you know, if we'd started this journey a year ago, we
- 37:32might have taken a very different approach to even how
- 37:34we're taking this today. So, but what was really
- 37:36important was the learning that the team went through and the
- 37:40expertise and the knowledge that they had today.
- 37:41If I had lost that whole year, I had to start from Ground Zero
- 37:45today. So I hear you, I and there is so
- 37:49much change, but if you don't develop a sort of course that is
- 37:52skills and competencies and a learning capability, then you're
- 37:56going to just so far behind a year from now.
- 37:58So I, I do believe it's important to jump in and learn
- 38:02and acknowledge that it's going to change and there's going to
- 38:05be some throwaway work along the way.
- 38:07Therefore, what is your tolerance for supporting AI
- 38:12initiatives before they before you expect that they deliver a
- 38:17measurable direct return? I suppose I do acknowledge I and
- 38:22I do that right up front. I acknowledge there's going to
- 38:24have to be a, an investment stage that we're going to have
- 38:29to make. And that's why I mentioned
- 38:30earlier, you know, some of these investments you'll, you'll make
- 38:33and then you'll realize a few months in this is not going to
- 38:36pay off. So you have to pull the plug and
- 38:38then redirect that capital elsewhere.
- 38:41So you have to be very vigilant in monitoring success and
- 38:46outcomes. And I think that's, that's why
- 38:47I'm going back to the same point about agility and the ability to
- 38:52be able to pivot and make different determinations.
- 38:55And you know, we have frankly in finance, I'm just going to talk
- 38:58very selfishly about the work I've done with my own team.
- 39:00I mean, we tried, we made some, we tried some experiments, some
- 39:03of them didn't work, Michael. So we pulled the plug and we
- 39:06went and made a different investment elsewhere.
- 39:09So we have been, there's not hasn't been a prescriptive road
- 39:12map out there. We probably went through some
- 39:14uncharted territory and it was a mindset that I did have to make
- 39:19some investments, but those investments were not, they
- 39:23weren't particularly sizable. They were manageable, I would
- 39:26say, in terms of my expense envelope.
- 39:28Across the industry, major players are financing one
- 39:33another's growth. Chip chip makers are funding
- 39:37model firms. Cloud providers are funding the
- 39:40startups that rent their compute as ACFO.
- 39:45How do you determine whether the demand you see is sustainable
- 39:51and based on real economic value or just financial maneuvering
- 39:58through this circular financing as it's called?
- 40:01I'd certainly say it's definitely been in the news
- 40:04lately, Michael. You know, as a company, as a CFO
- 40:07of HP, we're really focused on customer value and you know,
- 40:11that's sort of the North Star and the lens that we really use
- 40:15around our business. What about the implication for
- 40:20change and processes? You touched on that, but this is
- 40:27such a a core point and also the culture and the implications of
- 40:32AI on culture change. I believe that the whole area
- 40:38around management of change is probably the hardest part of AI,
- 40:42which is not, you know, with all that we spend so much time
- 40:46sometimes on the technical side of implementation and the pace
- 40:51that doesn't get the attention it deserves is the human side of
- 40:55change. You know, let's face it, as
- 40:57humans, Michael, we're somewhat, you know, many of us are
- 41:01resistant to change. And so taking an organization
- 41:03through a journey like this, I got to say it's no easy feat.
- 41:07And you've got to go into this with, one, acknowledging that
- 41:11that is going to be part of the journey.
- 41:14And two, you have to be very deliberate and intentional from
- 41:18sort of day one around how to really drive the mindset change
- 41:24of your organization. So, you know, it's interesting
- 41:27there is more being written I think today and and
- 41:29acknowledging that this cultural sort of empowerment from AI in
- 41:34an organization is really a big part of what's going to
- 41:38determine success. And I'd almost argue, Michael,
- 41:40you go back and read some of those early case studies about
- 41:43there were so many AI pilots and they all failed.
- 41:45What didn't get told I think was the whole story, which was did
- 41:49those organizations actually really start intentionally from
- 41:54the beginning with a whole sort of approach around management of
- 41:57change in their organization? I would say I was very fortunate
- 42:00because when I did a lot of early days of RPA
- 42:03implementation, I worked with some professors and we worked on
- 42:06all these theories. And actually we even wrote a
- 42:08book about it, Michael, on how to drive organizational change
- 42:13with, with technologies like RPA and ML.
- 42:15And I've used a lot of that learning today for, for AI,
- 42:18frankly. Those studies there, there was
- 42:21that one in particular from MIT that described pilots and how of
- 42:29the vast. Failed pilots.
- 42:31Yes, failed pilots. But I think the other problem
- 42:34with that particular study is the nature of pilots is they are
- 42:39designed to learn and to experiment and they will fail.
- 42:44I mean, I think you alluded to that yourself and as a CFO, it
- 42:49sounds like you're willing to accept a certain amount of that.
- 42:53I think you have to go into AI with that mindset from the
- 42:58beginning that you're going to experiment, learn.
- 43:02And particularly because we're still in the early days of sort
- 43:05of the maturation of AI, there's going to be lessons learned that
- 43:10will have big payouts and then there's going to be some that
- 43:12will not materialize. But I would say that even
- 43:15starting on the journey and maybe some of those pilots too,
- 43:18Michael, I'd say you would want to look back and see maybe they
- 43:21failed because they didn't really work on how you're going
- 43:25to address workflows in process. Because that part of the change
- 43:29is just as important as the technology itself.
- 43:31Like the, the discussion we had about my FP and a team going
- 43:35through redesigning that workflow, centralizing the, the
- 43:38actual work itself helped us then deploy agents much more
- 43:43precisely into the organization. So if you were just doing a
- 43:45pilot, you deployed an agent and nobody used it, well, you could
- 43:48say the pilot failed. We have another question from AJ
- 43:53Sharma on LinkedIn. You can see I prioritize the
- 43:57questions that come in from the audience, and I encourage folks
- 44:01in the audience. There's still time.
- 44:03We have a few minutes left, so ask your questions.
- 44:06AJ Sharma says. To what extent have you
- 44:09entrusted regulatory interpretation to autonomous AI
- 44:13systems, and what governance mechanisms ensure those
- 44:18interpretations remain accurate and compliant across diverse
- 44:23jurisdictions? To me, this is the same core
- 44:27issue of how to what extent do you rely on the AI to give you
- 44:34crucial information that's accurate.
- 44:37We help the AI in terms of we use the AI to help us speed up
- 44:41gathering the data or gathering the facts and one area black
- 44:44the, you know, text helping you collect a lot of the data that
- 44:46you need to make an out and me to determine an outcome.
- 44:49And when it comes to sort of the judgmental side, the human is
- 44:54absolutely engaged in ensuring that they're reviewing the
- 44:57information, particularly regulatory.
- 45:00I have a sort of high bar on a lot of regulatory interpretation
- 45:04because that is a bit more grey and that is absolutely where,
- 45:08you know, we may use the AI to help us speed up the analysis,
- 45:11but the human is then obviously determining the final outcome.
- 45:15So at the end of the day, accountability remains with the
- 45:20responsible person. Absolutely.
- 45:23I mean, you know, we're not at a point in this journey today
- 45:26where you can devolve accountability.
- 45:28And I, I look at my role, I mean as CFO, I'm accountable for, I'm
- 45:32the financial steward of the organization.
- 45:34So, you know, I'm accountable for those financial outcomes,
- 45:37Michael. And we have a question from
- 45:41Swami Viajianathan. And Swami says as you approach
- 45:47this initiative iteratively and have to start again in some
- 45:52cases, as you were just describing our teams, building
- 45:56organizational context, not just technology, but the organization
- 46:02and that context must be valuable from an overall AI
- 46:06transformation journey. But how do you put a number to
- 46:10that from a financial standpoint?
- 46:12Because the investment is truly valuable and you don't want to
- 46:16throw it away? Organizational learning agility,
- 46:19because what you're doing is actually learning through, you
- 46:24know, perhaps through failure. And that means that the next
- 46:26time you try to approach another implementation, you're probably
- 46:31going to go a lot faster. So learning agility is super
- 46:34important. And actually Michael, I mean,
- 46:36I'll be honest, I spent myself and my team spent a lot of time
- 46:39over the last couple of years literally training more than
- 46:423000 people at an organization. And we didn't sort of go out
- 46:45there, educate people once and then walk away and say hey,
- 46:48you're done. No, it is a constant process
- 46:51because things are changing as we talked about.
- 46:53So raising the bar on knowledge and sophistication on knowledge
- 46:57in your organization is like building.
- 47:00It's like building your house. You're just building that
- 47:02foundation and it's rock solid. So it means, you know, you're
- 47:06like an athlete, you're getting faster and faster each time you
- 47:09actually try to implement a different, you know, a different
- 47:13initiative or a different project.
- 47:15Marie, can you talk about the CFO role and how CFOs can and
- 47:24should in your view, relate to all of these changes that are
- 47:28going on? Your background is highly
- 47:32technical and unique. You work for an RPA company, but
- 47:38the average CFO doesn't have that benefit.
- 47:41So what should, what should ACFO do?
- 47:44Who's listening and knows they need to get up to speed with on
- 47:48AI, but they're not sure the best approach.
- 47:50The role of the CFO is certainly evolving from, you know, what
- 47:55has been a very much A1 where we were the guardian of financials,
- 47:59institute of capital to now having a really important role
- 48:03in AI inside a company. So, you know, sorry for myself,
- 48:07I think the, the background, the experience I gained over the
- 48:10course of the last couple of decades in tech has been, you
- 48:14know, I've been very grateful for having that background and
- 48:17it has helped me I think become a much more earlier probably it
- 48:21helped me become an earlier adopter of a lot of technologies
- 48:23like agentic AI. And I would encourage CF OS
- 48:26today. I think it's you, you can't just
- 48:28rely on your financial acumen. You need to use that plus your
- 48:32technology acumen to really help guide the enterprise because CF
- 48:37OS are becoming strategic business partners.
- 48:39And this is going to be one of the probably top three agenda
- 48:43items or top three items for any company in the next few years.
- 48:47So if it's in that list as a CFOI think you really need to
- 48:51develop expertise and acumen. And the good news, Michael, is a
- 48:54lot of things you can do to educate yourself part.
- 48:56What I would advise CF OS to do is, you know, learn as quickly
- 49:00as you can. The best thing you do is learn
- 49:02by doing sometimes. I mean, this weekend I'm hoping
- 49:04to go play with with Openclore. I ordered myself a Mac mini to
- 49:08sort of play at home, teach the kids.
- 49:11So I would encourage you to to learn by doing and experimenting
- 49:15and then, you know, really figure out what role you can
- 49:19play inside the enterprise. What can you do to help some of
- 49:22your business partners? What can you do to help the CIO
- 49:25in the company? So find a role for yourself as
- 49:28well where you can learn and grow and therefore, you know,
- 49:31build the acumen that you're going to to need in your role
- 49:34because it's going to be one of the top items.
- 49:36It's certainly a boardroom topic, Michael.
- 49:38Today I sit on boards and it's in the boardroom.
- 49:41Board members want to know how companies are really deploying
- 49:45and using AI in the company. So you you're going to get asked
- 49:48the question, so you want to get ahead of it.
- 49:50How technical does ACFO need to be today in order to be
- 49:55successful in this world? All that experience that you've
- 49:58had around financial acumen is it's bread and butter, it's
- 50:01absolute table stakes. But developing a level of
- 50:04technology acumen is going to become increasingly important,
- 50:08Michael, because boards are going to, you know, they're
- 50:12going to ask those questions and ACFO is going to help to help
- 50:16guide some of those answers. So developing a little bit of
- 50:20technology acumen, I'll just say a level of tech acumen that is
- 50:25relevant is going to become increasingly more important.
- 50:28Could we say conversational tech acumen?
- 50:32I think conversational, but I actually I'd push it, I'd say
- 50:36understanding. I don't think it's just being
- 50:38able to have a conversation about it.
- 50:40I think understanding like the question you asked me, like
- 50:43we're in the enterprise is, you know, a Jenik AI really having
- 50:47an impact. I think it's important for ACFO
- 50:49to understand that and be able to therefore guide the company
- 50:52and those investment decisions. Obviously, if you don't
- 50:56understand the logic and the dynamics behind it, how can you
- 50:59be effective that's guiding the company.
- 51:03It's going to be very difficult and then the board is going to
- 51:05ask you about have you, can you realistically, you know, measure
- 51:11those impacts. So you're going to have to be
- 51:12able to answer those questions with numbers.
- 51:14To what extent or to what level of AI education should board
- 51:22members or must board members have to be effective?
- 51:25I think it's super important for every board member today to have
- 51:30a good understanding of how AI is affecting companies.
- 51:34It is a question that most boards are starting to ask.
- 51:38And I think as a board member, you don't want to be left out of
- 51:42the conversation because you don't understand.
- 51:44So I think at a bare minimum, you have to be somewhat up to
- 51:49speed with what is going on and what's driving companies because
- 51:54the, you know, many industries, as you know, Michael, are going
- 51:57to get disrupted with AI. We're already starting to see
- 52:00some of that playing out the market with impacts on software.
- 52:03So as a board member, in terms of understanding risks and
- 52:06having a strong, you know, engagement and risk in a
- 52:09company, I think it's absolutely it's, it's essential to
- 52:12understand that today. Myles Sewer's question.
- 52:16Ian Beecroft said in his South by Southwest talk that we are
- 52:20heading for a world where agentic AI does the execution.
- 52:25If this is so, what does the finance organization look like
- 52:29in the future? How many people and what skills
- 52:32with retained finance will retain finance professionals
- 52:38have? Look, I do think if you consider
- 52:40that you know at genetic AI will do a lot of that foundational
- 52:43work. Hopefully finance professionals
- 52:46don't need to spend as much time on data gathering, data
- 52:48manipulation. They can spend much more time on
- 52:51the analytics. So I think you're going to see
- 52:53roles evolve and the skills evolve and the ability to
- 52:58actually apply the judgment that you and I have talked about
- 53:01through the talk today is going to be important.
- 53:02And the human side of your skills, whether it's the
- 53:05conversation we're having today or the sort of that the softer
- 53:09skills are going to continue. It's going to continue to be a,
- 53:12you know, a weight on those skills going forward in finance.
- 53:14So perhaps it means less technical, you know, less of the
- 53:18sort of technical specificity and maybe more of the general
- 53:23with judgement, which means that you've got to learn over time
- 53:26because otherwise judgement is not something Michael you you
- 53:29get the first year out of university.
- 53:31And Coffee on LinkedIn says as CFO, what corporate functions do
- 53:37you think will benefit the most from agentic AI?
- 53:41Yeah. I think we spoke a little about
- 53:42this earlier, Michael. I do think a lot of software
- 53:45coding organizations, marketing, IT departments are going to be
- 53:49some of the earliest places where we'll see AI have a big
- 53:53impact. And the final question I have to
- 53:56ask, junior finance staff have traditionally learned the
- 54:00business by performing that first round of analysis.
- 54:04If AI now handles those basic tasks, how do you prepare the
- 54:08next generation of finance leaders?
- 54:10That is our responsibility, Michael, you know, I take that
- 54:13personally very seriously. I look at my teenage daughter
- 54:16who's about to go to university. And my advice to to young
- 54:20graduates is your jobs are going to evolve.
- 54:23So what's super important is to get a foundational understanding
- 54:26of AI, but still ensure you develop the analytical skills
- 54:29because that combined with judgement is going to be
- 54:32important about how you interact with AI.
- 54:34So perhaps there'll be less of what I did many, many years ago,
- 54:39Michael, pulling data, building these pivot tables and much more
- 54:42around your analytical skills, the softer skills that come with
- 54:47it. So you know, it's going to be
- 54:49important to get that, that education.
- 54:51And with that, I'm afraid we're out of time.
- 54:55A huge thank you to Marie Myers from HPE, the chief financial
- 55:00officer of HPE. Marie, thank you so much for
- 55:04taking your time to be with us. I'm very grateful to you.
- 55:07Thank you, Michael, I really appreciate it and enjoy the time
- 55:09today. And a huge thank you to
- 55:11everybody who watch. Before you go, check out
- 55:14cxotalk.com, subscribe to our newsletter, and we have
- 55:18extraordinary shows coming up, so check it out.
- 55:20Thanks everybody. Have a great day.