Justin (0:00:00):
Slack had this AI feature that they were trying to sell as an add on to their existing SKU. And no one was buying it.
That's not good for Slack and Salesforce is the over overload because what each of these companies is experiencing right now is masks of pressure to show incremental AI revenue to support the insane amount of money that's being poured into the AI Capex build out. If the deleting software providers cannot show AI revenue uplift all of the excitement is gonna end very quickly.
Now what we're seeing is that AI revenue uplift is happening by way of bringing these AI capabilities into existing SKUs and not even giving customers an option of purchasing AI and instead being forced into new packages, new bundles that include these AI capabilities at this fifteen, twenty, thirty percent uplift from what they were paying last year.
Michael (0:00:58):
Welcome back to The Spend Table where we serve up cost and spend intelligence, fresh. And today's dish might be like the Thai food I have this week, and it might give you just a little bit of hard.
My name is Michael Shields, I'm the vice president of procurement here at Tropic, and I'm here with my partners and co-host, Russell Lester, our CFO, who, let's be honest, Russell, you've seen your fair share of poor spend requests, and then Justin Etkin, our COO, our co-founder, or diving head first into the topic that I can safely say, I think it's on everyone's mind. It's also on everyone's latest software bill, specifically AI.
It's almost like it's nineteen eighty nine, except not only is everyone buy it, but this bubble, if we can call it that, has a brain, it can write code. It can analyze contracts. And if you believe a lot of the headlines out there, it's gonna empower a lot of CFO and others to cut half their workforce.
Eighty six percent of companies say they're investing AI. And honestly, I think the first question that comes to mind for me is who were those other fourteen percent that are not and why.
Fifteen percent of AI decision makers say that they can show an actual EBIT lift, fewer than a third, feel like they can tie AI value back to P and L changes. We're gonna break all of this down today. Are we in the AI tax era? Are vendors using AI powered as a blank check to charge whatever they want, and what should procurement? What should finance leaders actually do about it?
Alright, Justin. I know I've heard you use the term AI tax, and, you know, I'm traveling right now. If you look at some of the bills, there are line items on there that I pay for, whether I want them or not, like, concession fee, Recovery, service surcharge, etcetera.
And so we live in the world of software, not travel. We're seeing vendors that historically have, like, pushed five to ten percent innovation uplift, but now these are coming in at twenty thirty percent, and they're using AI features as justification. So you're a data guy, maybe talk us through what you're seeing.
Justin (0:03:11):
Yeah. Absolutely. So you mentioned that by default, there's this known practice of uplift for software providers.
In the subscription model innovation needs to be priced in for the next twelve months or twenty four months whenever the term length is. And so most enterprise software providers get away with a five to ten percent increase in that table stakes.
Now what we're seeing is that this five to ten percent increase is giving away like you mentioned to fifteen, twenty, sometimes thirty percent uplift, and being chart up to massive AI improvements that are getting bundled into existing SKUs or in a lot of cases, new SKUs that look very similar to what people were previously buying, just with the addition of AI capabilities.
Now what's interesting here is we saw this with a couple different providers previously, you know, earlier in in the year, Slack is a great example. Slack had an AI product. We've all seen the the summary of channels and, you know, whatever it is and they had this AI feature that they were trying to sell as an add on, two their existing SKU.
And I think it was, you know, at, you know, twenty percent of whatever the existing SKU price already was. No one was buying it. It was getting a very low success rate actually being purchased a la carte.
And that's not good for Slack and and Salesforce is the over overload. Because what each of these companies is experiencing right now is massive pressure to show incremental AI revenue to support the insane amount of money that's being poured into the AI Capex build out.
If the leading software providers cannot show AI revenue uplift, the whole song and dance of the AI ecosystem being built out, is gonna quickly turn them midnight and all of the excitement is gonna end very quickly.
And so this pressure is being manifested in ways of how can we actually show this AI revenue uplift. And now what we're seeing is that AI revenue uplift is is happening by way of bundling, bringing these AI capabilities into existing SKUs and not even giving customers an option of purchasing AI and instead being forced into new packages, new bundles that include these AI capabilities at this fifteen, twenty, thirty percent uplift from what they were paying last year.
Michael (0:05:39):
Yeah. I saw that you posted on LinkedIn today actually about Microsoft as well where it was, you know, anywhere from a twelve to a seventeen percent increase. Is is that kind of in line with this same concept?
Justin (0:05:52):
Yeah. So Microsoft's is a great example, twelve to seventeen percent uplift, again, one that we would typically see in the, you know, five to ten percent uplift range. But they're not the only ones.
You know, I I mentioned the Slack example. You know, we've seen examples of them in the fifteen to twenty percent, you know, uplift range. Google has been pricing Gemini into a lot of their your workspace products seen thirty percent uplift from from Google AI being in Gemini being worked into workspace, Gong, a thirty seven and a half percent uplift in some of their existing SKUs and packages.
So it's happening across the board and becoming a very normalized way for how companies are getting credit for AI, leveraging their lock in with customers as opposed to, really powerful innovative features that people are excited to purchase to drive this innovation.
Michael (0:06:43):
Yeah. You mentioned companies like Microsoft and Google, let's be honest, sometimes they can do it, you know, just because of who they are. But then you talked about some other companies where maybe don't have that presence. But, so it's starting to kind of flow to the rest of the supply base as well, which is interesting.
Justin (0:06:58):
Yeah. I mean, I think this is this is the story of the enterprise legacy legacy software providers versus the AI native upstarts. And everyone is capitalizing on the AI boom.
You know, no one's looking to Salesforce, Slack, Microsoft to be the leaders of innovation. They're looking to them to be the kind of steady, reliable secure solutions that set the core data out of their infrastructure.
And so, yeah. I mean, there's gonna be locking in there. They're going to be able to flex their AI muscles by way of their bundling tactics they've used in the past. And they're gonna, you know, buyers are gonna be looking to more of the innovative AI native companies, the Cursor, the Harvey, the Lovable of the world that are building truly transformational AI native capabilities.
And relying on them to drive the massive productivity increases as net new technology that didn't previously exist.
Michael (0:07:57):
Interesting, Russell, I gotta get your take on this. What's your gut reaction when you see this, obviously, impacting budgets.
Russell (0:08:04):
I like when Justin said, it's about to be midnight. I thought about the story of Cinderella, and of course, in Cinderella, the crystal, carousel and horses all turn into pumpkins and mice.
And so I think that maybe that's the analogy. And because the math doesn't math. Right?
CFOs are not sitting there and saying, oh, it's budget season. Everyone take your software stack add twenty five percent next year to the budget. So think about what's going to happen.
If it's a land grab, and Wall Street is putting pressure or even private companies, investors are putting pressure on these AI centric companies to monetize all of this, because they're about to turn into pumpkins if they don't. Then inevitably, what's going to happen is that trade offs are gonna get harder than ever, and there's gonna be more scrutiny on software purchases, more pressure to consolidate, because the twenty five percent won't be there.
And by the way, I think sometimes the tax is even a little more hidden. When we get into the realm of credits and outcomes and agents and I mean, I've been doing a lot of work in AI recently with something I'm partnering with CJ for mostly metrics on.
And I I found myself sitting there saying, I've been doing this for a couple hours. Like, how many agents am I using? So the tax is not only there and forcing decision makers to make hard choices? There's not even good visibility and transparency around how much all of this is costing in real time.
Justin (0:09:35):
Well, I've been talking to some of the enterprise procurement leaders. I I remember one conversation with the leading social media companies out there, and they asked us point blank, "What should we be expecting for this AI tax?"
A good procurement leader is bringing proactive knowledge to the table so that they can support that budgeting process. I think this is on a lot of companies' radar as they've started to see individual instances of, whoa, this uplift came out of nowhere.
While it might be frustrating in a lot of cases for these these suppliers where, you know, they've whole... They're holding their buyers hostage. Yeah, they may not have a choice that they need to plan for it accordingly around what that uplift look like.
So I think it's it's changing the standard budgeting planning process that many companies have used in the past and how they think about the, you know, the growth and expansion of their software. And and starting to individual certain line items and expecting much more significant increases, which is gonna impact the amount of investment that can go towards other stuff across the the business for for more experimentation and innovation in the next year.
Michael (0:10:40):
I think about, like, the problem of shelf where you buy a tool for a group of people and a subset of the people are using that tool. I'm gonna say appropriately, but in a way that justifies the the ROI spent on it.
I think we're now seeing that in the world of AI too. When you look at the look utilization across the board, it's it's split. Right?
So meaning, you're getting some value out of it, you're not getting some value out of it. Some of it, obviously is what I call fear of missing out driven purchasing. How do we kind of push back on this formal driven purchasing?
And then how do we look at it to ensure that you know from a utilization perspective that it's not just creating value in pockets, but universally across the board.
Russell (0:11:23):
I feel like we've seen this story kind of play out in the past when the world was going from on prem to in the cloud. And the reason why the analogy is the same to me is many of the purchasers were like, you know, it was expensive to create a cloud based technology, and we needed to recoup those R and D dollars, which inevitably meant potentially charging more than what you would for on prem and trying to convert it into a recurring revenue model versus a one time purchase.
So we were trying to transform the way people thought about purchasing capabilities. But the core capability set of what they were receiving they perceive to be the same.
So I think this is also going to force go-to-market teams to better articulate why these agent approaches are better, the value it's delivering because the value is hazy, for a CFO. We're we're being forced with finding the budget dollars.
We go back to the teams and ask what capacity has this freedom, how has this improved your conversion rates, you know, pick pick your tool, what's the outcome metric that says, your use of AI has improved. And I think Justin, you've made the point in the past that the ROI is greater than what we're giving at credit for. But it's happening person by person, and the way they're able to accelerate their every day across all the AI tools they're using.
Justin (0:12:45):
The ROI question is a is a tricky one to answer right now. And I think that what we've seen over the course twenty twenty five is like massive test budget proliferation across a number of different AI tools.
Mandate comes down from the CEO or the CFO that says, we need to get more productive, let's start creating processes and and permissions to go rapidly test and explore different for AI tools and see what the benefit is across the organization. Now there's a couple problems with that.
One is, yeah, we're not starting kind of strategic or outcome first in lot those cases. We're not looking at, "Here's what we expect in the hypothesis around what we can drive." And let's set up kind of a rigorous text testing mechanism to see how we get there.
But rather, let's start with the solution and then see what the outcomes are, which is, like, you know, the hammer looking for the nail as opposed to the nail give her the hammer. Hopefully, I got that analogy. Right?
So we're gonna see some some issues around what the actual results of that AI testing is gonna look like. I was chatting was one of my RevOps friends and a at a at a big enterprise, thousand plus, you know, person sales team, and he said that when they're testing AI tools, they are only giving credit to like, revenue uplift as an ROI case as opposed to, human efficiency, potential cost reduction, cost reduction uplift.
Because to your point, Russell, it's very hard right now to, measure and evaluate time savings and how that time savings is actually being, felt across the organization because I as an individual may feel amazing time savings to do my normal day to day tasks and do them in fifty percent all the time, and take that fifty percent and go do other things maybe non-work related or you know, more work social related instead of filling that fifty percent with with more productive work activities.
And so it's much harder to evaluate those types of trade offs. And so instead, more and more energy is being focused on how can we use AI to incrementally improve the revenue, the productivity that we can drive as an organization, particularly in this this RevOps kind of person's use case, rather than make a a business case around continued investment based on efficiency gains that could lead to, you know, head count reduction or something along those lines, which is obviously less popular.
Michael (0:15:09):
Okay. So you guys are using the term... Actually I I don't know who said it, but you're both talking about it. This term called AI value.
Okay? Which which by the way reminded me, I saw a TechCrunch article, and and it says here are the forty nine US AI startups that have raised a hundred million dollars or more in twenty twenty five. Okay?
So talk about how AI value, like, Cursor went from two point six billion valuation to twenty nine billion valuation in under a year. That's not normal. And and here's what's kind of crazy is it's not like they're profitable.
It's not like they're they're they're they're raising all this money, but they're not necessarily generating all this money. So Russell I'll go to you. You know, what happens when, like, when this music stops.
Russell (0:16:04):
What you're describing is setting up a market for a bubble, and the bubble will burst, but those that are providing real value will still be providing real value. So the issue will be those that are just what AI wrapper, we call it, you know, the appearance of AI when there's actually not much happening under the hood.
There's not the depth of data. There's not the depth of insight. It just, it looks good on the surface it's, like, when you get AI to create images and at first, it looks okay. And then you look at it. You're like, oh, my gosh. It it's missing an eyeball or, you know, something crazy.
That's what many of them will be finding these agent companies that explode at first. How are they ever gonna get that value back to the investors? And so it's gonna create this vicious cycle of, they've got a claw back against the users, the users will push back and say, well, all of you can't be asking for this uplift.
One of you has to go. The AI bubble will burst in so insofar as the fake-ish or low value providing AI companies won't won't survive. But that's been true of any high cycle. Right? I actually think that one
Michael (0:17:10):
of the things that concerns me more than the hype cycle is the fact that they are losing money today, because their cost models aren't built appropriately. And, yet, we're all figuring out how to live life and and build our products, what happens when that cost model changed, and and what does that do from a a and a budget perspective in your mind. Should that, should that be scaring us?
Justin (0:17:40):
There was a a podcast. I listening to over the weekend with David George. She is a a growth investor.
One of the things I remember from that conversation was as an investing team right now, they're willing to overlook gross margin as a as a metric that matters for AI companies. And in fact, it's sometimes better to look at a growth company that has bad gross margins because it means that they are heavy, heavy investors in AI.
Because the stack and cost of AI tools to drive business is so expensive today. Now the thesis behind that is that over time, the kind of gross margin, the incremental cost of any AI call is gonna go down as the the frontier models continue to compete with each other, build more and better models to to get more and more efficient and and and cost effective.
But, again, that also relies on the massive CAPEX build out that's happening right now and going back to the original point, the clear revenue uplift that we can point to from what the benefit is of AI. So there's this circular reasoning that needs to happen.
I think that at the application layer, like, we're still very much proving out the use cases where these AI investments generate meaningful and clear ROI benefit. Software engineering coding, that's winner. You know, you see the you know, all the tools that have, you know, grown from that, and and it's clear that the investments there are paying off in in our worthwhile.
Legal tech support sales. Like, there are pockets where this is is very much evidence around the success and benefits of of these investments. But to Russell's point, the hallmarks of a of a bubble is you're taking those isolated success stories and then extra across a much broader market where the story may not actually be there yet.
And that's very much what's happening where any AI native company is raising massive massive, money because they have some early traction, but we don't know how resilient that traction is, within those individual companies and also across the entire ecosystem from those different functions that are implementing.
Michael (0:19:48):
I wanna go back to, like, where we first started off this conversation where you're talking about companies like, Microsoft and Slack, who are pushing, you know, really big uplift right now. Okay?
And and the reality is as a buyer, if you're dealing with Microsoft, it's really hard to push back on that. Some of the other vendors, like, you can have a sourcing strategy, some competition, but let's be honest, like, when it comes to Microsoft, sometimes just, they they can they can push that, and you you don't have a lot of options there.
So what happens when you have now built some of your workflows? Some of your tools on these products, they are successful. And they're losing money right now.
And so what now happens six months, a year, two years from now when they're like, hey, we can't lose money anymore. And is is it is it another Microsoft situation where we've put ourselves in a situation where we've intentionally, and knowing gone with a provider who is losing money who eventually won't be able to lose money, and we've built our business on that that model.
Justin (0:20:53):
I I can't remember said it, but there's a there's a quote that the best enterprise software companies hold their customers hostage rather than give them real choice. And, you know, we see that.
And there's definitely a a delicate trade off that exists between how much can you ask of your customers to pay incrementally beyond what you're already paying, before you start to encourage exploration to other solutions that can do the job and be worth the massive human capital cost of change management, rip, replace, etcetera, etcetera.
And, you know, we're flirting with that right now. I don't know what the history of uplift have looked like at these some at some of these companies. But at the I have to imagine we're flirting with some of the highest year on year increases that that some of these suppliers have ever asked for.
And I know that the procurement teams that I talked to are not looking favorably upon that. But it all becomes a, you know, a trade off discussion around is it worth the time to explore an alternative or should we be taking that time and investing in, you know, innovation from net new providers that sit on top of this reliable infrastructure stack.
I don't know. It's it's it's tricky, but I I think we are flirting with some of those dangerous thresholds around what uplift can look like and still retain those hostages in a in a resilient way.
Michael (0:22:12):
I know that it's definitely being talked about at the procurement level for sure. In fact, I was at a conference last month where it talked about, hey, it's basically like this this profit and revenue migration from big legacy customers to the Microsoft of the world companies the world, and so will that continue?
Russell (0:22:32):
Well, this entire conversation we've been having over the past five minutes reminds me of physics because I'm kind of nerdy in that way. So entropy is that things tend towards chaos that even when something is initially in order or organized, it will trend towards chaos.
Is this tendency for things to become messy more than for the for them to become ordered. Entropy is similar to SaaS spend, it's similar to bolt on all these capabilities where things don't tend to get cleaner and more organized, things tend to get more messy.
So the more companies that come in with AI, and the more that it proliferate. So we have more more models, more deployments, more capabilities across every vendor. It's pro the invoices.
Everybody's rushing for a land grab to try to get their fair share of it. And so we've got the laws of economic incentive taking place, competition. We've got even the regulatory environment, creeping in related to AI.
I think pulling back all these AI capabilities are gonna be hard because to Justin's point, people are training their day-to-day on using all these tools. CFOs are trying to cost justify whether or not they make sense.
They can't budget for all of them. They're gonna have to start shutting some of them down or providers are gonna have to accept getting less money. And so we are trending towards not more order, but more chaos, which by the way, I think bode well for procurement.
I think it means, I think it actually means more than ever they're going to need help navigating what do I buy? Which tool makes sense. Which ones are providing, about but I think it's gonna get mer and h than ever to figure out where the true AI value levers are and what is the true ROI? Because it's gonna get very diffuse very murky.
Justin (0:24:32):
Dynamic that I think is is true of this AI native wave that we're experiencing, which is very different from the legacy SaaS or on prem wave that that have come for it is that these AI native providers are inherently much more transactional.
The nature of this massive growth that we've been able to see from these different companies is that it is easy to switch on Cursor or Rep or Lovable and start getting value ASAP. You don't need to spend six weeks integrating with one of your core systems.
You don't need to train your user base for a month around how to actually unlock value out of the tool. So you can turn it on and immediately get value.
The the the opposite side of that is that you can turn it off just as quickly, and you're not gonna create massive user, confusion or efficiency loss, because the these tools are based are built based on an outcome oriented model, and and there is not the same sunk cost fallacy that goes into the implementation, training and enable it related to them.
So we see this huge growth that's happening across these AI tools. Many of them, that's resilient. That's good. That's that's gonna persist for for many years to come.
But for a lot of them, you can see an equally fast, you know, reduction and and termination of these same tests that didn't actually prove out. Now I think that becomes super fascinating for procurement teams because procurement teams have always wanted into five ways to drive the best business outcomes by way of the suppliers and partners that we're working with.
But historically, lock-in has made it extremely difficult to be nimble and to have credible rip and replace or, you know, change management approach. But with AI native tools, we're gonna be able to see procurement led growth initiatives where we can be thinking critically about how do we evaluate, you know, the Cursor, Rep or, you know, the Level versus Alley or, you know, whatever different tools that are out there and and not have to be stuck with these long-term agreements that would require six to nine months to actually do anything at that if you wanna make a change and be much more nimble around how you test and learn and evaluate ROI across the these solutions.
Michael (0:26:52):
I definitely love. And I think maybe that should be a a topic for a future podcast because, you know, I think AI in the world of sourcing is gonna really mix things up in in a wonderful way.
I think about, like, you know, for years how enterprise sourcing has has has optimized for, you know, certainty. Right?
Like, long sales cycles, heavyweight RFP, costly implementations that are really meant to, like, de-risk everything. And then in in in the world of AI, I think, like, it it kind of flips that logic. Right?
To be honest with you where, you know, it, you know, you know, you can you can map process as an hours you can you can build integrations in in in days. You can, you know, measure outcomes in real time.
Like, those are all things that I think are gonna help, you know, from from a procurement sourcing, which I I'm genuinely excited about. Let's kind of move into into the world of, how do we tactically kind of approach this?
Because, like, our audience, finance, procurement, you know, leaders, and they are staring down kind of these AI renewal these these AI request. I would say that, you know, once again, you know, Tropic was built upon this notion that, like, we we should help our customers be, you know, have more data and and be more proactive. Right?
If you think about like, the foundation of what Tropic does a lot of is helping our customers be more proactive. And I think that we need to take that a step further as procurement leaders. You know, we used to say, okay, how are we engaging in sixty, ninety days in advance.
I really think, you know, this idea of, you know, checking with key critical suppliers for for business reviews, is is super important. Maybe even at, like, the half point of the contract to say, hey, what does our renewal look like, not from a financial point of view, but from a from a line item point of view, from a SKU point of view, from what is your road map look like.
I think for a long time when things were kind of status quo, not things were ever status quo, but not the disruption that AI is offered. I I think that was kind of something you get about away with doing infrequently.
And I think that now with your strategic with your critical suppliers, you know, that is no longer nice to have. So, like, you're you're kind of putting things into, like, a two-by-two matrix.
And and the ones that truly are, you know, you know, your your revenue is dependent upon your product just dependent upon meeting with them and under stand it. And then and then how do you put contractually, you know, clauses in in the contract to do it ensure that, you know, SKUs are still existing like, hey.
You'll guarantee that I can still buy this thing that, you know, you you're looking at strong utilization data, and you're not, you know, paying for, you know, features that, you know, nobody's using. And then you're obviously trying to negotiate things like price protection clauses.
These sorts of things I think. And then you you analyze all of that from a sourcing angle. I think sourcing is one of those things that's strategic, and it takes a lot of work. But when you do it right, it pays off in space.
And so does, is that sourcing? Or is that sourcing muscle? Is it being utilized? Is it active? Is it strong right now?
These are all questions I think you should be asking yourself if you're in the world of procurement right now. Justin, I wanna go to you, you know, what's what's in the contract that, you know, people are missing right now and what what should they be looking for?
Justin (0:30:05):
I think that given the pace of innovation across all these companies, one of the the key capabilities that we we think about a tropic is, like, overlapping feature sets and and and offerings. And everyday, a new AI companies popping up that solves for some unique use case.
In that use case is undoubtedly overlapping with one of the legacy software new kind of AI features that they're trying to release. And so keeping a very current and up-to-date grasp on who's doing what and how is going to be a strategic advantage that the best procurement teams can bring to their their companies and to their business users in a proactive and current state.
Business users are very focused on their day to day with their growth. And maybe they has sit on an an E with one of their supplier account managers to learn about what's coming, but that's once every six months, maybe once a year, and the innovation is happening much more much faster than that.
So procurement teams that are responsible for constantly managing and keeping an up-to-date perspective on the business outcomes we're trying to drive, and the partners that we're that we're working with to drive them. They can be sitting at the frontier of here's what's changing.
Here's what's new, the latest and greatest feature sets to bring those insights back to the business so that they can be armed and thinking critically and proactively around detect strategy that's going to unlock those business outcomes over the next year.
Michael (0:31:40):
What about any, contractual win mines? Are you are you seeing any any insights that you can share with, you know, for an AI, you know, vendor agreement specifically?
Justin (0:31:49):
I think that as more and more of these suppliers are working towards outcome based pricing models, these, like, credit and morph concepts are are becoming much more pervasive. You know, even when we see, like, OpenAI contracts and the way that they are offering pool of credits and, you know, different types of user models and have different credits per user.
Like, what is credit? Can you measure it? How can I know how consume against it, how does one ask or one, you know, AI engagement, consume credits versus another?
All this stuff is super super murky right now. There's gonna be a massive opportunity to bring better visibility into that. And what I encourage our our procurement customers is to based thoughtful around how you're putting in place the right protections around credit utilization, and uplift protection around those credits so that, so they can find ways to bring better visibility and awareness around what is a credit and and create that expectation as part of their supplier relationship, but also make sure that they're not signing up for massive credit growth and proliferation on these agreements that they can't control and be intentional about consuming in the future?
Michael (0:33:01):
That lands really well. How does that land with you, Russell?
Russell (0:33:03):
Any any other advice you'd share from a CFO, angle? Yeah.
Well, when I think about how CFOs are thinking about all of this noise, this chaos that AI is bringing, I immediately think about Star Wars. In Star Wars, there's a scene where Luke is attempting to destroy the death star.
And he's flying low. He's flying fast. He loses a wing man. There, he's taking cannon fire from all sides and Darth Vader, who had previously killed Obi-Wan Kenobi.
Obi-Wan Kenobi gave up and became one with the force. He's pursuing Luke. And Luke has gotta stay on target.
And his guy that's over on the radio is, like, "Stay on target. Stay on target." And, of course, we know his instruments mess up.
And Obi-Wan Kenobi says, "Use the force, Luke." And he closes the instrument and he uses the force. He fires the shot, the shot lands true and straight, and he destroys the death.
CFOs have got to help their companies stay on target because what's gonna happen is everybody's gonna get distracted with all these agent capabilities and forget what they hired the tool to do in the first place. Don't get distracted.
Make sure the tools are solving the core job to be done. And you do need to use the force, use your gut, use intuition, but you need a good tool set to help you figure out how to stay on target and know that the shot you're taking is the right shot.
There's lots of shots to take. And you're gonna be taking fire, help your company stay on target. Don't forget your reason for being in business and the main goals that you're pursuing.
And so I think weeding through all that noise, all that chaos if things are, if entropy is true, and we need to stay on target. That's the role of the CFO.
Michael (0:34:57):
I love that example. I I like it more than I thought it would actually.
Well, look, you know what time it is data a la carte. Weeks brought a stat with you every week. Nobody knows what the others are bringing. And then we talk about it, you know, what whether it actually matters.
So Russell, you know, you had a great stat last episode with, you know, two hundred fifty three budgeting tools. What do you have this time?
Russell (0:35:19):
Seven hundred and seventy percent. And what's the mean
Michael (0:35:22):
of the seven hundred and seventy percent.
Russell (0:35:24):
So I was reading in the Wall Street Journal that the number of searches this holiday season that people are enact via AI to search for gifts is up seven hundred and seventy percent. And did you know there's actually a name for it. It's called vibe shopping.
So the old way is you go to Google or Amazon and you're like, "Holiday gifts for women," and it gives you this massive list and you're like, I don't know where to begin and you're scrolling and scrolling and scrolling. Vibe shopping is where through AI you describe with a prompt.
"Here's what my wife likes. Here's her hobbies. Here's her interest. I want something unique," da da, and it spits out and it gives you a much more curated list. It's actually amazing of things that would be good.
So that's called vibe shopping. So my brain thought, what is the equivalent of vibe shopping in the business context? How are people beginning to use AI now as their primary search engine to answer things like, "I need a tool that does this. I don't wanna pay more than this, it needs to have these features and capabilities."
What's the best fit? And so what is, what's the vibe shopping in an enterprise software landscape. That's what my brain's is thinking through.
Michael (0:36:41):
We went from Star Wars to holiday shopping. Here you're taking us, so. I love it.
Justin (0:36:46):
I got a typical Pareto principal data a la carte moment here, any twenty rule. And given the applies is basically everything, I will I won't ask you guys to guess when it could be for this for this specific instance.
But the CFO at McKinsey gave an interview recently, and he said that AI should be focused on eighty percent on on productivity for growth and twenty percent on productivity for efficiency. I thought that was an interesting dynamic because so much more, so much of the lower out in the AI world is around people efficiency and redundancy and reducing heads and, like, the fear around that.
And, you know, at this point, and, obviously, McKinsey is undoubtedly sitting at the center of many, many AI conversations within businesses right now, and that eighty percent of the focus should be on productivity for growth and not for efficiency. I think it's telling around how to drive AI adoption in a non-confrontational, non-controversial way within a business and remove some of the fear associated with AI at the kind of employee level.
And instead, make it a productivity and growth enhancing tool that makes any individual, a super human in their day to day to field growth as opposed to any human become replaceable by, you know, a more productive and more effective human doing the same job.
Russell (0:38:13):
No. You also kind of feel like growth and efficiency are not necessarily dichotomous or anti-hypothetical, like by being more efficient, can't you therefore drive growth. So separating them as if they're opposing forces.
The, sometimes they are straight up opposing forces. I just believe in a world where the P and L's interconnected. And if you become more efficient by giving more time back that frees up more time for generating grow growth.
Justin (0:38:39):
I think that's what the idea is, but, you know, if you think that the main metric for AI investing in growth and productivity is, like, you see that as like, a revenue outcome based on now one person can do one point five x more that's going to lead to more revenue versus the AI for efficiency is, you know, one person can do one point five x, and that means we need you know, two thirds the people to actually do the same jobs, and therefore, we may increase profitability, but it's not gonna have the same top line impact.
Russell (0:39:13):
Got it. Yeah. That's just half the story.
Justin (0:39:14):
Yep. That's what statement in the in the the the data is representing here. And I think it's a smart way to look at it, especially when you when you factor in this floor around AI leading to massive job loss. And instead, we can think about AI leading to massive market cap growth and GDP growth across these businesses because of better productivity.
Michael (0:39:35):
Alright. So I'll whole wrap up with mine. So, and I'm actually really excited about this one.
It comes for Gartner. It says Gartner predicts that by twenty twenty six, seventy five percent of large enterprises will have AI driven procurement tools. I kind of wanna add my own stat on there because I I I don't think it'll just be large enterprises.
Look, we're already seeing AI, you know, replace you know, everything from PO processing, you know, invoice approvals, basic spend analysis, you know, low value sourcing, that sort of a thing. But we're, look, I'm I'm bullish about Tropic, and I I can already see how AI is impacting me as procurement professional and how it's really enhancing how I do category strategy, how I do contract negotiations, how I think about, like, supplier risk management and how that is becoming table stakes for me.
So I think that procurement companies of all sizes are procurement teams of all sizes are going to have AI driven procurement tools sooner than later out of necessity. And and so we're already starting to see it happen, but it's really cool how starting to super charge procurement and I love that.
Justin (0:40:44):
Think it's probably worth like, the, like, the double click into that is does SAP releasing a an AI feature into Ariba count as AI being implemented at the procurement level. Like, I'm not placing my my bets on the latest and greatest SAP release around what AI innovation is gonna look like for procurement teams.
Michael (0:41:01):
No. I think you're right because I think so many companies are thinking about. Okay. Well, that that is the PO processing and the invoice approvals and and the basic spend analysis, like, but but that's that's that's so basic at this point in time that, yeah is happening, but that's not what's gonna move the needle.
What's gonna really move the needle is how it's going to impact the person. Okay. How it's gonna impact the per procurement professional with with, you know, once again, that the category management, the Spire negotiations, the sourcing side of the things.
And I think that's way more exciting. And I don't think that's gonna come from what you mentioned, slapping on some sort of AI. Functionality to, you know, orchestration.
Well, I think that's a wrap on The Spend Table. Episode two, we have covered a lot of ground today.
We talked about this AI tax, and it is real. Right? I think most AI purchases need more scrutiny.
And the procurement playbook is being re rewritten in real time. So AI isn't going away, but the way we're buying it, that needs to change.
And so I I think what I heard from, Russell, what I heard from, you know, Justin, you you brought some great stories. You brought some great analog.
You brought some date... Great data, but be curious, you know, not committed. You know, demand proof of the value, you know, build flexibility your contracts.
And also, you know, recognize that there's a lot of positives. There's a lot of good that's gonna come from. You know, I I'll admit that once again, it is scary for someone like, you know, me as as I think about, you know, not only the impact of job, but also the impact to the cost implications of a business.
But what I heard you say is this chaos can actually be, you know, a really good thing. So if you've found this helpful, subscribe wherever you get your podcast, drop us a note if you've got AI, you know, procurement stories, horror stories or success stories.
We definitely wanna hear them, Russell and Justin. Thank you for the debate and until next time, this is The Spend Table where we serve up cost and spend intelligence fresh.
And today, like a lot of our dishes out there came up a little a surcharge. This one was an AI surcharge. See you next episode.
