497. Weekly Pulse Check on Automation & AI News

Speaker A: Foreign. Welcome to Transform now brought to you by ssnc, where we explore the future of work through the lens of AI. Across industries. Leaders are rethinking strategy, culture and operations to harness AI's full potential. This podcast gives you the insights and ideas to transform your organization, stay competitive and thrive in a world driven by change. Let's get into the show.

Speaker B: Hello everyone. Welcome to Transform Now. I'm Brad Hairston. And I'm Michael Marchuk and this is the weekly Pulse Check. We're taking a look at what's happened in the news around automation and AI and giving you our perspective on uh, some, some hot articles that caught our attention. So we've got six today. Michael, ready to roll?

Speaker C: I am ready. These are some really good articles are.

Speaker B: They are. So the first one comes from Inc. Magazine titled SpaceX is already worth More than Amazon. So why are the analysts sounding the alarm? Well Michael, unless you live under a rock, uh, and are Too totally unaware, SpaceX, uh, in their IPO, which just happened recently, they essentially leapfrogged Amazon and Microsoft within a few days of their ipo. But then you know, they've kind of, the price has gone down a little bit. So now they're, now I think it's below Microsoft again. But Elon Musk did become the first trillionaire in history. Um, I don't think that has held, held true as well. I think it's now he's like in the 9, 900, it's not even a trillionaire anymore.

Speaker C: Oh my gosh.

Speaker B: So we'll all take up a collection for, for Elon. Um, so, so break down all the zaniness around this IPO and, and the, the points that the article covers, if you would.

Speaker C: Yeah, so this is obviously a very highly anticipated IPO in the market space because it combines um, their XAI got sucked into SpaceX as part of this whole pre launch venture here. And so there's tons of AI seeped into this whole idea of SpaceX. In addition to all the space stuff itself, all the launches, the technology they have for launches and whatnot, um, you add together all of the interesting things that Elon has been able to do and the past short period of time, honestly in terms of getting his reusable rockets that are now, you know, launching satellites left, right and center all the time. And of course most of those um, relate to some of the other businesses that he runs within uh, SpaceX with the thousands of small uh, satellites he have, uh, for his communications network. But it also opens up the doors to things that he's talked about before in terms of data centers in space and a bunch of other things, um, that uh, just make a very futuristic viewpoint of where things are going. So there's a plus and a minus here. One is the technologies that his organizations are coming up with and the future looking viewpoints that Elon has in so many different areas, um, really have been unique and have been coming to pass. He took electric cars and made them a commonplace thing now with Tesla. And now he's also made this idea of private space launches very uh, you know, successful and profitable. So there's, there's certainly cash coming in. The biggest issue I think the, that the market is why they're signing an alarm is the amount of income coming in for SpaceX and its related businesses inside that, that unit. Um, it pales in comparison to other more established businesses like Amazon and Microsoft and Google and others who have a large track record of, of producing a lot of money. And so uh, the valuation is where the biggest issue comes in for SpaceX. And the other part that kind of, this article kind of plays into a little bit is the way that the IPO is done with the internal employees as well. Mhm. Now I'm not going to get into the M major details of how lockups work and whatnot. You can read the article for yourself. We'll always have the link in the show notes of course. Um, but the idea is that um, usually there's a holding period or waiting period for people. They can't sell the stock until a long time into the future. Sometimes it's six months, sometimes it's a year, whatever the lockup period is. And in this case here, um, the lockup period is very different in and it comes much sooner. So it's uh, interesting because historically uh, IPOs generate a lot of buzz at the beginning and then over time reality sets in and the earnings come out and people start evaluating it based on the earnings. And therefore the stock tends to come back to some kind of normalization. Um, when that happens the people inside are still in that lockup period and therefore their valuation like you mentioned earlier, Elon, being worth over a trillion dollars then gets normalized. Uh, and to where the stock, it's actually come down to some level that is more, more real. And uh, and then the valuations for those folks inside who may have initially thought they were gazillionaires, um, comes back to reality and say well we may did very well but we're not nearly as rich as we thought we were on ipo. Day. So I guess that's where the alarm is coming from now is that the valuation is so overblown, um, and we've seen with so many AI companies, uh, overblown valuations, that this is just a major hype on top of it. So um, that's how I saw this particular article and some of the points that it made in it. But I think I agree with a lot of it.

Speaker B: Uh, yeah, if you think AI startups are overpriced, SpaceX was priced at 107 times sales. That's unbelievable. You know, just to think about that in the article Morningstar, a gentleman from there valued the company at 780 billion. Whereas a NYU professor who's very involved in assessing what's going on with SpaceX, he values them at 1.3 trillion. So that's quite a difference. Uh, when you look at that and they're that far apart, that tells you that maybe the market is pricing it as more of a story than uh, than as a balance sheet. Think of it that way. But yeah, it, this one is amazing. And if you're an employee and you can get rid of stock, you know, after the Q2 earnings release, that's pretty, that's, that is pretty unusual and I'm sure some of them will benefit from

Speaker C: that very much so.

Speaker B: All right, second article is from NoJitter.com title is Agentic AI's next challenge tackling Accountability. So this article weighs in on the very important topic of when an AI agent makes a decision, who is it that owns the outcome? So give us your thoughts on what you think this article says about that topic and you know, if you agree or disagree with it.

Speaker C: Well you, you already know that I agree with this. We talked about this quite a lot both on the podcast and away from the podcast that the one line here is, that is right on, spot on with this is accountability is not a thing you can delegate to software. The agent is an instrument and the person who owns the process owns the result of using it. So there, there's always a person that's um, attached to that accountability, accountable for the outcome, accountable for the results of using that particular um, process through the means they do it. And so with that human in that accountability seat, um, there has to be responsibilities that the human's willing to take on and understand what those responsibilities are. Um, you know, he makes a bunch of really good points in terms of um, the fact that like, um, like we've seen from like the automation and ARPA era, um, there was initial Thoughts process of allowing, you know, the agents or the RPA bots to be able to have access to do like whatever they needed to do. They were just software. Right. But the reality came back in when we started realizing, and this is a long time ago when this realization hit, and I think it's going to hit soon here, if it hasn't already for folks taking on AI, that then you need to treat each one of these uh, agents like we did when we started originally doing this with bots, which is limited authority, their own user IDs with there's just like another, another resource, the person, um, that's using the software. So uh, in this case here he indicates, hey, if you have an AI agent, maybe a specific AI agent can only authorize, um, you know, refunds up to a certain amount or whatnot. Um, just like you would give to perhaps maybe a junior employee who is in customer experience who they're doing this work and they're able to, they're authorized to um, provide a, uh, you know, a refund up to maybe $500. But beyond that there needs, it needs to go. It gets escalated to the next level or needs authority or approval from someone else who's able to do that. So similar to that, that's the kind of overlay that you could look at these agents in that same kind of light is what kind of authority do we want to be able to or do we feel comfortable with offering through an AI agent? And then what would we need to escalate based on whatever parameters those are. So it's coming through, rethinking the process as you would do it with any other, any other person in this category. If I hired a person to do this, what would I want them to do? If they were a junior employee and if they became a senior employee with a lot more experience, what more would I allow them to do, et cetera. And then where would I always want to have another person with maybe greater authority who's um, the escalation point to ensure that we don't have one person either. Uh, again from even a person standpoint, um, misusing or abusing that authority that they've been given. In this case here, I wouldn't see the abuse coming from AI, but perhaps, maybe it would be easily to manipulate AI in order to provide, um, you know, to get to that authority and offer refunds, but perhaps they shouldn't. So anyway, I think a great article makes a bunch of really good points in terms of where that line's drawn. The fact that there's always a person at the end of every outcome who's ultimately accountable and responsible for that particular outcome.

Speaker B: Yeah, yeah, I thought the article was right on point as well. I mean this, this, this idea of risk based autonomy just makes so much sense. The idea that you would size autonomy by risk, business risk, rather than transaction volume, you know, Right. Question of think about what's at stake and whether a decision is reversible. Think about those things versus just the amount of volume that the agent's going to do. And I like the way the article framed it up as you know, most companies have an authorization void is what is the way they described it. Companies have given AI agents the ability to act without ever really defining the authority under which they act. And that needs to short up for sure. So, yeah, good article. I knew, I knew what you would think about it. But hey, that's why we're here. Just to confirm, you know, when, when a good article comes along that we agree with those make them even easier to discuss. So.

Speaker C: Absolutely.

Speaker B: All right. The third article is from m.cio.com and it's titled Beyond Automation. How much does AI really cost? Well, Michael, this is the article every CFO wishes they'd read before they do an AI rollout. And the core premise is that it's really the boring kind of high frequency background stuff that often blows your budget when you're not really carefully watching it. So why don't you delve uh, into this topic for us?

Speaker C: I shall delve into it as you say. Um, the idea behind this is, um, they've actually got some kind of cost modeling pieces that they've put into a framework to kind of describe how tokens are being used, uh, in specific AI interactions. And um, the way they did it is through this model. I'm not going to go through the whole model. You can read through it if you'd like to in this article. Um, but the idea generally speaking is some, uh, users have a very heavy usage in terms of the number of tokens they go through in a session. And so it's very easy to say, oh, you know, I've used this many tokens when I'm, you know, doing some heavy research or I'm building software or doing something else. And that's very easy to classify and categorize as you're a heavy user. Um, the challenge that they're saying is it's kind of a paradox. The lower usage, lower session usage, um, is where you have the most risk. And it seems kind of counterintuitive. Well, if I'm only using a few tokens for each, each of these interactions, why is this an issue? The uh, biggest issue they're saying is because of the volume. And these are the ideas that we've all, you know, had piled into our heads for years now that oh, offload all these, these low value things to um, you know, to a repetitive automation and let it handle it for you on your behalf, um, so that you can do the higher value work. Which is exactly how these token things are kind of falling out. The higher value work, the more research, et cetera, that goes to the individual users with their sessions with these high token usage counts. And then the lower value work, if you want to call lower value work, the more transactional, repetitive work, has these low sessions. But the problem is the volume. So all that volume that you've pushed off into an automated space now starts to bite you because all of those tokens add up really fast. We have one user who's using a high, high number of tokens per session. Um, they're only doing that a few times versus an automation or set of automations that an agent is doing that could use thousands or tens or hundreds of thousands of, of runs or operations through these on every monthly basis. So even if they use a few tokens times a hundreds of thousands of times, it adds up very, very quickly. It does. So, so this is one of those discussion, discussions that you know, we've talked about and uh, we've been kind of working with our customers about. And that's related to the. When you look at a workflow, identify where, what's being done at every point of the workflow and make sure you get the right resource assigned to the right work. Because sometimes AI is the right place where you need to have some complex judgment made in an automated fashion to be able to extract data or to be able to do some kind of summarization and whatnot. And AI is the right choice. But sometimes going back to perhaps a uh, script or an RPA to interact with some other software that you had in the past, a defined outcome makes more sense, especially with the repetitions that come because running an RPA bot or running a script is far cheaper to operate than it is to run through AI where you're burning tokens and the latency even for getting that decisionless decisions made, which makes, you know, aren't, shouldn't change from run to run theoretically, um, because you're looking for the same type of outcome at each, at each point in these high repetitive kind of Tasks. So again, allocating the right resource to the right work makes uh, your cost basis. Then, um, tuned for the type of work you're allocating each one of your workflow steps.

Speaker B: The article cites a Goldman Sachs prediction that agentic AI is going to increase token demand by as much as 24 times current levels. So I echo your point about right tool for the right job. Uh, not everything needs to be executed by an agent. There's plenty of work that can be done, you know, through deterministic automation or some other method.

Speaker C: Exactly.

Speaker B: Um, and if you default and everything goes to the AI agent, then, uh, good luck. Good luck fitting that bill when it, when it arrives in the mail.

Speaker C: Well, and that's the problem there, these organizations, I mean Uber is one of the poster childs for this year that they blew their whole budget before the first half year. Uh, the year ended this year. So you, um, don't want to have your organization come with a surprise bill, even though they're getting some great throughput on their AI and it's doing exactly what they wanted to do. When that bill comes due, um, there may be some, some tough discussions on whether or not this was the right thing to do. So it's better upfront to make sure you do your due diligence on the costs.

Speaker B: Very true. Okay, the next one is from 2, uh, 47 Wall street com title is expert warns companies are AI washing by blaming layoffs on automation they haven't actually built. This has been a recurring topic. We've had many articles in the past about, you know, AI washing. Lots, um, of companies doing layoffs for various reasons and always kind of making the same narrative about AI and automation is what's driving it. So what do you think about. What do you think about this, this specific one? This is one of the latest pieces on this topic. Do you think that too many companies are using it as the fall guy when there's other things going on here? What do you think?

Speaker C: Hang on, let me give you my shit. My shocked face for you, those of you who are listening, you can't tell, but I had my shocked face on. Um, uh, the fact that this is something we talked about and we've brought up a long time ago that um, the amount of AI transformation, like actual transformation that's being happening, that's happening at these organizations is still low. Yet the volume of people that are being let go is disproportionate to the amount of value they're getting out of their AI. So I, I think there is, um, they Call it AI washing. I think, I think they are certainly looking for something to blame. And it seems like a, ah, I mean from a first blush it seems like, well, this is actually a good way of doing it. Let instead of, instead of saying we've got problems with our sales or we've got a drop in demand, let's just let go of people because we're becoming quote, efficient in our back end and we don't need all those people anymore. So look how even though, yeah, okay, we're losing the top outline, but look at our bottom line shrinking too. So that means we're being brought more profitable. Um, that kind of sleight of hand works maybe for 1/4, 2/4, but it does catch up to you. In fact, we've seen a number of companies who've reversed their decisions and said, holy cow, we need people to be able to do this work because our AI isn't ready. And that kind of, I guess opens the door to question a lot more about what that organization is doing. If they initially said AI is saving us all this and we don't need these people yet, they have to hire people back because either AI didn't save them that kind of money or they say they lied. You can't say they lied. Maybe they were overly optimistic in their projections, uh, to be able to go forward and actually, um, you know, save whatever thought they thought they were going to AI, like we just talked about another article. Um, it needs a person behind the scenes to be accountable for all the actions that are happening. So transactionally you may get some benefits, uh, at high levels of transactions if you've optimized the way things are working. But you can only optimize your, your work so much unless you're completely overstaffed and uh, are making no money per employee. Uh, it's very likely that you've already tuned, tuned that staff, that staffing level and that AI will support some levels of efficiencies beyond what you have right now. But, uh, the efficiencies aren't why you're doing it. That's part of it, but it's not the sole reason. There's so much more in terms of customers, the customer service that you're providing, the expanding in markets you're trying to do to be more profitable. Along each one of those transactions that you do have to be able to get different insights into how your customers are leveraging your own products and services. These are things that, these are why you use AI, right? And it's not why you lay off people.

Speaker B: Yeah. Well, speaking of people, one aspect of the article I liked was they talked about the importance of human infrastructure.

Speaker A: Mhm.

Speaker B: And the idea that as AI tools commoditize, the competitive edge is really going to start to shift more toward human elements. Trust between teams, clear decision rights. I mean things that are more intrinsically human. I thought that was a interesting spin on this and a little bit of a tangent from the main topic about AI washing, but.

Speaker C: Right.

Speaker B: Um, I always like it when articles kind of insert something there about the, the human aspect and how that factors into it.

Speaker C: Right, right. Well the only other part too that I think it makes sense is they are giving the same kind of guidance that anyone, anyone who's doing these kind of investing things should be looking at, which is, you know, scrutinize the 10Ks in the United States. We have these 10K earnings, um, reports that come out to understand which functions have been automated and what metrics have improved. Not just we've implemented AI and therefore we're chopping heads. There needs to be something that's attached to, to Show We've implemented AI and reduced our call volume by 86% because AI is supporting that all. And we have increased um, customer success rates or whatnot because of this particular AI, uh, interactions that we're working. That's where you start seeing okay, this is, this is actually applied somewhere and they're tracking it and they know how to report on it. Other than that it's probably air washing.

Speaker B: All right, well let's move on to Article 5. This is from customerexperiencedive.com it's titled How CX roles are changing because of AI. So this is a nice counter argument to the AI is wiping all jobs off the planet. It's really a nice piece about some of the new job titles that AI is creating specifically around customer experience. So what are your thoughts on it Michael?

Speaker C: Well, I think they did a great job in terms of um, exposing some of this and using ah, Airtable as um, one of their, you know, the primary view and how this is happening. And it's interesting to see the Airstore airtable story through this because they're looking to expand and create use AI for more activities. Um, so they wanted to hire an AI agent architecture and they looked for the unicorn to be able to do this exact thing. And the reality is that a lot of these roles that the needs they have, um, are difficult to find those specifically those specific people. So the good part is that you can start opening up those types of roles and allowing your own people to start, um, experimenting and working their way into a new career path beyond the call center. Um, allowing those folks to have a, uh, much deeper view because they're the ones who talk to the customers frequently. They understand the conversations they have regularly. They understand how the product or the service that you're offering affects the customers, any changes that they have in it. So when you have AI and the capability for AI to be applied somewhere, who best to ask than the people who've been doing the work, who've been having the conversations? So I think this is a fantastic way of, um, of, of. We talked about this a long time. Upskilling your current workers, it gives them a new career path, but also gives them some, it re. Reinvigorates their, their view on your company. Hey, they're giving me a shot. I, I didn't go to school for this. Maybe I didn't go to school at all. But they're giving me a shot because I know how to talk to customers and I, I m want to learn this and I want to, I want to be able to, to take this skill to the next level. So it's like it's learning together, allowing the company to allow their internal folks to expand their horizons while getting benefit for the company as well. And the customers at the end of the day are going to be the most impacted by this.

Speaker B: Yeah, it just makes so much sense that customer service reps would become AI trainers because as you said, they, they understand the customer journey and they understand the product side. Um, so I really like that, I like that trade off. I like the way that's, that's evolving. And you know, the title AI Agent Architect I think is a, it's an interesting new role that has emerged and it's not necessarily a highly technical one. You know, when you hear architect, you tend to think about a, an engineer, a coder. Um, but I think, I think a lot of, in this article talks to it. A lot of people that are more in the customer service or customer success kind of realm. There's going to be good opportunity for them in this new era to actually train agents. That, that makes 100% sense to me and I think that's something we'll continue to see take, take place more and more. Yeah. Uh, all right, let's go to the last article. This is from Forbes and it's titled the Future of Work isn't about AI, it's about us. Uh, and I think this is a headier piece. I really liked it. And you know as the title suggests, it's talking about the fact that, you know, the real future of work challenge is not the technology, it's whether leaders can stay curious and human centered. Um, while everything around them is the ground beneath them is shifting. Uh, so what did you think about it? What were your takeaways from the article?

Speaker C: I agree that there's um, so many good points to this. One of them I thought was funny is that, um, Dr. Dr. Kelly Monahan, who you interviewed on our podcast, uh, was one of the, uh, folks that she had identified, uh, in this particular article, which I thought was very interesting because we saw the same type of thing with the conversation you had with uh, Dr. Monahan. Um, it's interesting. They went through our education, they went through career progressions, they went through the connections between humans and when human and AI. I think again, there's many parts of this article that I could focus in on. I guess the one part I'd start with is the idea behind all of our technology usage. And I think they put it real clearly here is the idea is that you want to be able to be more human and have that human contact and human connection. And that is what grows our society, that's what grows people. And allowing that human connection to thrive is where, you know, the technology is really working in your favor. And they talk about some of the jobs coming forward and whatnot. Talking about generation Beta, which I had not even thought about, those are born in people born in 20, 25 and beyond. Now they're talking about the careers are not going to be anything like the careers that we've seen. They talk about baby boomers averaging approximately 13 jobs over their working life and sixth distinct career chapters. And looking back at mine, I'm, maybe I'm not, I'm not a boomer, but I've, I've had, I've had uh, a roughly that sort of job history, which has been interesting. But the way they're looking at it now, today's children are going to hold, they say, 20 or more jobs through multiple distinct career chapters with much longer working lives. So yeah, it's, it's a, it's a portfolio career with shorter tenures, things that in the past would have been frowned upon. Oh, you only spent eight months here or a year here, when in fact that might be a badge of honor because you were able to contribute eight months towards the profitability of a specific organization before moving on to another company where you were helping them, doing something similar and gaining additional experience. So there's all kinds of interesting dynamics at play with the way we look at careers and we look at the way, um, we contribute now to an um, economic viability of any organization that we're a part of. So fantastic article. I mean, I'm glad that all these are leaked in our show notes because there's such good reading and I would definitely encourage folks to be able to dive into this because there's so much to get out of these articles and the thought processes that we can only touch on a little bit, um, in our, in our overviews.

Speaker B: If the next generation is going to change jobs that much, I, I really hope they're getting trained to be resilient, adaptable, you know, flexible, uh, to be quick studies because there, that, that is, that's driving me crazy thinking about, you know, changing that many times.

Speaker C: And ah, one of the things they mentioned on here too, and I know that Dr. Monahan may have touched on it in your discussion, but, um, the quote here is curiosity, not certainty is the name of the game. Not expertise, not control. That's right, curiosity. Because things as you know, and if you're listening to this podcast for any period of time, you realize that stuff changes very quickly, especially in our current day and age, and it's not slowing down. So unless you keep that curiosity in place, where you're actually looking at things from a different viewpoint and taking a different perspective on it, um, you can get stuck pretty easily or frustrated. So it's an interesting perspective for sure.

Speaker B: Yeah, my, my favorite piece in the article, there was a section that said there's a, there's a simple kind of litmus test that you should apply before adopting any AI tool. Uh, does it, does it help people work faster? Does it improve quality? And most importantly, does it free people to spend more time in the physical world, connecting with, in the physical world?

Speaker C: Exact.

Speaker B: Love that part. I think that's amazing. And I, I'm gonna, I'm gonna retain that, I'm gonna write it down or, or keep it with me because I, I, I don't think we use that kind of criteria very much. And in this world we're uh, in, things are changing so much and we get enamored with the next great technology that comes along. But if we think about it in that way, I think that's, that's a healthy perspective and it's one that will benefit our teams and benefit our organizations. So I really, really did like that.

Speaker C: Yep. Great articles this time.

Speaker B: All right, well, thanks for joining us here on the Pulse Check. Remember, we publish another episode of our podcast every week with an incredible guest. Be sure and check that out. Otherwise we will see you back here next week on the Pulse. Check. Bye for now.

Speaker A: Thanks for listening to Transform now, the podcast from ssnc. You can find all of our episodes on your favorite podcast channel as well as YouTube. To stay on top of the hottest topics in the world of agentic automation, subscribe now. The future of work is here.