What's Up with Tech?
Tech Transformation with Evan Kirstel: A podcast exploring the latest trends and innovations in the tech industry, and how businesses can leverage them for growth, diving into the world of B2B, discussing strategies, trends, and sharing insights from industry leaders!
With over three decades in telecom and IT, I've mastered the art of transforming social media into a dynamic platform for audience engagement, community building, and establishing thought leadership. My approach isn't about personal brand promotion but about delivering educational and informative content to cultivate a sustainable, long-term business presence. I am the leading content creator in areas like Enterprise AI, UCaaS, CPaaS, CCaaS, Cloud, Telecom, 5G and more!
What's Up with Tech?
Future of Work: Inside the Agentic Enterprise Index
Interested in being a guest? Email us at admin@evankirstel.com
Salesforce's Chief Digital Officer Joe Inzerillo reveals fascinating insights from the new Agentic Enterprise Index, painting a vivid picture of how AI agents are transforming business operations at unprecedented speed. Within six months, we've witnessed changes that would typically unfold over half a decade with previous technological revolutions.
What makes this shift truly revolutionary isn't just the pace, but how fundamentally different it is from past disruptions. Unlike earlier innovations where humans simply needed to learn new tools, AI agents bring entirely different capabilities to the table – requiring organizations to reimagine work from first principles rather than simply replacing or augmenting existing roles. This creates a fascinating partnership where humans and AI complement each other's unique strengths.
The data reveals intriguing patterns in adoption and effectiveness. Customer service emerged as the natural first frontier, with travel, retail, and financial services leading implementation. Surprisingly, human escalations have increased (from 22% to 32%), not because AI is failing but because AI efficiently handles routine inquiries, leaving human agents to tackle truly complex problems. The result? Companies are seeing customer experience metrics improve by up to 200%.
Perhaps most fascinating are the unexpected behavior patterns emerging. Customers interact differently with AI than with humans – asking questions they might feel embarrassed to ask another person and engaging in longer, more detailed conversations. Meanwhile, employees are developing collaborative relationships with AI agents, having 35% more back-and-forth conversations and triggering 76% more actions month-over-month. As Joe describes from personal experience, working with an AI can feel remarkably like collaborating with a skilled colleague, transforming three-hour tasks into thirty-minute ones.
For leaders wondering how to navigate this transformation, Joe offers practical guidance: start with lower-risk applications like after-hours support to build confidence before expanding to more complex use cases. Looking ahead, sales applications represent the next major frontier, with early implementations already showing tremendous promise. The revolution is just beginning – are you ready to join it?
Crossing Borders is a podcast by Neema, a cross border payments platform that...
Listen on: Apple Podcasts Spotify
More at https://linktr.ee/EvanKirstel
Hey everybody, super excited, we're diving into the future of work with Salesforce's new Agentic Enterprise Index, a groundbreaking report on the use and adoption of agentic technology Top of mind to all of us these days. Joe, how are you?
Speaker 2:I'm doing well. Thanks for having me, Evan.
Speaker 1:Well, thanks for being here Really excited and we're going to dive in, but before that, maybe introduce yourself, your team within Salesforce and what was the big idea behind the report.
Speaker 2:So I'm the chief digital officer Actually, today is my six-month anniversary and I lead up the digital enterprise technology group, which is a bit of a fusion of your classic IT kind of computers, networks and all that sort of stuff, combined with us using our own products and in some cases inventing some products to make Salesforce the best version of the agentic enterprise we possibly can. The index, I think, is a great thing that we're trying to do now to really just give people some tracking. You know, the agentic stuff is so new and I think a lot of people are sort of thirsty for data and thirsty for information and it changes so fast that sometimes it's hard to track things day to day. So we really put our heads together and came up with this index as a way of sort of benchmarking so we can measure progress along the way and really do so in a much more explainable way.
Speaker 1:Really helpful, really useful. The report shows tremendous growth in just six months, the past six months. Are we at an inflection point, a real turning point for enterprise AI, or is this just the beginning?
Speaker 2:Well, I definitely think it's just the beginning. As far as inflection point goes, I think this curve is going to bend upward, there's no question about it, and sometimes it's hard to tell as you're on the ascent, like how high that curve or how steep that curve is going to be, but it's definitely. We know directionally exactly where it's going and, as you said, you know, last six months alone feels like about six years. So I think we've now determined that agentics are like a 10 to one kind of a factor here, but it's. But it's really interesting to see. I definitely think that people are getting and settling in and certainly we are internally settling in to really start to understand what's working very well and now in our second or third generation of that and at the same time discovering new use cases and really exploring piloting things like that and trying to figure out where there's other opportunities where we could drive real business value.
Speaker 1:Amazing. And Salesforce calls enterprise agentic really a complete business transformation, not just, let's say, traditional automation we've seen for many years. So what does that look like inside? You know, a typical company these days.
Speaker 2:Well, you know, it's interesting to me.
Speaker 2:I like to think of it where, you know, in all previous sort of technological disruptions or of any sort I mean, you know, go back to, you know, the cotton gin and things like that, anything that was really a technological disruption the people were pretty much the same.
Speaker 2:Right, like it's education gaps. Obviously people needed to learn about the new technology, but people much the same, right Like it's education gaps. Obviously people needed to learn about the new technology, but people have the same capability set, more or less, that we have had for, you know, hundreds and hundreds of years. This is a little bit different. When you start to bring agentics in and you now have agents that are working, they have a different skillset. They're really good at certain things and then they're, you know, shockingly not as good at other things that humans are great at. And so when you try to figure out how to make humans and agents work together, you kind of have to go back to first tenants and look at the jobs to be done, and that doesn't necessarily aggregate to oh, I'm just going to replace this or augment this position on the org chart with an agent.
Speaker 1:It's actually that the agents are doing a different type of work and therefore the humans have to do a different type of complementary work as well. Amazing. And with agent force use surging, what kind of patterns are you?
Speaker 2:seeing in companies that are using agents most effectively. Well, I definitely think that like call it Gen 1. You know it's tough to refer to generations of multiple core. You know that that take place in multiple quarters, but I think the first generation of this you saw really centered around customer service use cases, and so in our case, like helpsalesforcecom, I think that makes sense Because, if you think about it, most places were using a very regimented outsourced data model to handle a lot of these things. So there was already a lot of guardrails and metrics and documentation and things like that to train the humans that were doing that job.
Speaker 2:When you start to bring agents in, it gives you this rubric that you can at least measure yourself against and say is the agent doing a better job, worse job, et cetera, than the humans are? And so I think those were the first use cases. They're very, very successful. You know, at this point in time we've saved over a hundred billion dollars in costs on our health side of it, and you know we've been able to redeploy the humans that were on that side of it to do other things or go deeper in areas that we've always wanted to go deeper but we just couldn't afford to have a labor force that big to go ahead and do it. So I think that's kind of the first generation.
Speaker 2:The second generation we're really now starting to scratch the surface of doing things that, like previously, were not really even thought of as being automatable. So you really start to get into, you know, certain aspects of the sales job, as an example, where the agent can do a whole bunch of follow-up, they can do a whole bunch of lead development things like that, where in the past that was exclusively the purview of humans. But we're seeing that there's real value there and continuing to expand usage. And I think for ourselves, that's the part that I'm most excited about, because it's not just about cost reductions, really about delivering a much better customer experience or buying experience for the Salesforce customers that are out there Incredible.
Speaker 1:And it looks like travel retail financial services are leading the adoption Interesting. What's driving them ahead of other industries, or verticals.
Speaker 2:Well, back to the customer service use case. If you look at all of them, there's scale B to C kinds of operations and they have these massive customer service use cases and so there's no question that the early momentum there is, you know, based upon the job to be done. These are really good use cases for it. And I, you know, I think retail is a great example. You know we have a lot of customers where, you know, there's a lot of conversations that customers, their customers, want to have with folks in the retail space Where's my order? Can I make this change? Things like that.
Speaker 2:And you're really starting to see like this, this I do think you use the word inflection point. I think here customers are really starting to see the inflection point of just expecting more, like expecting you to be able to do things. And I think having a Gentics, having, you know, support available in multiple languages, having it available 24, seven, these are all things that companies generally didn't do in that space unless they were massive global companies, and even then probably not every language and probably not every day. But now we're just, you know, we're starting to get there where people are just starting to expect that this is the case that this stuff is going to be available, and I think agentics plays a huge role in how to scale that from an economic standpoint and still deliver high, high value from a customer impact standpoint and still deliver high, high value from a customer impact standpoint Incredible no-transcript.
Speaker 1:So that sounds like a real shift in how people expect to interact with brands.
Speaker 2:Yeah, I totally agree. I mean, if you think about it, we're really two, three years into this kind of pivot with chat GPT, kind of disrupting things from an industry standpoint, getting into the consumer zeitgeist. But prior to that, most people if they got a chat bot, they knew it was a bot and it was probably not going to be able to handle their particular problem, especially if it was outside of the 80-20 rule or 70-30 rule, of things that were highly scripted in the bot, and so it was often a very frustrating experience and so people generally wanted to talk to a person because they believe the person could help them out. I think, again, chatgpt changed that. I think the agentics agent force and things like that people are now seeing that the agent can help them just as well as the person can in a lot of these instances and it's not just going to give them these robotic answers, it's going to have empathy, it's going to try to solve their problem and I think just the overall experience is such that people are now just expecting it and they're starting to experience great you know, starting to have great experiences and therefore expect even greater experiences.
Speaker 2:And I liken this to the web revolution, you know in general, where a lot of the early websites were not that great but all of a sudden people started to make some really great websites. And when they made really great websites, then other people had to make really great websites to keep up because the customer expectation had just shifted. That transition probably took five to six years. This transition's like probably five to six months. It's just happening so much faster.
Speaker 1:Fascinating and speaking of human agent relationships, escalations to human agents rose from 22% to about 32% as AI gets better at routing. How do you see that balance between you know the human connection, empathy and AI efficiency kind of evolving.
Speaker 2:Well, you know, it's a really interesting thing because you might look at that statistic and say, well, that's actually a decline, it's a fail in the sense of the number went up, so more human presence. But we don't actually see it that way, because it's not really all about cost savings. In fact, it's primarily about the customer experience. And what we're actually seeing is, if you look at the questions that people are answering, the easy ones are sort of getting picked off by the AIs, so the ones that we know about that are common use cases agent force is totally handling, and so it's a little bit of a high pass filter that the ones that the humans get are actually much more complex problems. But the other phenomenon that we're seeing is, if you look at the queries that are coming in, people are actually talking to agent force in ways they would never talk to a human being, and I think that there's a couple of, you know, a couple of different variations of it.
Speaker 2:In some cases, I think you know, there are questions that people have that they're not sure of, and so sometimes they might feel embarrassed to ask that question to a human, but they have no problem asking the agent and the agent can give them the right answer to it.
Speaker 2:So sometimes people have a technical challenge and like, wow, I should know the answer to this and they feel weird about it. But they, they are definitely asking the agent more and using it more like chat, gpt in the context of uh, of a support window. So we're actually seeing that we're getting more done with each individual customer and I think that leads to the escalations, because then you wind up in a situation where the agent's talking to somebody and it turns out they really do have an interesting use case that's not super easy to solve and again that goes to the human and gets solved by the human then. So I think we're going to start to see this. I suspect that the escalations may actually go up, but the quality of the experience and the customer satisfaction is also going to go way up.
Speaker 1:Fantastic. So it's great to see, beyond customer experience, employee experience evolving as well. I see employees are having 35% more back and forth conversations with AI agents and you indicate actions triggered by agents are up 76% month over month. You know, are we getting to the point where AI agents feel more like your colleague in the cube down the hall?
Speaker 2:Yeah, I definitely think so.
Speaker 2:I mean, I think you know, just using it in my own personal life and also in my work life, sometimes you get into a mode.
Speaker 2:I was just working on a fairly technical problem the other day and I was having a conversation with an agent and we were sort of riffing back and forth and it very much felt like I was sitting there with another engineer that was helping me work this problem, and so you know, it's not always the case, but I definitely think you're starting to get into that Zen space with these agents in some cases and you're getting there more frequently and for a longer duration where it really does feel that you're working with a colleague. I'm always careful to not way over-anthropomorphize these agents. They're good at very specific things but, man, sometimes they can have just massive impact. As far as time saving goes, like in this one particular instance, the thing I was trying to do I would have figured out myself, but it probably would have taken me three or four hours and instead I got it done in 30 minutes, collaborating with an agent, which is pretty spectacular.
Speaker 1:That's fantastic. And despite all this amazing opportunity, there's still a lot of myths. And despite all this amazing opportunity, there are still a lot of myths and misconceptions Fudd around scaling AI agents. What are the biggest misconceptions that you see out there in the industry?
Speaker 2:Well, you know, I think one of the things that's fascinating to me is I start to see it like self-driving cars. So when you, when you look at it right now, like, by and large, self-driving cars are safer, like autonomous vehicles, are definitely safer than humans by a long shot, but they're also held to a much higher bar. And so I think you know every brand, every customer that we're going through, you know, when you're dealing with this agentic technology is as much as we put a lot into it to make it much more deterministic. If you, if you crank the knob all the way to full determinism, it sounds like a chat bot. And if you let it be less deterministic, then you sometimes get it to do weird things. And that's just the nature of agentics and the nature of where the technology's at writ large, you know, across the industry. And so you know it's interesting with some brands where you're like, okay, well, well, this thing is, this thing is here, and they're like I don't know if that's good enough, and it's like but it's like 10 times better than your people that are doing that job, and it's like, yeah, but the turning it over to an ai, uh, you know there's a little apprehension there. And so I think you know we're still early in this, like, while the technology may be accelerating at this clip that's 10x what we've ever seen in one of these other things it doesn't necessarily mean the people around it are accelerating that fast.
Speaker 2:So I think so every brand, every technologist, is trying to find their way about, like, what's the right level of this, is good or great, and sort of defining that for themselves. And I think you know each brand has its own sort of personality. It wants to read through, and you know people are trainable to that. So is the AI. But the question is like what's what? The parameters are acceptable there? And so I think for us especially, like I'll give you a case in point example when we started the help project, it was giving very accurate answers, but it also didn't really have a lot of empathy. We sort of forgot the soul part of the customer service rep or the help representative at the very beginning. And so people were feeling like, OK, well, this is good. But like, look, I have a problem right here, I have a several down, I need to figure this out. And so us figuring out how to make that agent sound more like a human again.
Speaker 2:You know the empathy you know I don't want to over-anthropomorphize it, but like the feeling of empathy that the agent can engender are things that we've worked with and now people really do feel like, okay, this agent has my back, it's going to try to solve my problem, and the feeling part of it, I think, is just one of those things that you know the digital on-off and feeling are not necessarily two things that go hand in hand, and so I think that's a big part of this is just the evolution of companies, the evolution of what is great look like and trying new things and figuring it out. So I expect you are going to see a bunch of photo around it and a bunch of things like that, but when you find the use case and it rings through and you get it dialed in, then you know it. It's one of those things you know when great, when you see it.
Speaker 1:You're like, yep, that is impactful, that's actually moving my business forward, that's helping my customers. Oh, that's wonderful. I love the autonomous driving analogy. My first Weibo ride in San Francisco was terrifying. My 10th it felt as smooth as butter. I think I took a nap in the back seat. So, it really is something you have to experience firsthand. Speaking of which, behind the scenes at Salesforce, how are you using insights from the index to improve agent force and maybe also help customers get better real results?
Speaker 2:Yeah, there's no question. I mean, data is absolutely the fuel that fires this agentic revolution. Without question, this agentic revolution, without question, having qualitative data, observability, understanding what these agents are doing, using that as part of continuous improvement loops, that's the way you get to great. It's very much like a direct-to-consumer product where there's a lot of testing, a lot of experimentation, a lot of analysis that's being done to just continuously improve these things. And you know, the good news is the rate at which they're continuously improving is also accelerating. We're getting better at figuring out how to tune them and better at figuring out how to deploy them.
Speaker 2:But it is so data-driven, you know, exceptionally data-driven and disciplined sort of exercise. It's just absolutely really important. And obviously you see the data then rolled up into the index. Obviously we have more data than just what's in the index. But you see the index as sort of a manifestation of that kind of like obsession about instrumentation, that obsession about observability, because otherwise it's just really difficult to figure out what these agents are doing unless you're that relentless about it. And it's a big part of what we built into AgentForce.
Speaker 1:Fantastic. So a lot of experts from places like Harvard and Stanford are saying businesses will need to redesign and rethink and reskill for this wave that's approaching. Wave that's approaching. What do you think about how leaders need to prevent a divide between the AI haves and the have-nots sort of new digital divide?
Speaker 2:Yeah, I mean, look, I think it's real. I suspect you and I certainly know for myself. We're of an age where we remember where not everybody knew how to use a computer, or not everybody know how to use a smartphone. I remember teaching my parents how to text as an example, which is now their preferred mechanism of communication with me, because they actually understand that it works really well and I can get back to them more readily. And that divide has always existed, right Like there's always a leading edge and a trailing edge of these things.
Speaker 2:But I do think that, in the fullness of time, that it is more of a reskilling thing than it is, this massive displacement Like people are going to figure out how to use these agents. You know, people know how to Google things. You know, I remember talking to my wife once and she asked me something, and this is a while back, and she's like well, you should Google it because you're better at Google than I am, and it's like how can you be better at Google? It's Google, you know, and I think it's going to be the same thing with these AIs. Yeah, people will have more aptitude, some people will actually develop the agents, but as far as like being able to use them.
Speaker 2:I think it's really important that companies, you know, invest the time to do the enablement and the training for for their employees. But I also think that's, you know, one of the things at Salesforce that we excel at and one of the reasons that we're so focused on these enterprise use cases, because I also think a big part of that enablement is companies having the tools and having something like Salesforce probably already in-house and the vast majority of these companies us bringing the enablement with us. When we bring agent force, we can not just give them the technology, we can also give them the best practices and that customer zero use case. That is really the absolute catalyzing. Part of my job is to make sure that we're not just figuring these things out, but we're figuring them out in a way that's explainable, so that we can go out and evangelize to our customers hey, this is what worked for us. You may want to try it for you.
Speaker 1:Oh, well said. Indeed, some companies are seeing customer experience metrics and net promoter scores up by 200% with AI agents. It's really phenomenal. But yet we all experience those places where you know CX is falling short in our everyday lives. Where are they falling short and why? How can we help them?
Speaker 2:Yeah, look, I think a lot of it is. You know, first and foremost for me is accessibility. Like I just in my personal life bought a piece of furniture and then I had to deal with the company and you know it was sort of like Fred Flintstone slide down the back of the dinosaur at six o'clock on a Friday and I was like wait, what do I do? Like I can't talk to anybody until Monday. And I think you know a lot of it is that companies have practical financial constraints and you know people are not inexpensive and having off-hour support off-hours especially multilingual support is a very expensive proposition. So I think part of it is using something like agent force. We can really just extend that window to the point where, like if you were closed, you know the bar of what the agent has to be able to do is a good place to start because you were closed, so your customers had nothing to do.
Speaker 2:Once you start to get those use cases and use it as a bit of a test bed and I like the after hours use case for folks that didn't have after hours before is you really are able to develop these things and then you start to build confidence.
Speaker 2:With the technology, you build confidence with the agent, then you're like, okay, well, now maybe I can have that agent take over other parts of the day or augment my team. Or maybe I should rethink the way I'm doing IVR altogether, like maybe there isn't an IVR, maybe you just talk to the agent and then the agent can route you if it feels like it can't solve your problem. So I think really figuring out how to get on the journey is the biggest thing that we can help companies do. A lot of people are just it's the paralysis by how do I even get started? And that's one of the reasons that we continue to work and we'll have some really great stuff to talk about at Dreamforce, about really easing the onboarding. It's where people can really just get started with it and start using the technology and understanding it and then, like I said, build confidence and then, you know, transform, you know it's sort of. Those are the stages that I see people go through repeatedly and I've seen us go through internally.
Speaker 1:Amazing. Speaking of Dreamforce, you have your hands full over the next few weeks. What else are you excited about over the upcoming weeks and months?
Speaker 2:Well, I'm really excited about the sales use cases honestly, not just because obviously you know it's kind of in the title of the company, it's an important thing to us, but also know you see a technology and you're like, oh yeah, oh yeah, that's where it's going, that's definitely going to happen. And I think we're really starting to see that now on the sales side, where you're like, wow, this is, this is going to be impactful. And you know, I think the the thing about AI is it's been so rapid, like this agentic revolution has happened so quickly. I actually think that the you think that some people are like it's overhyped. I actually think the hype is proportional. It just may have been a little bit too much of a leading wave of it, and so we're now starting to get to the point where we're seeing especially, like I said, internally, I'm seeing some of these sales use cases. I'm like, yeah, that's going to live up to the hype, no question.
Speaker 1:It's just going to take a little bit of time to catch up. Well said, well, that was quite a mic drop moment. Thanks so much, joe, for joining and, as always, really insightful conversation. Appreciate the time Great. Thanks so much for having me. Thank you and thanks everyone for listening, watching and sharing this episode. Take care.