← The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX
The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX
#622: Balancing AI optimism with realism, with Stefan Weitz, HumanX Conference
00:29:55
Loading summary
Transcript36 lines
- [00:01]
Greg Kilstrom
What's the state of your website? If you're looking for a new digital experience platform or DXP to run your digital experiences, I have the book for you. The Agile Brand Guide to Digital Experience Platforms is part of my best selling series of MarTech books. In this book I explore and demystify DXPs and look at the roles a digital experience platform should play, the different types on the market, as well as how to initially evaluate platforms, then how to best implement a DXP once you've selected it. The book also features a forward from Rupali Jain, Chief Product Officer at leading DXP provider Optimizely, as well as several other thought leaders in the industry. Learn more and get a copy in print or digital now by going to the Agile Brand guide website at www.agilebrandguide.com.
- [00:48]
Stefan Weitz
The Agile Brand.
- [00:57]
Greg Kilstrom
Welcome to season seven of the Agile Brand where we discuss the trends and topics marketing leaders need to know. Stay curious, stay agile and join the top enterprise brands and martech platforms as we explore marketing technology, AI, E commerce and whatever's next for the Omnichannel customer experience. Together we'll discover what it takes to create an agile brand built for today and tomorrow and built for customers, employees and continued business growth. I'm your host Greg Kilstrom, advising Fortune 1000 brands on martech, AI and marketing operations. The Agile Brand Podcast is brought to you by Tech Systems, an industry leader in full stack technology services, talent services and real world application. For more information, go to teksystems.com to make sure you always get the latest episodes, please hit subscribe on the app you listen to podcasts on and leave us a rating so others can find us as well. Now onto the show. With AI adoption on the rise, how should businesses balance their optimism with the real challenges of regulation and ethics? Today we're joined by Stefan Weitz, the CEO of the Humanx Conference. Stefan brings a wealth of experience on how organizations can leverage AI to enhance their business capabilities while navigating the challenges of ethics, regulation and societal impact. The Humanx Conference is taking place March 10 through 13 in Las Vegas and features conversations and speakers at the forefront of shaping the future of AI in business. Welcome to the show, Stefan.
- [02:29]
Stefan Weitz
Thanks so much for having me. Appreciate it.
- [02:31]
Greg Kilstrom
Yeah, yeah. Looking forward to talking about this with you. Before we dive in though, why don't we get started with you giving a little background on yourself and your role at Humanx?
- [02:40]
Stefan Weitz
Happy to all keep it brief. I have been a nerd since I was a small child. I'VE been doing robotics and AI since before I could probably do proper math somehow and went from there into Microsoft for many years and built a bunch of products there, things like the Bing search engine and parts of Windows and then left Microsoft to go do some work in the logistics space with some private equity folks to really figure out how to shift things more effectively across the planet and then sold that company. And I've been starting companies for the last few years in all different types of areas but really my passion has been continues to be AI and I was just so excited when the world discovered the more the broader uses of AI that we're seeing today because finally my mom now understands what I love to do versus just have to talk about collaborative filtering models and things that she doesn't understand. So yeah, very very exciting.
- [03:30]
Greg Kilstrom
Nice, nice. Love it. Well yeah, we're going to talk about quite a few things here today and we're also going to be citing some research that Humanx and Harris X collaborated on together. So I'm going to bring up a few highlights. There will also link to that from the show notes so others can can see the report for themselves. So I want to start with AI adoption in the enterprise. Something certainly we talk about a lots of people are talking about balancing this AI optimism. I'm certainly I consider myself an optimist when it comes to technology in general and AI specifically. Your study with a Harris X found that 62% of business decision makers believe AI will enhance their work capabilities and 81% expect it to positively impact the quality of their work. Why do you think there's so much optimism about AI in the workplace?
- [04:23]
Stefan Weitz
Well unfortunately I think it's because many people have been not fooled but certainly been misled about what AI actually is doing at this point. So everyone sees ChatGPT and they see these incredible responses or all the various large foundation models and they see what appear to be human like responses. And it's very exciting. And people have anthropomorphized these things and said wow, we finally have AI. And the reality is not to take away from the amazing work these companies are doing, but it's still basically applied statistics. It's basically just trying to figure out what word comes after the next word. And so it's now granted, you could argue that human brains kind of work the same way, but it's lacking that kind of true understanding of the world and physics and objects and those sorts of things. It'll get there at some point, but something but my point is I think executives have been. Have been seeing the excitement with something like an anthropic or a chatgpt and rolling it forward into, wow, this is going to be amazing for my organization because we finally have thinking machines, and it's going to be a rude awakening. It already started to happen when these executives realize that indeed, it is still nascent.
- [05:36]
Greg Kilstrom
Right, right.
- [05:37]
Stefan Weitz
And just because it looks like a human, sounds like a human, and can operate at a thousand X the speed doesn't mean it's a human, doesn't mean it can actually do what they think it's going to help them do.
- [05:49]
Greg Kilstrom
Yeah, yeah, it's really good at some things, but it's not what a lot of executives, I think may hope that it is.
- [05:57]
Stefan Weitz
Not yet. Not yet. Not yet. The vision's there, but it's not there yet.
- [06:01]
Greg Kilstrom
Right. So along those lines, you know, we're not only are we in early stages of really, AI maturity and things like that, but also in adoption of AI. So the study also showed that nearly 75% of organizations have a strategy to adopt and utilize AI, but many are still in the early stages of that implementation. What would your advice be to those companies that are, you know, they've started, but maybe either they're stuck or they need to get moving faster.
- [06:37]
Stefan Weitz
Yeah, this is like, to me, a classic. They need to slow down to speed up. What we saw in the last two years was really this explosion of pilots in companies. And it's because these tools are sometimes very inexpensive to begin and groups had autonomy. And so you saw just this, literally hundreds in some cases at large companies of AI pilots. And as you can imagine, they weren't run from a central PMO or they weren't run from a central center of excellence in many cases. And so you have, you know, various varying levels of success, and even the measurements of success are, I would say, in some cases questionable, because the teams that are doing this maybe aren't the one, aren't the best at actually defining a business ROI and calculating it. They might be a tech team or a CS team that's just doing something to get ahead. And. And so that's the first thing I would say is I would say take a look at the portfolio of things if you already have them out there, and really understand what's been happening. And then look at your bills, too, because things can start cheap and get really expensive very quickly. And then so once. So get a handle on your landscape. And then I would say again, as much as I don't love the notion of a pmo, I think they, a project management office. I think they can be really burdensome. You need something, you need some kind of traffic copy that looks at a rubric or a framework to really understand where should we be applying AI in the company. Because what a lot, lot of times what happens is people see a bright shiny object, a cool tool or a cool system and they bring it in. It could, it could have impacts, could have positive impacts, no question. But is that the highest, best lever you can pull to really drive business impact? And that's, that's what's missing now is people saying even though it looks really cool to have a CS automation, if I'm getting my butt kicked on supply chain and can't figure out my, my, my lines over there, even though this automation thing looks, I'm sorry, this, the CS thing looks really cool, really, I should be spending more time, more money and going a little slower to fix my supply chain forecasting issues. And that's going to have the highest benefit for that company. So that's, that's what I would say is really think through and go slow and you'll be much happier than we're seeing people now who are regretting allowing the pilots to flourish.
- [08:50]
Greg Kilstrom
Well, yeah, I mean I would also add to that when things get too fragmented or maybe down the chain too much, then you don't benefit from AI being something that I see as a very horizontal capability when done well. Right. So it's like not only are you maybe not allocating resources in the best way possible, but you're also just missing opportunities. Right. Where to cross functional stuff. Right?
- [09:18]
Stefan Weitz
Absolutely. Yeah.
- [09:20]
Greg Kilstrom
So, and I mean to that, to that end, you know, the, the other part here is, you know, we've all been big data was what the buzzword like 15 years ago or something like that. And so we, we started collecting like mountains and mountain lakes and mountains and whatever of data. And what I'm seeing in the, in the orgs that I work with is it's AI adoption is really shining a light on, well, what have we actually been doing with that data? What's the, what's the actual quality versus quantity? You know, what's the actual quality of that data? So the report as well highlights the importance of high quality data in the training of AI models. What are your thoughts there? Again, storage doesn't necessarily seem to be the problem, but accuracy and all that is.
- [10:09]
Stefan Weitz
Well, storage hasn't been a problem for a while thanks to all the very cheap kind of storage options we have these days. What's really a Challenge is you've got many cases, heterogeneous data systems across an enterprise. So you have marketing has their silo and you've got ops has their silo and whatever, and sales has their silo and they're all in their different tools. And so of course the data lakes were made to kind of pull these things in a central location and normalize them, make them usable and cross functional. But what you're seeing is that that data lake, which was in place for a very particular purpose, which was to collect everything and normalize it, wasn't the same level of perfection that's required to really train AI models effectively. So you end up with these big, big, big lakes of data that is good enough for quick and dirty analysis, are good enough for a crosstab, but not good enough to actually deploy into production with an AI system. And so that's where I see organizations who are winning these days really stepping back again and saying, okay, before I even get to forget AI, before I, even before I even think about AI, my first step is to understand my data posture, what I've got out there, where pipeline's broken, where pipelines missing, and create that comprehensive view of the enterprise data. Then you can roll into, okay, now, great. What do my AI systems need and how do I modify or use my lakes to provide depth on them?
- [11:41]
Greg Kilstrom
Yeah, yeah. And so as far as investments go in, I would definitely say, you know, agree that investing in the data infrastructure is, is, is really key. 78% of leaders, according to the research, expect their AI investment to increase in the next three years. Smaller organizations maybe lagging behind a little bit in that. But assuming that investments in data infrastructure and quality and all that is kind of the foundational thing, you know, what advice would you give? Let's say an organization has that, that data infrastructure, kind of, at least it's well on its way. Where would you recommend that an organization invest next?
- [12:23]
Stefan Weitz
Well, if it's, if they have a good handle on their data, I would go back to what I said earlier, which is really, you know, don't be persuaded to go chase that shiny thing, which, which sounds amazing. Again, that could be the high, it could be a great ROI for you. But really the next step is to do that. Forget AI. Look at your, look at your business. You know, your, your kind of operational framework, like what are the areas in your business that are struggling, period. Forget about any systems or process or systems or technology. Think about where your process is. Is it, is it. Your sales pipeline isn't tracked carefully enough so you can't figure out what's closed and open. Is it, you know, is again, is your ops not tracking manufacturing successfully and therefore you're leaving, you know, finish ras on the floor, whatever, all those. I mean, I could go, I could think about a hundred different things that I've seen in my time. I would start there, like start, start with the core business problems and then, and say, wow, if I could fix this, I could increase margin, I could decrease my costs, I could increase productivity. Then from there, that's when you go into your AI solution. Because in some cases it may not have a solution, it may not be a solution, that's great for it. But other cases, like, think of like, you know, like accounts apar type stuff that we're seeing some really cool tools now to help reconciliation at an automated fashion. It's like a, it's like a bigger version of RPA we had five years, 10 years ago, but it's truly. You could save literally hundreds of hours in a month from an accounting department by letting the machines actually match invoices to, to payments and whatnot. So there's just got, you got to pick that highest best use, otherwise it's just kind of wasting money.
- [14:04]
Greg Kilstrom
So then, moving to next topic, want to talk a little bit about, you know, AI and societal impact? You know, there's been a lot of very positive and very like doom and gloomish, you know, stuff being talked about AI. You know, as I mentioned, I'm generally an optimist when it, when, when it comes to these things. And you know, according to the findings of the report 3 and 4, decision makers think AI will have a positive impact on society, especially in areas like science, education, healthcare. Where do you see and where do you think the most meaningful societal changes will be brought about by AI?
- [14:42]
Stefan Weitz
I think if I had to place a bet, which I do with my other venture hat on that I wear, I'm betting on AI really helping us with our energy. I think if you look at what's being done out of Yale with their Tokamak reactors, the fusion reactors have been talked about for years and years and years. And they're amazing, amazing, amazing, amazing. But they have this very nasty habit of the plasma field that contains the reaction, can sometimes be penetrated by the reaction itself and the whole thing shuts down. And as an example, the Yale folks trained in AI to be able to modify the magnetic field in real time to keep that plasma contained. One small example of where we're seeing dramatic potential for fusion. And really, if you have fusion as an example. And you really have solved for all intents and purposes the energy usage from around the world. So it doesn't have batteries to store everything else. But really if you could have unlimited, basically unlimited, basically very, very clean and basically free, minus all the capex and put into it power that's going to dramatically change the planet. And so that's where I get really excited, things like that. Also things just like edge computing, edge AI. You think about certain areas like parts of Africa and we don't have the best connectivity or don't have a ton of smartphone penetration. I think, you know, the necessity is the mother of all invention. And you saw it happen, the payments back in Africa 20 years ago that they let they just leapfrog the US and North America around mobile payments because they kind of had to. They had feature phones, they had Nokia phones. How did you pay someone that way? And pace it came out as a result. And I think you may see the same thing with AI that these, that the, the countries and regions that don't have access to a limited bandwidth and have very, very high standards or high wealth to go pay for these models and whatnot. I think you're going to see some real creativity come out of that and you're going to see those developing markets I think probably take the lead in some cases because again they have to.
- [16:38]
Greg Kilstrom
Yeah, yeah, that's fascinating. And with that growth and fairly diverse growth as well as you're saying, also comes the need for regulation. Right. So the survey said seven in 10 respondents believe that AI should be more regulated. I know there's some things kind of same as with consumer data privacy. Europe is, I would say leading the charge in a sense, although there's still a lot of stuff up in the air. Right. So should governments or companies or both lead the charge on AI regulation? And why would you say so?
- [17:18]
Stefan Weitz
Well, yeah, I think the only way this is going to work is if regulators look at outcomes and not the systems themselves. So I think the answer is if you look at the study, what you did to the same thing that most people said there should be a blend of these. Everyone appreciates checks and balances in these organizations and leaving it just to companies to go run amok doesn't sound like a good idea. Trying to have a very large bureaucracy regulate a very fasting technology has never worked very well. So really it's more. I saw the California bill, they were going through Congress there and they were trying to. And even the executive order from President Biden around limiting petaflops and these things are, they'll be out of date and they already are outdated, frankly. And they're not really getting to the core root of the problem, which is these things can have dangerous outcomes. And that's what you have to regulate against. Kind of like the drug companies, FDA does not necessarily regulate the internal processes of how drug discovery and development works, but they certainly can regulate the end result and make sure it's appropriate. So I think that's kind of the model we're going to have to see here because there's just no universe where only industry can do it or the government can do it in isolation.
- [18:32]
Greg Kilstrom
Yeah, makes sense. There's also significant concern about unethical AI use in things like elections, social media, policing. How can, given kind of what we're talking about here is this need for both areas to regulate, how can businesses proactively address some of these concerns while continuing to push boundaries and innovate?
- [19:00]
Stefan Weitz
There's good examples of companies. One company I love called Trupik, they've developed with all the major tech companies this thing called C2PA, which basically is a. Think of it as a digital watermark that describes an image and its provenance. So basically you imagine someone takes a photo and then that photo gets edited. If they're using C2PA compliant tools like Adobe is putting it in their systems as well, it'll actually have literally a blockchain record of the modifications that have been made to it. And so why is that important? Well, if you think about companies who, who rely on, on visual data, it's insurance, it's media, it's the big tech companies, et cetera, making sure that they are promoting things like CTPA and using that and really making it much the same way today as people expect to see the SSL lock in their browser when they go to their bank website. And if they don't, they go, that doesn't seem right. I think we're going to get to a point not, and not, not too distant future where people will begin to look for, okay, is this image CTPA certified? Can I hover over it and see a lock or a check or. I don't know how they want to do it, but I think that's a good example of where it's a great privately held company doing amazing work with all the big tech companies. And they are, I think potentially helping really solve a huge problem, which is the inability for anybody really at this point to detect the fakes or the modifications that are happening with AI. So think tools like that, that's how I say companies need to be thoughtful about that, go proactively seek those solutions. And that doesn't hamper innovation at all, actually helps it, but it allows them to play it responsibly.
- [20:39]
Greg Kilstrom
Yeah, yeah, definitely. Yeah. And I mean that's a win win, right? I mean there's, there's a, there's a commerce component to that, but there's also an ethical and it benefits us.
- [20:49]
Stefan Weitz
All right, Exactly. No, it's a great one.
- [20:52]
Greg Kilstrom
Yeah. So looking ahead a little bit, certainly there's lots to look forward to at the Humanx conference. We'll talk about that in a second as well. I'm certainly looking forward to attending a couple things on the horizon here. So community generated AI, that's another thing that was brought up in the study, has been identified as a way to improve data quality. Can you talk a little bit about how this works and why it's an essential part of future AI strategies?
- [21:23]
Stefan Weitz
Well, I don't think there's a great definition for it yet, honestly, but you're seeing a lot of people, whether they're using RAG or fine tuning, who are taking existing models and attempting to kind of in a specific way morphed into their particular use case. We are seeing open source, which I think it is also being often talked about as a community generated kind of AI foundational models. You're seeing some of the most impressive, like the OM model from Kai Fu Lee over in China or even Lama obviously is doing incredible work too with open source models. So I think it's hard for me to kind of put my finger on exactly when people say community gen AI because I don't really totally know what that means. But certainly if you think about the old folksonomies in the old days or even going something as simple as a Craigslist, certainly that wisdom from the masses and with control is what we have seen work in the past. Doesn't always work perfectly, but it works pretty well. And so I think depending on which angle you're looking at Unity generated AI as a solution, it's going to. There are many tools that are being built right now to actually help companies use AI internally and leverage their employees, leverage the larger networks that those employees have to tune data to be appropriate. But yeah, it's still early days for all that.
- [22:39]
Greg Kilstrom
Yeah, yeah, definitely. Well, so, yeah, let's talk about the conference. So HumanX, it's March 10 through 13 in Las Vegas. What should attendees look forward to?
- [22:51]
Stefan Weitz
Well, a few things, I guess. First thing I'd say is you have almost 300 of the world's best speakers in AI. So if you look at our speaker list, it's really anyone who's in AI at the senior level. So it's not just the program managers, it's the CEOs and founders. So you have incredible speakers. The attendees themselves are looking really incredible as well. So a lot of connections you can make. That's why people go to conferences, is actually to make connections and find people and find partners and find solutions to problems. I love our programs. Solution Bridge, Adventure Connect, these are two programs we have. Solution Bridge really matches buyers and sellers. So buyers come in and say, I have these six problems and then sellers on the other side come and help solve them in real time at the event. So that's a huge thing as far as ROI is concerned. Leaving an event with a solution in hand. Your most pressing problems, that's really exciting too. And just the event itself is going to feel really special. We've got some great surprises, some great surprises in store. And the event, if you'd been to any of the ones that my partner John Weiner has done in the past, Money 2020 or Shop Talk or HLTH, the health conference, it's the same DNA. So the level of production quality, the level of convenience, the high quality food and bed, these are small things, but they make a difference. If you're not running around looking to work, where can I use my drink coupon to get a coffee or stand in a huge line to get a crappy taco? This makes a difference. And so that's what we're doing, all of that. The last thing I'll say, which is exciting, is that every single session is being recorded, transcribed, summarized by AI, and then it will be made completely searchable by our own custom AI. So literally any attendee after the event can say, were there any sessions where we call it community generated AI were discussed? And even if they weren't, those sessions, we'll, we'll find where it was mentioned in sessions and drop you right into that conversation. So really, you, you can be everywhere, everywhere, all at once at Humanx because of our technology we're deploying. So I'm really excited.
- [24:44]
Greg Kilstrom
Love it, love it. Well, yeah, looking forward to it. We'll have a link to register for the conference as well in the show notes and looking forward to seeing everybody there. So, Stefan, thanks so much for joining today. I've got one last question for you before we wrap up. I like to ask everybody, what do you do to stay agile in your role and how do you find a way to do it consistently?
- [25:08]
Stefan Weitz
I think the easiest way for me to think about that is to take on more than I can handle. Which may sound like a counterintuitive idea, but I mean, I wear many hats and it forces me to prioritize and forces me to order things in certain ways because I can't do everything. So that's how I stay agile is that I'm constantly pushing myself to learn new things by taking on things I probably shouldn't be doing in the first place, but then also assigning those the right amount of time so I don't end up spending way too much time on something which isn't important. But to me, agility is all about learning. So if you can just keep pushing and keep learning new things, that is where agility tends to come from. The second you stop that, you're just necrotic.
- [25:51]
Greg Kilstrom
Yeah, I love that. Yeah. Someone I know very well told me, if you need something done, ask the busiest person you know. So it reminds me of that. But no, I love that. That's true. Well, again, I'd like to thank Stefan Weitz, CEO of HumanX. To learn more about Stefan and Humanx Conference, you can follow the links in the show notes. Thanks again for listening to the Agile Brand brought to you by Tech Systems. If you enjoyed the show, please take a minute to subscribe and leave us a rating so that others can find the show as well. You can access more episodes of the show@theagilebrand.com that's theagile brand and contact me if you're interested in consulting or advisory services or are looking for a speaker for your next event, go to www.gregkilstrom.com that's G R E G K I H L S t r o m.com the Agile brand is produced by Missing Link, a Latina owned, strategy driven, creatively fueled production co op. From ideation to creation, they craft human connections through intelligent, engaging and informative content. Until next time, stay curious and stay agile.
- [27:07]
Stefan Weitz
The Agile brand.