Our Tools Can Think Faster. That Doesn’t Mean We Should Think Less

Check out the full episode below! Enjoy The Exchange? Don't forget to tune in live Friday at 12 pm EST on the Greenbook LinkedIn and Youtube Channel!

AI can summarize faster. Analyze more. Surface patterns we might miss. Take work off our plates that used to take hours. Great.

But one of the clearest messages coming out of IIEX.AI last week (get on-demand access here) was that none of that transfers responsibility away from the person using it. As Karen Lynch put it during Episode 149 of The Exchange, the machine can produce the output. You still own the decision. And one line from the IIEX.AI town hall captured the tension particularly well: “Cognitive offload is fine. Cognitive surrender is not.”

That idea followed Karen and Lenny Murphy into the rest of this week’s conversation, from shopping agents and embedded insights to synthetic research and data quality.

Thanks to our producer, Karley Dartouzos. 

Use code EXCHANGE to get a 15% discount on your general admission IIEX tickets!

IIEX West 

Stay Ahead of the Curve! Subscribe to The Exchange Newsletter on LinkedIn Today! 

Transcript

Karen Lynch: Are we just live we are live we're just live we're just live we're just live you know I feel like I feel like if if we were actually a live like or something there'd be like music or something so there'd be some sort of an audio cue that we're live maybe right right instead we're just talking we need something like that so you know it's it's what it's our charm So, happy Friday, Karen.

Lenny Murphy: Happy Friday, everybody.

Karen Lynch: Oh, happy Friday for sure. Happy first Friday of fall. I guess. I don't know. Does it feel like fall? It feels like fall. Yeah, it definitely does up here.

Lenny Murphy: Yeah, I'm in Atlanta right now and it's in the 60s. So that's wonderful.

Karen Lynch: Yeah. Yeah. Crazy cold. Crazy cold. So, yeah. So we are. All these disclosures. So Lenny is working at an Airbnb in Atlanta. So let's hope that Airbnb Wi-Fi works as well as his at-home internet. Emma is still subbing in for Carly. So let's hope that that goes off without a glitch. All the links will be published after the show, again, like she did last week, not during the show. So if you're. Checking us out on LinkedIn, you can feel free to engage in the text thread there. That's how we're doing the show again this week. Carly is still off in Europe somewhere having a grand old time. I don't know if we have any other caveats, but those are the two big caveats.

Lenny Murphy: I think that's it. I know we are on a time crunch today, but we didn't build in, and we should quickly debrief. From IXAI.

Karen Lynch: Yeah. Well, let me tell you, it was, yeah, amazing. I won't spend too much time. It was an amazing two days. It was, it was, it was packed with learning to the point where I really. I don't always, when it's a virtual event, I don't always just sit kind of listening the way I was and like watching the slides. But this time I really was. I was, it's hard not to multitask when you're at a virtual event. And there is a little bit of that that happens over time because it's hard to sit and focus, I think, for so many hours. But there was a lot that I really just sat and focused and listened to. It was really amazing. High level, I think there were some really big themes coming out of this one, which is different from what I've been hearing before. You know, yes, there were case studies, yes, there were methods. So, I encourage everybody to watch on demand, catch up on case studies, catch up on all there is to learn about AI moderation and synthetic data and the way people are using AI. But yes. Ethics and governance, but there was a big, there was a bigger story, you know, a higher arching story, which is like really in three parts. One, I wrote this down, so it's gonna sound a little bit like written down because it is. Accountability doesn't transfer to whatever machine you're using, whatever AI you're using, right? Can produce the output, but the person, the researcher, still owns that decision. You are accountable for whatever the AI is doing. Remember that and own it. Like, remember, you are accountable as the professional, not the AI. That was a very loud and clear theme that we have to stay mindful of as professionals. That is what's going to ensure our professional credibility. Let's not lose it. Let's not lose our professional credibility to AI. That was a really important theme. Another theme was that. And this is related: our job is shifting, our job as insights professionals, shifting from kind of checking the work, which we have to do, but to checking the thinking about how we're using the work. Pam Forbes at Mondelez spoke first thing yesterday morning, and she said, Remember, we are AI's teachers, not AI's users, which was really, really like one of those sound bites that you're like, yes, we are teaching AI. And then at the end of the day, Lisa Cordade from Oregon joined us, you know, one of our oldest fans, which is really fun to have her on. Always awesome. She said, cognitive offload is fine, but cognitive surrender is not. So, like, remember, we are, we are doing. The thinking there, like it was just that was that point was just loud and clear throughout this event, also. And then the last thing that kind of came off loud and clear also is the value, like you know, speed, like that's becoming table space. Everybody can do things faster now, end of story. Um, you scale, volume, that all can, you know, that all can happen. Um, you know, cost, that's all coming down, like the value equation is changing. But what is worth paying for is the judgment, the critical thinking, the aid to decision making, the consultative thinking. You know, we were talking at the very end of the day at the town hall that what is worth paying for if you're on the supplier side is, you know, that almost high-level McKinsey-like thinking. Like you have to really, you know, dial it up because people are going to be paying. You for your ability to give clarity on business decisions being made. So anyway, that's my high-level takeaway of the event. Sarah Snunda recorded a full debrief for the podcast, releasing it next week. So she and I will dig deeper. And yeah, we'll have a deeper, more than five-minute debrief later today. And those themes will probably be the top ones we dig into. Super cool stuff.

Lenny Murphy: Good. And those are for what it's worth, right? Every conversation I have now with everyone, everybody.

Karen Lynch: Yeah.

Lenny Murphy: All that's echoed, right?

Karen Lynch: Yeah.

Lenny Murphy: You know, there are the stacks we've always talked about. There is data, there is process, and there is thinking and impact. And at the pyramid, it's the thinking and impact that's at the top. We may get more volume from a business standpoint on those other ones. Absolutely. And businesses do that.

Karen Lynch: But the critical piece that everybody is prioritizing is it doesn't matter how much infrastructure and speed and efficiency and cool stuff we do, it's the humans, so yeah yeah just adding jazz jazz is listening to that was echoed through the case studies, uh yeah yeah just just sharing with the world jazz's uh takeaway also jazz cognitive offload is okay cognitive surrender is not was the perfect mic drop moment at the end of yesterday's town hall for real. It was really, really great. We are with you, Jazz, with you, 100%. So, yeah, really good stuff.

Lenny Murphy: And at least it's good at those mic drop moments. Yeah. She's, uh, I can't believe they are. She is, yeah, she's great.

Karen Lynch: Yeah, it was really, really good. Barry Jennings was there, and Dave Dugan from Dish Network was there. And yeah, it was a great town hall. So, yeah, catch it on the on-demand replays. Really good stuff. It's there. It's going to be on-demand for about six months. You can still catch it. So, really, really cool. So, next up, Sciax West. Yeah, it's great. It's really exciting. Not just AI. Like, yes, there will be AI on the agenda, of course, but lots that's not AI on the agenda because we're switching into AI just being a tool in the toolbox and just one of the things, but not the thing. I think it was at the very first session, also, at the AI event, like where his name was, you know, Abran Maltonado. He was saying that, like, if you're, you know, if you're a creator, if you're, you know, if you're, if you're a director, if you're an artist and you're making a movie, you know, you're not, you're not showing what tools you use to create the end product. You just look at the end product. You know, nobody's saying what kind of camera you used to film that when you're sitting there watching, you know, the IMAX Christopher Nolan. You're not saying what tools you used. You just want to, you just want to look at it on the big screen. That's what the researchers are, right? We don't, sure, AI is one of your tools, sure, but we need the outcome to be really what we're looking at here. And anyway, there is AI as a tool, and we're going to be talking about a whole lot more at IAX West. So get there, for real, get there.

Lenny Murphy: Yep. Well, and speaking of tools, that's yeah. Yeah. Well, it's probably a segue for the uh to get into the news and get into the news.

Karen Lynch: As always, there is always. Yeah, there's so much. So why don't you just start with, you know, our good old friends?

Lenny Murphy: Sorry, it wouldn't be an exchange unless we a friendly bid for the U group and receipt jar, basically receipt scanning out of Australia. Absolutely makes sense. We've talked about how these four behavioral data are emerging in importance. Utility now to buy something like that. And I will make a prediction. I will make a prediction. It has been widely buzzed for quite some time that Numerator would be something that Bain would do something with Numerator, Bain and Cantar, soon. And the question is: who would buy them? Because it would not be, it would be a big, you know, like multi-billion dollar acquisition. Yeah. And I. And NIQ. Or Sirkana. Here's my prediction. We'll see what happens next year if that occurs. But my bet is NIQ or Sirkana. All right. Logical buyers for numeria. Because NIQ just showed, you know, yes, get this data, pull it together. Yeah. And it makes an awful lot of sense.

Karen Lynch: So you heard it here, folks.

Lenny Murphy: Heard it here. You heard it here. And I know nothing.

Karen Lynch: Yeah. This is not inside our knowledge.

Lenny Murphy: Not inside our knowledge. I am just, I'm just guessing.

Karen Lynch: I'm just speculating.

Lenny Murphy: Yep. That's where the pieces fit.

Karen Lynch: Yeah. Well, the team enumerator is probably sick right now. They're like from your lips to God's ears.

Lenny Murphy: Yeah. Yeah, we'll see. What about our friends and Susie? Do you want to talk about that?

Karen Lynch: Yeah, well, so, you know. So two things going on here. New leadership, right? New leadership, new funding. Growth partners led a new round of funding for them. But the new leadership, I think, is really interesting. You know, I feel like we've talked about them recently. I don't know when, but, you know, they kind of were ripe for a refresh, in my opinion. You know, kind of ready for like, all right, like we haven't been a little bit quiet, you know, like sure, on-demand consumer insights, this has kind of been in their wheelhouse. So that's what they do. And so this funding is going to help kind of continue their AI first product evolution. But what I like about this leadership change is Brian Silverman started a company 11 years ago. This isn't bringing somebody in from the outside. And he started in sales, right? Sales development. So, you know, Matt Britton, who was the founder, is kind of moving on, non-executive chairman. And they're bringing the business really well up and into the position. And so it doesn't mean massive change, right? It means somebody with deep knowledge of the company. And it also speaks to the company values, right? So, anyway, I applaud them. I think there's something about that. Anyway, to me, that says more about who the company is and the deep, deep knowledge of the kind of the, you just self-awareness and moral fiber of the company. So, you know, wishing them luck with that new funding and for how they are evolving the company.

Lenny Murphy: Yeah. I thought exactly the same thing, right? And shifting into, of course, I guess we should mention that they want to be the integrated hub and arguably they have been. I mean, they have lots of capabilities and are putting more AI tools around that to make it more operational. So yeah, look forward to seeing what they do. And congrats to Brian.

Karen Lynch: Yeah. Yeah. Cool stuff. Cool stuff. So yeah, speaking of contract, congrats. So Zappy, I just think that there's a story out about Zappy right now that I thought was so interesting because it's venturing into another world, right? And anyway, so what I'm talking about is, you know, Pixart, right? So we're talking about how Creative Production has now partnered with Zappy. They've collaborated on something called Vera, which is an AI agent now in the kind of PixArt agent.  Marketplace letting creative decisions be kind of grounded in data and consumer insights. So research kind of moves out of just research into creative execution. And I just think this is really cool. So the way to go zappy is kind of the way to go zappy. Way to go zappy, and just another example of the way this industry is evolving into, you know, other areas embedding insights into their workflows.

Lenny Murphy: Yep. Yep. I think we're going to see more and more of that. Forever, I've thought, why Zappi, Steve, and I talked about it. Why are you not in the middle of Shopify or whatever, right? I mean, you could just see this, and that's happening. So these applications are being embedded. Into workflows.

Karen Lynch: So very cool. Well, and it reminds me of, you know, the conversation we had with IDX, with Crispin Beale recently. It's like we should, all of the decisions for any kind of creation, whether it's copy or packaging, like we've been talking about, Insights has informed this stuff, you know, for as long as we know, but they've contracted insights providers, right? Whereas now let's just hard bake it into the people doing that work. So think about what those possibilities are. Think about the possibilities. Anyway, it's really cool to think about that evolution. I really like it. It was such an interesting time.

Lenny Murphy: There was so much about AI agents and shopping and measurement and even more the hit today that'll go into next week, I guess.

Karen Lynch: But yeah, yeah, yeah.

Lenny Murphy: But Dunnhumby is investing. In uh in a measurement for AI shopping agents, which makes perfect sense. Um, I'll tell a quick story. This or did I tell this last week? No, I don't think I did. I trying to make a purchase for business, business purchase, and was using my agent to do that. And it, the site didn't work with the agent. And I got frustrated and said, no, then go to another site because this pain in the ass. Did I tell that story? I don't know. Anyway, the point is. So this is happening. I've now experienced this directly on what a pain in the butt it is when an agent cannot complete a sale that I wanted to do. And it has to be measured. So I've spent how many years doing that along with like 20 other companies.

Karen Lynch: Yeah. It's such a big deal. Yeah, it's a big deal. And, you know, for. If you think through this, Dum Humby owned by Tesco, which is a multinational grocery retailer, think through what this all means: that if you level this all up, that means this retailer is thinking about the AI agents at play. And that is a signal to everybody that these retailers are. Like fully aware of agentic shopping disrupting their businesses. So God knows if the retailers are focusing on it, the brands are too, right? Like we know this, right? So, to all the listeners in our audience, if you're on the brand side, we know you know. We hope you know. And if you're on the, you know, kind of supplier side, you know, start to think about how you're going to help your clients sort through all this.

Lenny Murphy: In my and and tracking that is important and with the my understanding is that the the meta launch of their their personalized system their agent system that that's going like gangbusters so and we hadn't seen that yet we hadn't seen an agentic system designed for consumers yeah leave it to meta to launch that so I I think we'll see pretty rapid expansion now uh from that standpoint, try align it. So, yep. And groceries are real, that's a low-hanging fruit.

Karen Lynch: Yeah.

Lenny Murphy: Yeah.

Karen Lynch: So, yeah. Yeah.

Lenny Murphy: Very cool.

Karen Lynch: Yeah. It is very cool. I don't like, I could, I could segue into that meta thing right now, but maybe we hold off, or do you want to go down that rabbit hole with me right this second?

Lenny Murphy: It's up to you. I'm watching the clock for you. I know that you've got hard stuff.

Karen Lynch: Well, I don't know, folks, you know what we're talking about. We're going to talk about this later, but did you, did everybody? I did not. I'm Tim. Tim's all in on Meta right now. And Muse is the AI agent that Meta launched this week. I'm conflicted. I'm very conflicted by this. But it is very user-friendly. It is, somebody was saying this Apple should have done. This is a very Apple-esque, user-friendly AI agent tied to. You could do avatars, right? You could do avatars. You can. It's very friendly. It can, you know, show in your email. It gets to know you from your email ecosystem in about a split second. It knows you so well. I'm not sure I want all of, I'm not sure, full disclosure, that I trust that organization with so much of my personal information based on past history. Tim was like, I'm good. Good. So, I probably know all about me anyway. That's right. That's right.

Lenny Murphy: And he knows what Tim thinks about you.

Karen Lynch: Exactly. Yeah. So, so yeah. So, but you know, but apparently it's, it's kind of a big deal. And a lot of people are chattering about it. You can see it online. I'm sure Reddit's having a field day. I haven't gone down the rabbit hole yet, but that's what we're talking about when Lenny and I are talking about Muse.

Lenny Murphy: Metas, Metas in the Game.

Karen Lynch: Metas in the game. Metas in the game.

Lenny Murphy: And they doubled down on wearables this week as well. That didn't make it into the news and it doesn't need to, but think about the implications. You have a personalized agent, and now they continue to have this vision that wearables is going to be where that interacts. I certainly could see the opportunities. I'm walking down the road and I see something. Oh, agent, find that for me, whatever.

Karen Lynch: Yeah, yeah. Um, I just have to message our producer a minute in Slack. Um, anyway, um, yeah, we're getting the spammers again, so yes, yes, just uh, I'm gonna check Slack. She's on it. Okay, so yeah, so we can talk more about that, but I haven't explored it. If you've explored it, send us an email at theexchange at greenbook.org and let us know your experience with it. Or, you know, tag us on a LinkedIn post if you've explored it, because I think Lenny and I would probably be a little slower to explore Muse, but we know it's a thing and we know it's a big thing and it's a big deal, but we both have those, you know, anyway. Yeah. But it's a big deal. Early adoption is out of control.

Lenny Murphy: Utterly disconnected from all the meta.

Karen Lynch: Yeah, exactly. Exactly. So, but yeah, like you're gonna share, you know, yeah. 730 downloads in five days. It's a big deal. You know, there are 730 million consumers. 730 million downloads. I mean, it's a big deal. Consumers. We talk about B2B quite a bit and what's happening in the B2B space, but this is consumers. They are going to adopt this much faster than some of the other platforms because it's made to order for them, showing up in places where they are. It doesn't, it doesn't take much. I mean, I'm on Instagram, I'm on Facebook, and like, hey, do you want to do this? You want to do your muse avatar? And I'm like, no, no, I do not want to do that. But it looks very cute right away. They made it very easy for me to consider it. I just didn't want to because I was aware of what was happening there. So anyway, live without a cat. My dogs are barking.

Lenny Murphy: Right. Well, let's talk about the

Karen Lynch: Product launches. Um, yeah, yeah, you start. I'm gonna have to give a dog a bone. Quite literally, give a dog a bone.

Lenny Murphy: All right, all right, Stravito, which we just talked about a few weeks ago. They're on a repair. Um, we had kind of promised that we were working on MCPs anyway. We are, um, but it is indicative. Uh again, this shift towards workflow integration and Stravido has done that with their glass box approach, which I think is really interesting. Integrate directly in ChatGPT, Claude, co-pilot. So that was interesting. But then Numerator, right on top of that, Numerator launched their conversational access to their intelligence. Inside client environments. They haven't done that before. It was a very different type of buying process. Now the numerator is integrated into the flow where client heat is. The ask nicely, which I had not heard of

Karen Lynch: Yeah, yeah, no.

Lenny Murphy: Yeah, kind of a CRM platform integrated directly into the cloud with an AI assistant. That's just, it's non-stop. Yeah. That is now rapidly becoming, I was talking to the CEO of a very large supplier in the industry yesterday and who, well, CEO, anyway, senior person, who said they were now being flooded with brands wanting the agenda procurement process. And I told them the story of like, yeah, about 18 months, they started asking 18 months ago, they were saying. You know, be ready. Here we are, right? And this is all part of that. Because the MCP by default becomes an agenda procurement process. So, yeah. So, all part of that.

Karen Lynch: Yeah. Yeah. So, no, I know. I think it's, I think it's good stuff. There's, yeah, it's so interesting. And yeah, I don't know what to say about this MCP stuff. I'm sorry we were going back on our word, but you know, yeah, it's the way it's the way it is. It is the way it's what's happening, just like you know, what con what CommScore is doing by you know, measuring advertising inside AI answers. So, like chat-based advertising, man, and people are figuring out how to measure it. So, that's, you know, we know that that's happening. So, um, you know, there's just different ways of measuring it, you know, showing up left and right. So, um, yeah.

Lenny Murphy: Sorry, Karen, I didn't even talk over your point.

Karen Lynch: No, no, we're just going to see more and more capabilities, I think, of how to measure this, how to track this. You know, until we get solid measurements and solid benchmarks to kind of gauge what's happening in this advertising, I think, you know, sponsor-based, chat-based advertising is what's differentiating. Can you stop? Anyway, that's what's differentiating this one, but it's all right. I really want to talk about Marvin next. So, anyway, what else you got about ComScore? What's your other thought on ComScore?

Lenny Murphy: No, that's yeah, right. We talked about measuring agentic purchasing. How do you get there? Right? Pat's purchase starts with the search and the And our definition is previously we think, oh, bots are coming onto my platform, and that's bad. No, it's actually now it's not. I mean, it depends. There's good bots and bad bots, but the reality is that we're going to see through this process, you know, far more agentic discovery. Search, procurement, and activation bots that are managed and higher paths are purchased. So, and this is a piece of that. Yeah, let's talk about, hey, Marvin.

Karen Lynch: Hey, Marvin. So, listen, here's the deal. You know, at first it was like, we don't need to talk about another, please stop barking. You're ruining my group. Enough. She's a little devil, this dog. At first, I thought we don't need to talk about another AI moderating platform, but then I really looked at this very closely because I'm like, or do we? Because this is in the moment, intercept AI moderation. So the deal is automated interviews right inside digital products and websites. Kind of, let's get into it. You it's a little different than just kind of pop-up surveys on the site, right? Like now we're talking about moderated qualitative feedback in the buying experience, in the online buying purchase experience. It's not, it's not recruits, right? This is this real-time kind of intercept, intercept in context research, another kind of shift in the, you know. Online AI moderation. Um, it seemed like, yeah, it's a natural, natural iteration of what we're doing there. So, um, check out what Marvin's doing. I'm like, that makes perfect sense to me once I really dug deeper and thought through it.

Lenny Murphy: So, yeah, yeah, it's the exact same thing that makes perfect sense. The, yeah, we already get that with, oh, I'm your agent, I'm here to help you, right? On most websites, like, okay, well, hey, give us feedback.

Karen Lynch: So, why not? Absolutely. Speaking of CX and feedback. Speaking of that, Yes, exactly. Yep.

Lenny Murphy: I think that the Qualtrics model is becoming clearer now. We see what they're cooking, I think. Do you want to talk about this?

Karen Lynch: Well, sure. So, you know, Qualtrics is, you know, pushing deeper into healthcare. They've been talking about this for a while: combining Qualtrics tech with Prescani data. AI-enabled experience platforms. What I thought was interesting when you read this press release about it, I'm trying all right, what's new from what they've been talking about for a while, this deeper push here? Jason Maynard, who's the CEO of Qualtrics, says, when AI changes what people expect faster than healthcare organizations can change how they deliver care, the consequences extend far beyond financial or competitive risk. They affect trust, health outcomes, and people's lives. So I think that's really the story here: just how serious the implications are and the industry awareness that Qualtrics is really approaching it with. This is high stakes and of, you know, some might say more importance than almost anything in, you know, almost every other category out there, right? We're talking about life and death, quite literally. And I think that, you know, if we can close that gap, I think more power to them, you know, so. Let's  take that data and do some good with it. Good luck, Jason. Um, I wish you well. Let's do some good in our healthcare, healthcare ecosystem.

Lenny Murphy: I think Qualtrics wants they're going, they're choosing to be the infrastructure, um, the data infrastructure, at least from a customer standpoint. In that entire category. And, you know, you go back, it's even like the SAP, when the SAP first bought them, right? The idea of linking these data types made sense on the surface. Now they didn't actually ever execute. I don't think they ever pulled that off at the level that they wanted to.

Karen Lynch: Along comes AI, right?

Lenny Murphy: I mean, it's the technology that enables this. And I have, I have little doubt. That they will pull it off and it will be a good thing for everybody, especially the patient, to have that information there. So, yeah, go Qualtrics.

Karen Lynch: Go Qualtrics. You know, one of the things that came out of the town hall yesterday is that. What buyers are going to be looking for is industry expertise. Like they really want people to know, they want the people that they pay, they want the consultants that they work with, they want the vendors that they contract with to have deep industry knowledge. They want them to know their business because they are the insights profession, the internal insights professionals are charged with. Strategic business decision making at this point, right? Informing the C-level executives, making strategic business decisions. So, if you think about that, right? The healthcare institutions who have to act on this data, this Prescani data, and make really big changes, they need Qualtrics to know their business. So, if Jason and his team are going to double down on healthcare right now and really become that. Experts in that industry. It's also addressing the kind of that buyer needs, which is you better really know our business. Absolutely.

Lenny Murphy: And know our customers. I mean, bear that in mind. This is XM. This is experience management. And let's face it, healthcare is a shitty experience. I mean, it's not a great experience for so many reasons, right? Some are not even within their control, but it's just not.

Karen Lynch: Great.

Lenny Murphy: Anything that helps optimize that and have a better experience that should lead to better outcomes, et cetera, et cetera. So that's a win-win across the board.

Karen Lynch: Qualtrics just wants to mention on the pit the way Prescaney got in that first episode of the pit. I don't know if you have no idea what I'm talking about. There, but like season one, episode one of the pit, they mentioned prescaney and you know, so Qualtrics just wants to dive in.

Lenny Murphy: So on the pivot is basically just a sequel to ER, but a little more green. Is that kind of, yeah, all right. No wireless connection.

Karen Lynch: It's like, I know it's, it's, it's like, you know, a cross between ER and 24 with Jack Bauer, basically. Okay. Anyway, ER means Jack Bauer. Yeah.

Lenny Murphy: I had to quit watching the art because it's like every, I would, I would just cry like a baby.

Karen Lynch: Oh, yeah. Well, don't watch the pit then because you know, no, like what kids?

Lenny Murphy: No, anyway. All right. All right. The uh tech development.

Karen Lynch: Yeah. Well, we already talked about Muse, so we can kind of take that one off, but you should talk about these other two. They're big.

Lenny Murphy: They're like OpenEye, Anthropic, Grok, you know, all release new models. And I saw something the other day. It was tracking the time compression, right? Used to be like six months, right? Yeah, and now the time between release of new models is like 17 weeks, yeah. Um, the uh so they all three released new models; they all three. The prices came down, yeah. So let's keep that in mind. Something that I was on to like, oh, the computer's going up like no it's not no we are rapidly getting towards something that is approaching zero on the computer bus. Yeah. So, and all the intelligence, it just amplifies capabilities. I'm sure you see it every time there's a new release and it's like, damn, this is really for the work.

Karen Lynch: Yeah, better. I agree.

Lenny Murphy: Yeah. So, so, which is challenging from a business standpoint, right? Well, it just built on this model. And now I have this model.

Karen Lynch: I know, don't get me started. All my custom GPTs, I have to migrate them now. And are you kidding me? It's just making me a little bit crazy. And, you know, this is what's going to force us to switch between the platforms. I'm like, all right. And Lenny, I might like joining the all-in-on perplexity projects because I have a few of those now too. I might be changing everything. So, whatever. We'll see. All right. I will. I know. I know.

Lenny Murphy: You gave this shout out. I will give that shout out too because I don't have to worry about it. So I know. Perplexity, choose the right model for the task.

Karen Lynch: I know. I know. So, anyway, but no, I have some projects there that are going very well for me too. So, anyway, we'll see. Switching is hard. It's almost like switching banks at this point.

Lenny Murphy: I I know I I get it I I get it 100 imagine if you were like suddenly to twist it's like we've got all this project work and you hate to lose but I mean hate to leave but anyway all right I well I hit what one tip i've gotten in the habit if it's a really important thing I download the markdown file oh yeah google drive yeah yeah so I always have access to the historical stuff just in case I know. And on that note, yeah. And now we talked about Jev, but we got to talk about it again because the TypeSafe AI is Jev, the decision AI, right? It's not about chatbots.

Karen Lynch: Yeah.

Lenny Murphy: The Versal AI Gateway's fastest adopted model with nearly 13% of paid teams trying it within 24 hours. So that. That shift of now we're getting into managing decisions in workflows, and that's all that it does fast and cheap. I think we're going to see within six months, that will be just a standard built into all these models, and we'll just see a massive acceleration into decision enablement. And then productivity is really going to take off.

Karen Lynch: Yeah. Yeah. So, yeah, crazy. Crazy times. It is. All right. Quick, quick, because we're going to run out of time. We've got three things that you all can like to read deep on, right? This is good stuff. This pew, this Pew research, just when you think we're done talking about data quality for a little while, it pops back up, right? So this was good research. How deep did you go into this study?

Lenny Murphy: Yeah, I. I definitely dived in. And Pew's the gold standard, right? I mean, for them to be able to say, watch your data. A lot of it's unreliable. No matter what you do, there's still a certain amount of shit that still gets through the filter.

Karen Lynch: And I mean, I think you just said it, no matter what you do. I think the point of this Pew study is that it. There is no single screening method that is reliable, whether it's, there's fraud, there's bots, there's AI-generated participation, which is, I guess, a way to explain people using AI to help their participation, like, people who are not frauds and not bots, but they're using AI to assist the AI-assisted participation. The data quality challenge is just getting. Stronger, and there's not just one thing to look out for. So, anyway, you know, that is, you know, it's just intensifying. The data quality challenge is intensifying.

Lenny Murphy: Nelson was, Elsa and I were slacking last night.

Karen Lynch: Nelson's research director on grit.

Lenny Murphy: Nelson's our research director on grit. He's, he's more than that. He really is the grit guru now. Anyway, point is: going through the data, and it's like, the amount of bots the amount of ai even within grit researchers you know deploying this stuff yeah I mean that you would think that we would be somewhat immune to it nope yeah there's still those challenges exist within uh even within the grit survey yeah yeah yep I uh this article article um that I thought this is the type of independent research a validation benchmark on synthetic users. The 86% recall, that number keeps coming up a lot. That seems to be, I know some company, and I totally buy that in specific use cases on specific high-quality data, that number can go up. But this seems to be about. The level that most folks are operating is about 60% accuracy compared to reality. And that just means, so recognize that. There's a lot of stuff where 86% is just fine. So, but if you need higher accuracy, synthetic, you know, at least the broadly available synthetic models may not be the right model.

Karen Lynch: To deploy. We used to say all the time we'd have these qualitative caveats in every report that we did, which is directional. You know, it's, you know, this is directional. And, you know, we just probably have to, and all of our, all of our work we're using synthetic is just let's get our caveats in place. You know, this is not necessarily this, you cannot glean statistical inferences from this the same way. And we just have to. Lock that down pretty early and find the right language. And what is the proper way to communicate the risk that we're taking? We cannot project this to populations necessarily the same way we project quant data. It's just not the same.

Lenny Murphy: I'm saying. Low consequence decisions.

Karen Lynch: Yeah.

Lenny Murphy: The lower the consequence, the more appropriate these things are, especially discovery and early stage ideation, all that good stuff. The higher consequence, we still need to make sure that we've got rigor. And that I thought this last piece, this Ethan Moloch article, was just really timely because of the gap between what AI can do and what we as humans and organizations are ready to do in transformation challenges. Yeah, what was your takeaway?

Karen Lynch: Yeah, no, the same. And it's a great bookend to how we kick this off, debriefing the AI event. Ethan Malik is talking about AI's overhang, which is really this capabilities overhang, talking about this gap that Lindy's talking about, like what AI is, what these models are capable of doing for us and what we're actually doing with them. It's like this big gap. So, you know, I think that the constraint is our ability. Not what they can do, but our ability to have the judgment to discern what it can do for us? How do I want it to work on our behalf? How do I guide it to do what's truly valuable for me? You know, how do I, how do I, how do I stay, you know, how do I, how do I? How do I put my judgment over it? How do I make sure that I am the one using my critical thinking skills to make sure it's delivering what I need it to deliver? How do I stay, you know, how do I stay, you know, the consultant over it and serve my end client, whether it's an internal stakeholder or an external client? How do I serve my client best by using this as a tool? And push the boundaries of what I'm using it for, because it can do so much more. I think it's pretty cool. So, God loves Ethan Mollick for doing that sort of thing.

Lenny Murphy: I mean, yes, you know, and every week, experience that. I mean, even the, you know, the audience that Karen is experimenting with how we come up with our summary, right? Things and the balance between compounding knowledge, right? Because of these systems, we put all the links in and then filter. They help us filter down. And that I always check myself, like, wait, I need to make sure that this actually fits and works. And it's getting better and better. Most of the time, I don't have to change it at this point. Like, yep, that's right. I agree with that. That's the way they do that. But it's still. It's a constant learning process.

Karen Lynch: Well, what's interesting about it, and you know, again, we can share. My filtering is based on my audience at Green Book. I serve Green Book solely, right? Like, it knows who the Green Book audience is. You serve a different audience at both Gen 2 and IOV now. So, like, you serve Green Book, yes, you know, but your primary work is. Is slightly in slightly, you know, you have it like a Venn diagram. So I think that's interesting, right? Because our AIs learn from the work that we do. And I just find that fascinating. And so, yeah, I mean, we are always evolving, we're always iterating and we are always learning. And I think it's great also to have thought partners like each other and that we're sharing this knowledge. So we are thought partners not only with our AIs, but with each other.

Lenny Murphy: Oh, 100%.

Karen Lynch: I learned so much. Great way to end that show. So, as I wrangle my dogs and get ahead of the inevitable doorbell ring, as somebody comes to pick up some babies that are in my house.

Lenny Murphy: Ah, all right. Well, then, on that note, and apologies for the audience. I'm aware that there's probably some lag.

Karen Lynch: Good. All right. It wasn't so bad.

Lenny Murphy: Have a wonderful weekend. Talk next Friday. Enjoy the rest of your afternoon.

Karen Lynch: I hope you're not too cold now that it's officially fall, everybody.

Lenny Murphy: And I was so happy to dig out. A jeans and long sleeve shirt and boots guy. That is my, that's my most comfortable, but I can't wear that. Six months out of the year. No, I know.

Karen Lynch: I'm like a jeans and t-shirt kind of girl. So I don't necessarily, I'm not necessarily sweater season, but I like a good cardigan because I can put it over a t-shirt. So I'm, this is a sweater sweater. I'm officially, it's fall, man. I have to wear a sweater. So I don't love it. Winter's not my favorite and it's coming a little bit too fast for me. I must, I love summer, but it's fine. I'll only be sad for the next, what, you know, nine months. That's fine. This is the time. You get to watch out the window and wait for the changing leaves. I think I have one tree out there that's starting. You can see a little hint of a tree out there that's starting.

Lenny Murphy: Yeah, they've started Kentucky with a few. About two weeks ago, actually, things.

Karen Lynch: Well, here in Connecticut, we're holding on to some green. That's probably 'cause we have maybe more moisture up here.

Lenny Murphy: But look, I still got home and moved the grass.

Karen Lynch: So, anyway, all right. Have a great weekend, everybody. We'll see you next week.

Lenny Murphy: Thanks, everyone.

Karen Lynch: Take care. Bye-bye. Take care. Bye. Peace out.

Links from the episode:

dunnhumby ventures invested in Azoma 

Stravito Launches MCP Server for Enterprise AI Tools 

Numerator Launches Conversational Consumer Insights In-Platform and Across Client AI Environments 

AskNicely Launches Claude Connector, Bringing Live Customer Feedback Directly Into Claude 

Comscore expanded AI Intelligence to measure sponsored advertising in AI chats 

Marvin launched Live Intercept 

Qualtrics launched Healthcare XM 

New models from OpenAI, Anthropic and xAI continue to expand the AI model market while reducing the cost of the technology powering AI workflows 

Introducing Grok 4.7 

TypeSafe AI’s Jev became the fastest-adopted model in Vercel AI Gateway history 

Meta’s Muse AI agent reached 730,000 downloads in about five days 

Pew found that no single screening method reliably solves the problem of bogus respondents in online opt-in surveys 

Articos reports that its synthetic-user research achieved 86% theme recall against expert research findings across 46 studies 

Ethan Mollick examines the “overhang” between what today’s AI can do and how much people and organizations have actually learned to do with it