Inside Morning Consult: AI and the Future of Insights

In this CEO Series conversation, Leonard “Lenny” Murphy finally sits down with Morning Consult CEO Michael Ramlet to unpack how a scrappy, $30K-funded startup became a global real-time data engine reshaping the insights industry. Ramlet shares Morning Consult’s origin story, how they now run 30,000+ interviews a day across 43 countries, and why owning the sample supply chain is central to data quality. From the “Bloomberg terminal of public opinion” to MorningConsult.AI and unlimited questions on demand, they explore what AI, synthetic data, and always-on tracking mean for brands, vendors, and the next generation of insights leaders.

Transcript

Leonard Murphy: Hello everybody. It's Lenny Murphy with another edition of the CEO series. Thank you for taking time out of your day to spend with myself and my guest. And today it's kind of special…

Michael Ramlet: Thank you for

Leonard Murphy: because for many years there's been this company called Morning Consult that I really wanted to talk to the CEO and it didn't happen. And then all of a sudden over the course of the last few months, Michael and I did have a chance to finally connect off the record and I twisted his arm and said, "Come on, We got to have a conversation for other people because you're such an interesting company." So, Michael Ramlet, CEO of Morning Consult, welcome.

Michael Ramlet: For having me. Thank you for the persistence. we try to keep a low profile for kind of the first decade, but I think we're excited to tell our story and talk about sort of the future of the industry.

Leonard Murphy: It seems like you've been flexing quite a bit lately. So absolutely. let's get into the origin because I think a lot of folks probably are not as familiar with Morning Consult as some of the other companies. you said you've been around for 10 years. so give us your superhero origin, Michael.

Michael Ramlet: So, the origin story, humble beginnings, we took 30K in Angel Capital.  It's not a traditional venture start. I had two incredible founders, one in Kyle Drop, who had just finished his PhD at Stanford and at the time was technically a Dartmouth professor. and then I convinced another co-founder, Alex Dulan, to leave his Chicago PhD program and his background was computer science. And honestly, it started with a problem. We were obsessed with what we thought was bad executive decision-making on limited data. So whether it was government leaders, we're shocked at what felt like this sort of big data revolution that was talked about hadn't actually materialized. And what we were obsessed with was this idea that you should know what people think in real time, whether you're making decisions about a business outcome, the state of the economy, a kind of geopolitical environment.  And so we had this obsession with you should have real-time data and that became a mission to inform better decisions. And so that really launched, we've talked before our obsession about the supply chain and the importance of getting higher quality interviews. And in the emphasis on quality, we discovered how much inefficiency existed and that actually also allowed us to remove the financial barrier to going every day. So I kind of think about the company in three stages. The first stage was bootstrapping. So first five years you learn really quickly the difference between cash and accounts receivable. and the realization was once we could go every day in the United States to collect data. We also then want to take that capability around the world.  And so the second stage of the business kind of second five years we raised outside capital from three incredible investors and allowed us to go from daily data collection US to today we're doing over 30,000 survey interviews in 43 countries and every day we're asking folks deep audience demographic questions. We ask them what they think about a number of brands. At this point it's now over 5,000 brands in those 43 countries. And then we ask them about the economic environment and variables where they live and the geopolitical environment. And so all of it really traces back to this real-time information on what people think to make better decisions. it's been a series of really exciting events along the way.

Leonard Murphy: Yeah. I think many folks, maybe I'm projecting here, but when you first launched, I think, that's really cool. That's a new approach to polling. and I bucketed you wrongly obviously in that kind of geopolitical space. And then just kind of checked out a little bit. So right the I mean Mhm.

Michael Ramlet: We're grateful. We're grateful. I think one of the things that we were shocked by was the lack of investment in the industry and really understanding the supply chain and the tech stack necessary to make the advancements both in terms of quality, speed, and frankly at the end of the day efficiency, cost. and so I think we were really scared that anyone else would follow what we're doing. today we're probably the world's largest end buyer of sample.

Leonard Murphy: Yep.

Michael Ramlet: And I say end buyer because there's a lot of buying and reselling that happens between the actual person who's doing the research and the sample respondent. And we removed all of that friction. And for us it's our reputation at the end of the day. We work with a quarter of the global 2000 companies.  We work with almost every major central bank, pretty much every major hedge fund, and when we're giving data to the Fed, they're making decisions on interest rates that are highly consequential society. And so, we better have the highest quality data and just a overwhelming emphasis on our end around quality. And that allowed us to, I think, make the investments and care about the details that are oftentimes missed. There's too much sort of polyianish head in the sand. Data quality but data quality isn't a press release. It's actually building the text stack and having the behavioral data to validate it.

Leonard Murphy: Yeah, the audience what again Michael I did have a chance to talk about two months ago and I did get a kind of a privileged view I think to an extent of a little bit behind the scenes and I'm telling you the machine that they blew my mind. it's like, "Wow, you've done all that. That's cool."

Michael Ramlet: And I think obviously current events sort of necessitates to come out is the issue isn't like there are some incredible panel companies and panel executives out there that are building really great businesses. but the idea that the onus on quality is strictly on them I think is unrealistic and frankly misplaced. we have to own quality organizationally at morning consult. one of the things I was, showing Lenny and I think is important for all of your audience to think about is it's not just about the steps that you can control quality around,…

Leonard Murphy: Mhm.

Michael Ramlet: Identifying bad respondents or inattentive respondents. Ultimately, what matters with quality is can you predict business outcomes. And so, one of the things we were very focused on is there are real world behavioral outcomes that we can benchmark against. You're thinking about making a purchase or staying at a hotel before you actually go.  So what we did is we took all of the 5,000 brands we laded them up to the publicly traded tickers and we back tested all of the financial metrics. So SEC reportable financials and you can see how predictive our tracking of purchasing consideration might be for hotel hospitality brands and how predictive it was of the next quarter's earnings call and results. And that's what I think gives so much more currency to the insights function to executives. you should be able to walk into the CFO's office and have currency and trust and that's going to improve investment in the insights industry on the brand side. It's going to improve the quality and operations of I think the vendor community that just didn't make those investments. There's some very big I don't get invited to the holiday parties for the industry for very good reason. I just don't think that legacy companies did what they needed to do over the last 20 years to actually bring this industry forward from an innovation standpoint.

Leonard Murphy: And, so much of that was based on not to defend, but there was a business model that many of these companies were built on. and that's the point of innovators like you that come in. you don't have the inertia and the incumbments of trying to protect an existing business model. Instead, you're making a better mousetrap. and I would say that's the story of

Michael Ramlet: And I think and I think that's now the challenge for the brand side leaders is you're going to have to make decisions that have high consequences for your organization going forward. for 20 years there just wasn't much in play. this is how we did our tracker. This is how we did our operations. And now in the age of AI, there are highly consequential decisions from not making the right technical choices, the right vendor choices.

Leonard Murphy: Yep. Absolutely. And that is so as we record this in the tail end of July, this is going to be live for October. But the trend that's manifesting now and will continue to I think roll out and expand as we go forward into the remainder of this year and the next is that the impact of AI not so much on the research process, we all kind of get that, right? But on business processes and the speed and efficiency and imperative to do that that is happening at the buyer level. and it is fundamentally different than anything that we've done before. And to your point, it is based on not just speed of data but quality of data that is on demand in real time fundamentally to inform decisions…

Michael Ramlet: Right.

Leonard Murphy: Because the pace of change continues to accelerate. there is it in all things. So I think I remember vividly and I'm sure you saw this as well in 2020 around May we realized benchmarks and norms don't really matter now we're in uncharted territory we need to now look at developing new norms to predict behavior because consumers were just changing and it hasn't stopped and I think that there is assumption of this normaly bias to think that no people just still keep doing what they were doing because the new normal is constant change across the board.

Michael Ramlet: Change and a piece of change that acceleration around it. I think we kind of look at it through three different lenses. So's how is AI transforming our business operations. So that could be everything from the commercial function to the finance function, all the ways in which we run the business. There is AI and how it will transform our own research process and then there is AI and how it will transform the use case and underlying access for the client and I think in every one of those ways the rate of change is accelerating and…

Leonard Murphy: Yeah. Yeah.

Michael Ramlet: That becomes both to your point earlier the innovator's dilemma the adage it's really hard to turn a battleship on a dime like it really benefits scale because those things compound but The reality is the nimleness or the flexibility to move at the pace of change.

Leonard Murphy: So, yeah, recently you launched your kind of Bloomberg terminal for insights and that was public and I'm not sure if I saw the private version, but regardless either one, it was really damn cool. And to that point of you obviously were always in the business of turning data into an asset that had multiple use cases and applications that could be productized in one form or fashion.  And now moving at least in my view moving towards that ultimate view that is aligned to the era of AI that you have a data feed that is specific around the use cases that you are focused on while also now having the ability to augment as needed with additional data feeds. So you become kind of the central hub of insights for your clients within your purview. And what a powerful and amazing thing is. and far different than the companies that are now this idea of synthetic sample. and I read something say there's so people caught up with the idea that that's just based off of everything that's online. yeah but that's why that's not good. What's good is based on firstparty data asking the questions and I think that's going to be an advantage for everybody going forward and you are so far ahead of the curve in building that capability already your clients you're just giving them a new feature effectively

Michael Ramlet: I think when we look at it today, back to our superpower origin story is that we're better than anyone in the world at collecting a high quality interview for the lowest possible cost. And so that' How do we use our superpower for good? And so I kind of look at that as the first inflection of the company. The second inflection moment is we're going to go and collect data longitudinally because up until that point it was prohibitively expensive. You couldn't do it in all these markets real time.  And so that was the sort of first primary use case was okay and became the second inflection. We're going to go collect at this point now it's over 80 million interviews. This year we'll do another 20 million. So the rate of acceleration on interviews is growing. So we've collected now 100 million proprietary interviews across those 43 countries. It's updated every day. So before either of us woke up this morning, it was already going around the world. That data programmatically processes at 1:00 am and it becomes immediately available in what became the first kind of full-scale technology platform for us and that's what we call morning consult intel. And the entire vision you're saying before is it's a Bloomberg terminal of public opinion marker research data.  this idea that you pull up any you could pull up any economic indicator, any geopolitical risk indicator, and you could go through and understand any sub geography, you could combine a question on a question, and all of those things are really powerful in a cloud-based platform. That's very much sort of the SAS software as a service approach. What became exciting about AI is, when you're building that platform, it's really hard to think about every individual user. Even though they might be in the finance function, the sight the strategy function, and a professional services firm and a hedge fund, all of those different users, it's hard to design one platform that meets every single use case and then to train them on what every knob and widget does. And so I look at the third inflection moment for Morning Console and…

Leonard Murphy: What? Okay.

Michael Ramlet: Frankly for the industry being this AI revolution.  And so two years ago, and we talked about this, this is, kind of that innovator's dilemma you're talking about with legacy firms is you face these inflection moments where you have to be willing to at least conceptualize the idea of burning it all down to make the investment in the future. And so it's not that, the power of morning console intelligence is incredible, but it's recognizing that you have to know how to go in and use the platform, and not every use case is going to have an executive that is going to take the time to do that.  And so what AI became for us and this is what morning consult.AI right now is and anyone that's listening this can go and create an account and access those 80 million interviews but it's recognizing if I don't know how to program in SQL or I don't know how to use the platform I maybe just know how to ask really good business questions and so be able to ask a natural language question like how are personal finances in the United States today or how has that changed over time. I don't need to have a background in economics to ask a follow-up question which is how about among my customer base. So you saw this with the CEO of Home Depot a couple weeks ago on the earnings call. they're not talking about the Michigan consumer sim index which is only 700 interviews. It's not reflective of the Home Depot customer. They're talking about what's the consumer confidence of a Home Depot customer. And you can pull that up on morning consult.ai just asking business questions. And so I think what we see as really exciting about that public version of the AI chat application is that it democratizes who can work with insights data and it's going to create a problem for industry stakeholders that want to be gatekeepers. The gates have fallen down. They're burst wide open. Now it's about hey…

Leonard Murphy: Yep. Yeah, how do you take what about high quality data, statistical analysis and improve the decision-making. It should empower you to be more relevant to the CFO or the CEO or other functions. And that's sort of I think the inflection moment for the industry we're at. I could not agree more. we're in violent agreement. So, the right

Michael Ramlet: And the question though to your point though is So what is synthetic data? I think this is a other part of the industry's sort of critical moment is really defining what are high quality use cases for synthetic data and what's garbage in garbage out. And I think too often you see folks pretending like, synthetic data is going to take my small sample size and make it seem great or it's going to take my old data and make it seem valuable. And I think the reality is that synthetic data is going to present really compelling opportunities to solve problems that were otherwise not possible and…

Leonard Murphy: Yeah. Yep.  Mhm.

Michael Ramlet: At a speed that wasn't possible. But it doesn't obiscate the need of really valuable high frequency. It has to be updated. the number of people I see selling synthetic data with 20 24 respondents like Joe Biden was president in the minds of those respondents and there was no trade war. So how could they possibly give you a sense of their purchasing decision-m in today's economic environment? So it's high frequency large scale it's just going to get better the more scale you have. But I think what works about things like synthetic data is how do you connect a custom study to a syndicated data set that becomes possible at a higher confidence interval and that delivers more value to a customer or how do you expand crosswalks with other data sets like credit card transaction data. The work we do with the Chicago Fed is combining the Visa, Mastercard, credit card transaction data with all of the attitudinal survey based data on commerce and retail spending. Those are really exciting unlocks with synthetic data. It's not trying to make garbage a little bit more valuable.

Leonard Murphy: Yeah, 100% agree and from the business standpoint right and think about the industry itself the looks more like something like a subscription and…

Michael Ramlet: Mhm. Yes.

Leonard Murphy: That's the first stop to ask the business question. and then it becomes an iterative buildoff of process for asking new questions or filling the gaps of information and creating this virtuous cycle of data. And no wonder that many people are trying to offiscate and castigate the idea of synthetic data because it is a fundamental threat to the business model of traditional research.

Michael Ramlet: It's going to be like this is the irrefutable trends.

Leonard Murphy: Of course it is.

Michael Ramlet: It will be faster in this industry.

Leonard Murphy: Yep.

Michael Ramlet: It will have larger scale and it will cost less. It's not like pick one. All three are happening and you have to be ready for that future. that's all we're investing is…

Leonard Murphy: Absolutely. And not future,…

Michael Ramlet: How do we either deliver higher quality scale, how do we deliver it faster or how we deliver it less expensive.

Leonard Murphy: I maybe in the William Gibson sense,…

Michael Ramlet: It's right. Yes.

Leonard Murphy: Right? The future's already here, just not evenly distributed. But it is here and that is happening. And I think it creates a really interesting now opportunity so what happens to other players in the space the large leaders in the category we'll pick on Neielson for the moment because they just went public right so okay all right so Neielson's public again and they've got a bajillion dollars to spend how are they navigating this are they looking at all right we need to be part of that Bloomberg terminal with morning consult in some form or fashion right is that we building an ecosystem of partnerships that are based primarily on data and…

Michael Ramlet: I think they're two-way currents,…

Leonard Murphy: Data feeds rather than an ecosystem based on kind of competitive ancillary process components for data collection and I think that could be really exciting and interesting. Yeah.

Michael Ramlet: there's certain things that especially our brand customers…

Leonard Murphy: Yeah. Yep.

Michael Ramlet: They're not going to want to share highly sensitive data with us for understandable reasons. They don't want to share it with anywhere outside of their four walls. And so in those cases, how do we build the APIs and MCPs necessary to make our data and the functionality around that data accessible to that brand client to be able to pull in and solve the business problems they need to solve in their four walls.  at the same time whether it's maybe smaller organizations that don't have those capabilities or knowing what we know about our data and the ability to create even more compelling value proposition for customers where do we make investments that transform the data use case and so that was the impetus behind morning consult.ai once we realized, collecting 80 million interviews costs hundreds of millions of dollars. So we spent a lot of money to collect the data, but the actual efficiency of the retrieval gener augmented generation architecture that goes by rag as an acronym is it puts in guard rails. So one of the big concerns with chat GBT or other LLMs is that the concern about hallucinations. what's great about this rag architecture, it eliminates the potential for hallucinations.  you type in an actual business question like what are the key brand metrics on McDonald's compared to Chipotle by the Chicago market, it would go and retrieve that data because we have that data. If we didn't have that data, it wouldn't come back with a response. But those are the types of things where all of a sudden the compute power is so efficient that we realize we should open it up as a free beta because we're going to learn more about what are the types of questions that a user might have that we should be making an investment in to be able to answer. And I think those are things like what can we do better than anyone else with our data and then what should we be doing to open up that data so somebody else can do great things with it. We have a lot of management consulting ad agency communications firms that they're trying to build their own AI tools. great. Take the API and leverage it and use it in transformative ways, but we're going to need to do something different than how you're doing it. And we need to make sure it's value added for the customer.

Leonard Murphy: Yeah, that I was actually going to ask the So, I've been thinking about this a lot recently, past few days. So, I'm going to, let's see what you think about this. I call it the browser wars in my head, right? It's who is the central agentic operating system both for consumers for brands and I think brands is obviously more interesting and there's this vision of I'm a research manager but fundamentally today my job is managing projects so it's a huge chunk when it should just be getting to the business answer and helping to guide the business and executing on that. Okay, we're moving into that world now where that does become the primary aspect of my job and I'm accessing the tools to do that via Azure, right? Because I'm already using Office and Windows and that is a series of API integrations through platforms like yours and some cases I need to get new information but off first place is going to be I'm pulling it from morning consult and others to answer those questions and build off of and that is such a fundamental different way now so here's the question I don't think in that world the operating system world that buyers are I think they're going to get to a

Michael Ramlet: Okay. Yeah.

Leonard Murphy: Point where they don't even recognize that they're getting the data from morning consult. Okay. U because it's baked directly into their annual license with Azure or OpenAI or whoever. it's just kind of in the cake or maybe there's an intermediary like a Salesforce or kind of large infrastructure players that still play a role but fundamentally they're just transaction processors in an ecosystem. I don't know if that takes us two weeks at this point to get there, but what do you see that look like in terms of the engagement process the buying process of creating revenue for morning console of how that's going to shift.

Michael Ramlet: Yeah. Yeah. So, I think maybe I'd sort of in every one of these paradigm shifts, it's like an curve and you see some very early winners and then you see a different evolution of winners beyond that.  and what at least took place I think cloud is probably the right analogy where you're absolutely seeing value creation for the hyperscalers so AWS Azure Google cloud and those pieces but the reality is that I think in a future agent world there's going to be the broadbased use case kind of to your point there's a certain level of scale that you're going to want to be able to access data pieces within.

Leonard Murphy: Yeah.

Michael Ramlet: Think kind of the way you're alluding to this the ERPs, so Salesforce, others where you're going to want to have data in that space, but they're not going to build out the vertical specific use cases. They're going to, I think, facilitate the build out of the application layer. So you saw this with cloud, those layers had a lot of value at the foundational level, but the largest at least market cap valuation creation was at the application layer.  And so I think from that standpoint, this is where we're focused on building agents. So the chat functionality for us was a really great way to create a discovery tool to begin the sort of evolution of thinking about what's possible with AI and a highly structured data set like morning consults. What we're excited about is you can now build functionspecific agents. So I was giving the example before of Home Depot and the earnings call. For every publicly traded company, they begin preparing for an earnings call 8, nine, 10 weeks out before the earnings call. why not design an agent that's a somewhat similar process? It's different for every category, and what the analyst community on Wall Street is focused on is different to reflect that, but we know generally speaking, what are the core elements of understanding a consumer customer base? why not have an agent that generates that report uniquely for that business and have it do so on the schedule of which they prepare for that earnings very specific use case. Similarly with brand reputation right now you've got these legacy brand reputation trackers. you might see them do a quarterly brand reputation study. They're spending three $400,000 on this program sometimes more. Gosh you better pick the exact right dates in that quarter.

Leonard Murphy: Yeah. Yeah.

Michael Ramlet: Think about everything that just happened in the last three months across the trade space across geopolitical operating environments. You get a hot war in India and Pakistan. all of these things are pretty dynamic events. So the lack of real-time data but why not have an agent that's designed for your brand reputation analysis that's real time that could alert you when you're going through.  I did a nice job of knocking a coffee cup down my justicipulation but can alert you in the moment it happened and that's the real time aspect of it's speeding up. It's not like I'm waiting for the report at the end of the quarter or three to get it. I need the alert right now. hey, there's a problem going on in Mexico for an American brand and it relates to the tariff and Americanism like issue. they need to know now so they can address it whether it's through advertising or other public affairs, but waiting for the next quarter isn't fast enough. And that's where I think it's going to be very use case specific.

Leonard Murphy: Yeah. Yeah.

Michael Ramlet: I agree with you you're going to have the hyperscalers. They'll be big winners, no doubt, but they're not going to be able to design every single use case and every vertical for every function.  Okay.

Leonard Murphy: So, as of today, I am a big fan of perplexity. because I just trust the profound research for data. There's no hallucinations. it gives me what I want and need in a very utilitarian way, And then Grock 4 from more of a production creativity perspective, right? Those are my to tools right now. Grock is interesting and I'll get here in a second. set aside all the other stuff around Elon, that's irrelevant. The idea of an AI that's trained on real-time news and sentiment, via X and other sources. That's really interesting. And now integrating to poly market for that prediction market component. I think that's a really interesting piece but I thought often I even made an attempt to get in touch with Elon a while back to say you need survey data in here but he didn't acknowledge me. So in that world of a value creation for some of these systems XAI I think they're building a very specific approach to the type of data that they want and how that is used overall.

Michael Ramlet: Yeah. Yeah.

Leonard Murphy: Do you see an opportunity to engage directly with the LLMs to say it's great, yeah, you're sucking up all this other stuff, but we have a very specific data set that is highly predictive on things that actually people pay for, brands, this addresses business issues. What do you think? Have you had those conversations on where you fit the big

Michael Ramlet: I think you're bringing up what essentially is going to be some of the quintessential questions in the next three years for enterprise AI. So, one thing I like to distinguish here is most of the development has been with large language sets. So they are consuming text…

Leonard Murphy: Yes. Sure.

Michael Ramlet: Video, audio. Pretty much 9% of the world's language has been consumed by the LLMs. And it is different to design those systems around language than what it is to design those systems around numerical data sets. So I think one of the things that sometimes lost is people think, I'm just gonna throw my data at the AI. It doesn't work that way. And Right?

Leonard Murphy: No, it sucks at it actually.

Michael Ramlet: Math they're known as large language models, not like large math models,…

Leonard Murphy: Right. Yeah.

Michael Ramlet: Even though there's obviously incredible amount of math behind them. The issue that I think is exciting for the industry and for anyone that's collecting structured high value longitudinal data is that for most enterprise AI applications, they're going to have to go beyond we use chat GBT wrapper on our qual interviews. that's already commoditized.

Leonard Murphy: Yep. I know.

Michael Ramlet: It's kind of been amazing to watch how many VCs have just sank money into the same exact solution.

Leonard Murphy: And there's people spinning up. I actually spoke to a CEO this week who's like, "Yeah, we're going to give that away." I spun it up. I live coded it this weekend when we have it.

Michael Ramlet: 100%. Yeah, it's Table stakes. Yeah. And essentially the compute power is going to zero or…

Leonard Murphy: Yeah. Yeah. Yeah.

Michael Ramlet: The compute cost is going to zero. What's going to matter is what can't you do? And what you can't do is go back in time and collect that data. And I don't think there's probably enough appreciation in the industry is we're not talking about banner tables. We're talking about we have pabytes of respondent level data that is highly structured and that data infrastructure I think has been an Achilles heel for the industry it's too often we generate a quick topline cross tab for the project all this project based work versus thinking about it in terms of a long-term infrastructure and…

Leonard Murphy: Yep.

Michael Ramlet: I think for us we're really excited about what will become the enterprise AI opportunities so there whether they're with the ERP PS like Sal for Salesforce or Adobe the world and informing those whether they're with the management consultancies and the advertising professional services firms or if they're with platforms like Bloomberg how do you ingest so by the time this will have aired we are already Bloomberg's already ingesting the consumer confidence data and…

Leonard Murphy: I agree.

Michael Ramlet: Other economic data directly into the platform and so you're going to see I think it's an ecosystem but there will only be a greater premium placed proprietary data and what becomes I think the opportunity for the brands is to think differently about collecting data that becomes valuable to the organization. So the other idea I think by the time this airs that we'll be public with so it can be a little bit more cavalier here is unlimited questions on demand.  Basically, we've driven down the unit cost of collecting an interview to a place in which we can open up the concept that you shouldn't be limited by time or cost to collect more data that's relevant to the decisions you want to make. We're going every day regardless. How do we extend that capability to agency partners to brand clients to just ask questions? And that might mean, hey, I want to come up with some tracking data sets. That might mean I want to have some thematic elements, but it's no longer, hey, I want to do one project. I only have one shot at it. It's this emphasis on continuous data collection. And I think that's going to change how hopefully the end brand client thinks about the data set they're creating, how it plugs into AI applications in their organization.

Leonard Murphy: I Yeah.  It's no wonder you don't get invited to the holiday parties. the right right and…

Michael Ramlet: It should be better. It should be faster and it should be cheaper. It is in an mantra

Leonard Murphy: And yeah so the industry was built fundamentally on service and then we engineer these processes that were also revenue generating attached to enabling the service. when it was always just supposed to be here's a question and here's the answer. that's all the buyer cares about. Everything else we built these edififices of academic purity and all that stuff is important but irrelevant now. Yeah.

Michael Ramlet: And in isolation irrelevant does it actually connect to the business outcome. So that sort of ability to predict consumer behavior and business performance enhances the role of insights in an organization if it actually is elevated to the CFO to the business leaders in the way that today I think too often insights of the last two decades allowed itself to get siloed there are data analytics teams that don't identify with insights that are doing insights work and…

Leonard Murphy: Right. Yeah. Yeah. Right.

Michael Ramlet: That should be a change in the industry. We should see insights empowered with the right perspective on…

Leonard Murphy: Yep. Yeah. Yeah.

Michael Ramlet: How I can support business decision making across multiple functions.

Leonard Murphy: There's a piece that I'm going to be publishing in the next day or so. It'll be old news by the time this goes live, but where I imagine this of the day in the life of an agentic research manager basically. and' that vision of look we thought automation etc etc would get us to this point where we could be consultants on the brand side right if I'm a resource manager but we just did more managing of the process. Now we are at that point where yes we can reach that point…

Michael Ramlet: Yes. Yeah.

Leonard Murphy: Where we are consultants in the organization to help deliver business value through the answer to the question not through managing the process. Yes.

Michael Ramlet: How do we make sure that the best human insights leaders are doing what humans can do better than what the automation can do today? it doesn't devalue the importance of a novel creative question. It actually enhances the value of creative thinking and…

Leonard Murphy: Right. Yes.

Michael Ramlet: The ability to ask really sometimes difficult or novel questions. And that is going the pursuit of truth is even more valuable. But you actually have to be committed to it's not about the process. It's actually about the pursuit of the answer.

Leonard Murphy: Yes, agreed one again by line agreement. And I think that there's certainly a world where we get to where the data becomes a commodity itself as well because it is ubiquitous and available. But the ability to engage with a human to fill in the gaps of information it's the unknowns that mechanistically systemically is…

Michael Ramlet: Yes, right.

Leonard Murphy: What will still drive value for any organization at least in my mind you have built a system to engage with humans and to get information from them. Now the types of information We may see more behavioral, whatever that certainly is reasonable and we'll see that happen. But it's still not going to get to the why. And I think increasingly that we will now be enabled to focus on the why. this is what they're reacting to this new ad campaign, but what is why? And to your point, how predictive is that of future behavior? Because that's the killer app, right, that the brand is really interested in.

Michael Ramlet: And recognizing though that consumer behavior is going to change,…

Leonard Murphy: Yes. Yes.

Michael Ramlet: Minds change. And I think this is going to be the next level challenge for just leaders generally in a business setting is that the old stock market adage of prior performance is not a predictor of future success.

Leonard Murphy: Right. Yes.

Michael Ramlet: That's where I think synthetic data falls down in particular is outlooks change. If you're a leader and you have conviction and idea, nobody's, the old adage around like Steve Jobs and the iPhone, like nobody's asking for an iPhone because they can't imagine a world in which there's an iPhone.

Leonard Murphy: Right. Absolutely. Yes.

Michael Ramlet: You actually, I think, have this opportunity now to know exactly what's going on in the current minds of the consumer, but it still necessitates you to have conviction and idea about what's possible in the future. And it allows you to track your performance at educating or moving that future. So it's more real time. It allows you to refine it and see what's breaking through. So it improves the decision-m. But you still have to have vision. It doesn't replace that.

Leonard Murphy: Yeah. Yeah. Agreed 100%. I mean it seems like we are living in paradigm shattering times in every sense of the word,…

Michael Ramlet: Yes. Yes.

Leonard Murphy: Right? The, things that we thought Things that we thought were not true, we realize are. There's, behavior changes we didn't expect. What? Kids are going back to flip phones. what I mean there just so many different to your point unexpected and maybe there's an algorithm that could have predicted it but I sure as hell didn't and I haven't seen these unexpected changes in reactions in behavior due to just the world that we live in and yeah anyway I was not going to old fogy territory back in my

Michael Ramlet: I think what I mean though is like…

Leonard Murphy: Yes. Yep.

Michael Ramlet: if you're coming into the industry now and starting your career, these are incredibly exciting times because you have this break in a paradigm shift. it's an opportunity to think differently about the world. And I think you're going to need executives that are comfortable with the idea that you might not know the answer today, but you probably have pretty strong conviction that the way we operate today isn't going to be the way we operate tomorrow. And being willing to empower and give the autonomy to individuals to operate. And I think, That's individuals coming into the industry early in their career and really exposing themselves to experience perspective and experience perspective embracing that change and doing so at a rate of change. the industry has not been known for. one of the greatest things we learned over the last three years was that,…

Leonard Murphy: Yep. Go.

Michael Ramlet: We've brought in some incredible talent from some of the legacy firms and they've totally transformed what we're capable of doing. But the reality is it's not a nineto-five sleepy industry anymore. it can't be and…

Leonard Murphy: No. Yep. Yes.

Michael Ramlet: It will fail to transform in the direction needs to go.  And we've gotten to the point where we need to acknowledge the intensity at which we feel we need to operate to achieve our business goals, but also frankly to just achieve our clients goals.  And that's just a paradigm shift that I think the industry is going to probably have a lot of nashing of teeth over.

Leonard Murphy: But the good news is so I was talking to another platform play right I mean purely technology focused platform and there's a concept out there of services software and that model and I think I've even shared that with you in one of our email exchanges but point is it's interesting that they are saying look we are really going into the era of aqua hires for all intents purposes because we need smart people who do provide some layer of service of consultation around this even though that's not really how we make our money but it is how we create strong client engagement overall so I think we're seeing a path where even some of the existing folks in the industry rather than saying my god what it's all changing no they're the core skill sets of being smart critical thinking

Leonard Murphy: Knowing how to ask questions, what are the right questions to ask? To your point earlier, whether it's intuition or experience or that mismatch of those things to the sense making component of information to help guide decisions that aren't just, something that we could easily display, right? That there's a strategic element that we need to think through. that gives me hope for the industry to transform rather than be solely disrupted.

Michael Ramlet: Yeah, I think I'm for all of sort of identifying the changes and undercurrents, I'm long the insights industry. I think that it's going to have an entirely different look to it, but the importance of informing those decisions has never become less valuable. The importance of knowing what people think has never become less valuable. It's just that it's not going to be about process and sort of back office management. It's going to be about how do I, creatively design a question or an approach that answers the question or gets the answer you're looking for how do I communicate that? So the human aspect of that and…

Leonard Murphy: Yep. Mhm.

Michael Ramlet: The trust building and credibility building of that is going to be more valuable with every passing day. the issue we talk about it framing it this way. If you go to morning consult.ai right now you still have to ask it a question that's still the starting point. Now…

Leonard Murphy: Right. Yep. Absolutely.

Michael Ramlet: Granted that's why we're building agents and function specific so it removes sort of the blank screen but if you can go and still whether it's writing whether it's asking those questions like it's still deeply powerful to be able to ask a novel question.

Leonard Murphy: And anybody who's experimenting even with just the basic LLMs that exist today should be seeing that, right? I found myself now that it's a very iterative process going back and…

Michael Ramlet: Yes. Yes.

Leonard Murphy: Forth with my two favorites. because and my questions are getting really long and detailed. So,…

Michael Ramlet: Yeah. the prompt engineering.

Leonard Murphy: Yes.

Michael Ramlet: I think that's one thing that folks don't realize is on a prompt engineering side it is multiple layers and it's the way in which you're design it's almost more like you're developing software or thinking about the construct of software development but it's the critical thinking and structured thinking that leads to better outcomes and some of that will change one of the challenges we really push on our team to think about is don't think about where the pick your LMware is today. Think about a future ideal end state and assume that somewhere between now and some uncertain period of time they're going to unlock the ability to get to that future end state.

Leonard Murphy: So I want to be conscious of time…

Michael Ramlet: So design for the end state recognizing that somewhere six months from now, six years it's going to be possible to do things that we didn't think were possible six days ago, six months ago, six years ago.

Leonard Murphy: Because you and I have proven now in our second conversation and…

Michael Ramlet: I did.

Leonard Murphy: We can talk for a long time. but it's something I think about an awful lot is in the kind of like the browser of the LLM wars as a week or so ago Grock was supposed to be the fastest. Next week GPT5 is supposed to come out, And it's supposed to be the smartest. And so, as you're building a business with these things baked in,…

Michael Ramlet: What? Yeah.

Leonard Murphy: How do you spread your bets so that you don't get caught into the Betamax, version of things? so that you're okay.

Michael Ramlet: So it starts with being LLM agnostic. So for us we use every one of the models you mentioned before em we use Claude. different models are better for different tasks and obviously you can't control for some of the events that might happen that change the availability of a model.  And so we've been LM agnostic from an architecture standpoint from the beginning. our external platform is morning console.ai. We have a separate internal platform. You'll appreciate this code name Flynn from Tron. So all of the internal GPTs that are built by our staff, whether it's improving questionnaire writing, whether it's programming questionnaires, whether it's fraud analysis and analytics. All of those things again are model agnostic but we're developing both for external facing applications as well as internally. And so for us to be able to do unlimited questions on demand the reality is that we built all the infrastructure to be able to go and collect the data. We know how much it costs to collect a respondent in one of those 43 countries at any moment in time on that day. And so how do you then build out the applications around that?  And you see a future in which every one of those applications then goes through an orchestration of working together and that's where I think you end up with larger full-scale agents today they're GPTs whether some of them may be MCPs but I think that it's sort of recognizing it's about architecture and it means making investment decisions today that scale so infrastructure matters I don't think the industry did a good job in the last two decades of investing in infrastructure and the reason why I know is we couldn't possibly have launched morning console $30,000 in angel capital and reach hundreds of millions of dollars in revenue if there wasn't a lot of inefficiency. We were able to do that because of really great people and we're able to do that with really great systems and tech. But the reality is the industry can be much more efficient than it currently is.

Leonard Murphy: Yeah. And it's going to be forced that way. So, it from that standpoint without sound like doom and gloomers, I don't think that's the case.

Michael Ramlet: I think it's exciting. It's Yes.

Leonard Murphy: It is exciting. it's existential in a good way. It's like, all right, we're going to get through this and we're going to shift and change. And some people may succeed more, but we all can. Yep.

Michael Ramlet: it's a mindset and I think that's the thing is I think we're being much more honest about the intensity we think it takes to operate at morning consult and how we have to operate that intensity for clients but it also is the continuous learning this commitment to it is not going to be static you cannot stand in the way of progress and I think that's going to be a client existential question is for insight  leaders. Do you want to be a pioneer in progress or do you want to be run over, in this new state? Yes.

Leonard Murphy: I'm not even sure that is I would usually argue against the slowmoving pace due to inertia. But the business imperative simply from a cost-savings standpoint the procurement thing, I know for a fact how much labor cost is wasted for people that are making, $150, $200,000 a year in going through the procurement process to identify a supplier for a specific project, right?

Michael Ramlet: Yes, this is unlimited questions on demand.

Leonard Murphy: And suddenly if we can wipe that out or at least condense it way down the purely business level efficiencies,…

Michael Ramlet: the number of times we sos or…

Leonard Murphy: Right? Yeah.

Michael Ramlet: Totally new procurement processes like what if we did it once and you could ask unlimited questions anytime you want anywhere in the world at a moment's not status like that's the type of paradigm shift we see coming in the future  Yes.

Leonard Murphy: And it's just an annual license, I assume. All right. so we really could go on. I love that you guys have come out of stealth mode. Thank you. you were at IAX North America. You spoke at Europe. we hope you continue to engage in the industry ecosystem at the level that you are now because everybody needs to hear this because you're an example of that change in action, right? So competitors out there, you think, morning console, screw them, they're going to, no, this is inspiration and this is why I wanted to have you on, right? You're an example of a company that was building something differently because you could and now the era has arrived and you're tacking and as needed to adapt to that and every other company in our space can as well, every other leader. So I hope you take this as inspiration your audience on those things because yeah the future can be bright and you're exemplifying that. So thank you.

Michael Ramlet: Thank you. It will be bright. It's going to be a great industry. Thanks, Lenny.

Leonard Murphy: Mike, anything else that you want to bring up?

Michael Ramlet: I think the only thing I'd add, how we organizationally and as an industry engage, we should all be more transparent about ways we can improve.  I think that's the sort of like call to arms is, too much hand ringing behind closed doors and too many press releases about what we're doing. I think the reality is that there's an opportunity to be really transparent about, things like synthetic data and what is actually beneficial and value creating. and I think that also is really exciting for folks who are younger in their careers, especially with academic backgrounds. this is the type and opportunity we have as an industry to take a major step forward. And I think that that's a mindset. And so I appreciate your leadership on the industry and the week in week out exchanges this is an area where we should all be talking more and trading the best ideas and challenging one another to be better.

Leonard Murphy: That's awesome. Thank you. Where can people find you?

Michael Ramlet: They can find me at morning consult.ai. So that's the best place that I would go. As much as I love, the idea of talking to me, I think you should experience her data first. They can follow me at Michael Ramlet on X. And similarly on LinkedIn and then also obviously Morning Consult on X and the morning consult page on LinkedIn.

Leonard Murphy: Great, thank you to our audience for taking time and listening to us pontificate. hopefully you got value out of it. It was a fun conversation for us. Hopefully it will be for everybody else. thank you to our producers to Bridget to the folks behind the scenes who manage all the audio and all of our platforms and to our sponsors. But that's it for this edition of the CEO Series everybody. Take care. Bye-bye.