Precision GTM: The RevOps Lever Stack

03:29:05.700 –> 03:29:15.069
Aaron McReynolds: Yeah. So head to the website, alysio.ai, and then please just message me on Linkedin. I I try and get back to every. DM, yeah. alicio.ai.

03:29:16.420 –> 03:29:23.450
Julia Nimchinski: Thanks again and moving right along. Let’s welcome Jose Romero.

03:29:23.640 –> 03:29:27.080
Julia Nimchinski: maybe of product and platform at X Factor.

03:29:27.200 –> 03:29:32.059
Julia Nimchinski: They’ll be presenting precision. Gtm, their revops lever, stack.

03:29:32.370 –> 03:29:34.160
Julia Nimchinski: Say, how are you doing.

03:29:34.160 –> 03:29:36.479
Jose Romero: Good, good! Thank you for having me here today.

03:29:36.720 –> 03:29:38.010
Julia Nimchinski: Our pleasure.

03:29:39.010 –> 03:29:55.748
Jose Romero: So it’s a pleasure to be here that last session with Aaron was very good, because I think there’s a lot of overlap between what you’re gonna hear me talking about and in his point of view. So I think we agree in a lot of things. But

03:29:56.330 –> 03:30:08.850
Jose Romero: so today we’re going to talk a little bit about moving away from, you know. Gtm, firefighting to strategy to talk a little bit more about a new framework for running Gtm in a very surgical way.

03:30:08.880 –> 03:30:28.110
Jose Romero: using AI. So what you know what Aaron was saying, I had a chart for the Gartner has identified that about 13% of a Cro’s work week is spent firefighting, sorry doing strategic work. 87% firefighting. So

03:30:28.480 –> 03:30:41.500
Jose Romero: Gtm teams today are very reactive by nature. So that that is a big problem. And and it’s it’s something that needs to change in order for for teams to be able to meet and exceed their goals.

03:30:42.150 –> 03:30:43.110
Jose Romero: So

03:30:43.380 –> 03:30:54.279
Jose Romero: what does that mean? More? More precisely right? So Rev. Ops, it’s really a 3 legged stool. Right? It’s sales marketing customer success and their challenges.

03:30:54.620 –> 03:31:12.350
Jose Romero: If we treat things separately as 3 different functions, this stool starts falling apart because it’s 1 interconnected system. So, for example, if we decide to push sales to close faster. It destabilizes the whole system.

03:31:12.370 –> 03:31:29.100
Jose Romero: creates bottlenecks, wasted resources because the data and in the actual system is not disconnected. It’s all interconnected. So it drives inaccurate results, inefficient outcomes. So

03:31:29.620 –> 03:31:55.550
Jose Romero: what are these levers that we talk about? Right? I mean, I think a lot of a lot of us on this call are very familiar with them. I mean, I’ve listed some of them here, but it’s a mix of marketing sales customer success in financial levers from top funnel qualification, velocity, conversion, retention, expansion. So the problem is, you know, they’re all intertwined. So how do you handle that? What ends up happening

03:31:55.890 –> 03:32:02.167
Jose Romero: is that we end up creating what we call a growth guest gap. The

03:32:03.100 –> 03:32:05.170
Jose Romero: the you know.

03:32:05.950 –> 03:32:25.819
Jose Romero: as as I’ve been both building AI and Ml. Systems at companies like Amazon for over the years and also been in go to market for companies like Verisign. I’ve always bumped into this problem. We build AI models, whether we were building AI models to build

03:32:25.820 –> 03:32:38.599
Jose Romero: to running Gtm teams. The the problem is always the same. We’re forced to take multi-million dollar guessworks in with, you know, using brittle spreadsheets. So

03:32:38.960 –> 03:32:49.389
Jose Romero: the question is, you know, how do we do this differently? I think we can and must move from, you know this guesswork to a more engineered approach to growth.

03:32:50.020 –> 03:32:53.629
Jose Romero: So what do we mean by that? Right? So let’s just take an example.

03:32:54.320 –> 03:33:03.040
Jose Romero: So your your you know your team does a heroic job of shortening sales cycle, and we all think that’s a win.

03:33:03.320 –> 03:33:06.510
Jose Romero: But if the marketing lead velocity doesn’t change.

03:33:06.630 –> 03:33:10.019
Jose Romero: then your high performing reps are now idle for a month

03:33:10.160 –> 03:33:18.549
Jose Romero: of the quarter. So that’s what I call a bottleneck in action. Right? So it’s a Gtm system it’s

03:33:19.140 –> 03:33:37.559
Jose Romero: the system is is, you know, is as only as fast as the slowest part. So you speed up sales cycle. Now you find what? Where your next bottleneck is at, and that’s the thing you got to fix next. So how do you? How do you find those bottlenecks? How do you proactively adapt to those things

03:33:37.960 –> 03:33:38.800
Jose Romero: so

03:33:39.070 –> 03:33:54.629
Jose Romero: as a way to help move things along. I’ve put together a Pdf. That I can share with the group. So there, there is a way to start closing this gap today

03:33:55.050 –> 03:33:57.025
Jose Romero: and take a if we

03:33:57.951 –> 03:34:27.170
Jose Romero: take a couple of examples that I’ll share on this. Pdf, you can use a tool like cloud desktop, for example, in a salesforce, Mcp. Server, and then, instead of just chasing your reps or relying on run rates, these 5 prompts that you see here leverage generic llms as a smart calculator in a way to provide far deeper insights into pipeline health, lead, conversion, stall opportunities, engagement, etc.

03:34:27.610 –> 03:34:30.750
Jose Romero: to have a more nuanced projection on revenue impact.

03:34:31.490 –> 03:34:49.799
Jose Romero: So these are practical, immediate things you can do today. And at the end of this session. I’ll share that. Pdf. And you can take a quick look. What I’ll do is I’ll just pivot over really quickly, so you can take a take a quick snap at it. But again Julia has the Pdf. And she’ll share it. So these

03:34:49.800 –> 03:35:14.579
Jose Romero: prompts, for example. This is a pipeline health prompt if you connect. Let’s say you use cloud desktop or any tool that leverages. Mcp. You could connect your your salesforce instance to it. You create an integration user and you’re off to the races, and then you can start asking questions. So in this, in this example, I’m doing an analysis trying to figure out the gap

03:35:14.580 –> 03:35:17.539
Jose Romero: in opportunity creation. So

03:35:17.760 –> 03:35:27.130
Jose Romero: I’ve you know, asked it a few very pointed things like, you know, the last 12 months I want to exclude bottom and top

03:35:27.600 –> 03:35:51.439
Jose Romero: outliers by amount and sales cycle. And I want to predict what new opportunities need to be created between today and the end of the fiscal year in order to hit my target. And I also want to predict the number of opportunities that will be created between now and the fiscal year. So basically, what am I targeting as a opportunity generation

03:35:51.440 –> 03:36:04.830
Jose Romero: right now? Versus how much do I need? Am I going to have a gap? And it’s, you know, this is different from what you can do with salesforce. It’s salesforce will only go so far right? It cannot do forecasting. It cannot do predictive analytics.

03:36:04.880 –> 03:36:09.829
Jose Romero: And so when I ran this particular one.

03:36:10.210 –> 03:36:39.429
Jose Romero: It gave me a full analysis. And you know, we’re using mock data on a salesforce instance. But it it went deep, right? It went ahead and sat and identified that, you know, I give an example of an 8 million dollars target. It’s identified how many opportunities I needed. It identified what my rate of opportunity creation needs to be versus what it is. And I figured out, you know my, what my opportunity creation gap is.

03:36:39.610 –> 03:36:51.839
Jose Romero: So this is what I mean about a growth guess gap. So this is using mathematics to come up with a mechanism to identify where your bottlenecks are at.

03:36:52.030 –> 03:37:07.730
Jose Romero: So with this, there’s like, I provided 5 different examples, one for lead, conversion, bottleneck, identification, and many others be, you know, I’ll share this with a group, and if anybody, of course, has any questions on the document, be more than happy to answer them.

03:37:08.879 –> 03:37:16.870
Jose Romero: But so this is great. I mean, this is fantastic. It moves you away from, you know the the crystal ball and the prayers, and

03:37:17.480 –> 03:37:35.709
Jose Romero: in the end of endless scrolling of of spreadsheets. And you know, megabyte size spreadsheets that, you know are break down after certain usage into a more modern, faster way of understanding not only what’s happened, but what will happen.

03:37:36.700 –> 03:38:01.939
Jose Romero: So you know. Growth. AI, I think, is certainly a better way than using spreadsheets. But then there’s still some challenges with using off the shelf. AI, right? So it’s a huge step forward from spreadsheets. But there are some limitations that you can see right. The systems aren’t talking to each other. All of this data is not necessarily speaking to each other. If it’s not sitting in the same place.

03:38:02.140 –> 03:38:23.910
Jose Romero: you still have to provide some context, and you know, in the form of averages and other things that it may not have available to it immediately. And most importantly, these tools cannot simulate what we call second order effects on your decisions. They can’t answer questions like, what happens to my sales capacity. If this marketing campaign succeeds

03:38:24.100 –> 03:38:29.359
Jose Romero: right for that, we need a new class of AI something that is beyond

03:38:29.670 –> 03:38:55.239
Jose Romero: what a standard off the shelf AI can do. So, you know, here’s where your AI becomes much more. It’s not just any. AI. We’re talking about. Not a generic Chatbot. We’re talking about an AI that is purpose built for the revenue engine. So it’s trained on specific dynamics for sales, capacity, pipeline velocity, marketing campaigns, etc, etc.

03:38:55.460 –> 03:39:09.110
Jose Romero: This is what we call a new emerging category. We’re calling growth AI, and this is designed to overcome all these limitations that generic ais don’t have by understanding the causal relationship in your Gtm system.

03:39:09.570 –> 03:39:20.390
Jose Romero: So what do I mean by causal relationship? Right? The 1st level of of AI is predictive. AI. It’s the 1st level of intelligence is powerful

03:39:20.680 –> 03:39:30.479
Jose Romero: in the the Pdf. That I’ll share has leverages that in many of the examples that I share.

03:39:30.710 –> 03:39:34.700
Jose Romero: It analyzes your historical data and forecasts what will happen next?

03:39:35.290 –> 03:40:05.189
Jose Romero: It can give you more accurate sales, forecast or identify customers at risk of churning, but it has a critical limitation. It operates in correlation, and a speaker earlier talked about correlation versus causation. So I’m going to hit that same topic. Now, a predictive AI might tell you that a rep who say flies to meet clients has higher win rates, but it cannot tell you if the travel is, what caused the win

03:40:05.320 –> 03:40:16.780
Jose Romero: right is is that’s the difference between correlation versus causation. So what caused the the the improved winds? Is it the fly, the flight? Or is it something else?

03:40:17.460 –> 03:40:25.320
Jose Romero: Versus causal? You know this is where the the big changes. It goes deeper to understanding cost and effect.

03:40:25.760 –> 03:40:30.049
Jose Romero: So you think about the question, if I pull this lever?

03:40:32.010 –> 03:40:40.500
Jose Romero: what is true and causal impact of to the entire system, what will actually, by moving this lever. What is the trickling effect?

03:40:41.009 –> 03:40:44.300
Jose Romero: So this is how we move from forecasting systems to simulation.

03:40:44.530 –> 03:41:04.199
Jose Romero: So we can model complex trade-offs like, you know, hiring more reps versus increasing campaign, spend and see the probable outcomes before we commit a single dollar. This is how we find out which lever truly moves the number, because at the end of the day the business succeeds. If they can hit the revenue goals.

03:41:05.490 –> 03:41:31.060
Jose Romero: So you know, let’s go a little bit deeper into simulation. And what I mean by simulation. So let’s take a real world example in a complex scenario. So you have a Cro. Say, that’s you know, worried about hitting an annual number, right? And ask you, as a revops person, to model the impact of 4 different improvements across the entire. Go to market engine.

03:41:31.821 –> 03:41:38.920
Jose Romero: In in this example, you know the the they’re asking to modify attrition.

03:41:39.230 –> 03:41:52.320
Jose Romero: retention, win rate, and Mql. To SQL. Conversions so and go figure out. If that’s the right combination. If you try to do that on a spreadsheet, it’s pretty pretty much impossible.

03:41:53.221 –> 03:41:56.470
Jose Romero: The interconnected dependencies are too complex.

03:41:56.980 –> 03:42:24.870
Jose Romero: But if you, if you use an AI system that is built purpose built for for revops, you will be able to achieve that. So you’re not guessing anymore. You’re simulating. You adjust these levers, and the system instantly models the compounding effects across the whole revenue engine. It tells you not just the potential revenue lift, but also that you’ll close the gap in 9 months, that you will need

03:42:25.200 –> 03:42:39.750
Jose Romero: an 8% increase in pipeline to support the higher win rate. It’ll factor all of these things things for you. So this is the big difference between a panic guess. For because the Cr asked you to do this to a precise data-driven approach.

03:42:40.340 –> 03:43:04.199
Jose Romero: So you know, what does this look like in real world? I’ll just pivot to an example. Here we’re looking at a capacity performance dashboard, where you can simulate. What could you could do? You know what would happen if you say increased your average sales price, or you decreased your attrition, and you. When you move these levers in real time, your your outputs change.

03:43:04.460 –> 03:43:05.360
Jose Romero: You’re

03:43:05.540 –> 03:43:28.200
Jose Romero: you’re seeing that that by modifying your your new business or increasing your new business average sales price by a percentage. It. It changes the whole system. You modify your let’s say your your America’s sales attrition. You reduce it. It also modifies the whole system.

03:43:28.270 –> 03:43:40.380
Jose Romero: You decrease your RAM time. It modifies your whole system. So it’s a it’s a way to, you know. In one place be able to model all these different possible scenarios.

03:43:41.850 –> 03:43:42.630
Jose Romero: So

03:43:44.380 –> 03:43:53.430
Jose Romero: But we go. We need to go further, because there’s so many levers that kind of more than 30 30 levers. What is the right combination?

03:43:54.043 –> 03:44:05.999
Jose Romero: What are? What are the right levers to move? In which way? Right? There isn’t a single answer. To be honest right. The it all depends on what the business can achieve.

03:44:06.040 –> 03:44:30.110
Jose Romero: given its resources in its, in its scenarios, and where it’s at right now in their history. So if we, if we then take this approach and build a multi-agent system across it, it can actually decipher a set of potential scenarios and provide them to the user. So this is ultimately, this isn’t a solution to replace revops.

03:44:30.640 –> 03:44:45.679
Jose Romero: It’s a solution to elevate revops and even amplify it. So you’re moving away from the mundane manual analysis to something more automated. So it frees revops teams to focus on high value, strategic work

03:44:45.760 –> 03:45:01.839
Jose Romero: by designing, say, better playbooks, handling complex exceptions, even fine tuning, some of the AI responses. So your your job as a revops person becomes now you become the chief architect of the Company’s entire revenue generation engine

03:45:03.060 –> 03:45:26.419
Jose Romero: something to another interesting point to share. This is a slide that we put together with the revenue operations associates. They are the company who put together the Revops book and are offering the revenue operations certification, and they are in this chart showing the top reasons why Gtm teams are failing is because revops are not focused enough

03:45:26.580 –> 03:45:32.939
Jose Romero: and don’t have the right tools to drive these top growth drivers that are listed in here. So

03:45:33.280 –> 03:45:52.019
Jose Romero: with a growth AI system. You can now spend more time on strategy and less on firefighting. So these are the things that that could be more impactful if you if you had if you could move away from firefighting into more strategic planning and becoming that that chief architect or chief architect of revenue?

03:45:53.652 –> 03:46:03.989
Jose Romero: So you know, you’ve you’ve heard through today a lot about agentic AI and systems that execute tasks. And you know that’s the future, and it’s powerful.

03:46:04.200 –> 03:46:10.889
Jose Romero: But agentic execution is only as good as the strategy it’s executing. So the focus of growth. AI

03:46:11.110 –> 03:46:15.879
Jose Romero: is to go one level above it’s empowering you, engineer. The strategy itself.

03:46:16.180 –> 03:46:25.900
Jose Romero: though other systems are going to focus on automating the how we use causal AI to help you master the why and simulating what ifs?

03:46:26.380 –> 03:46:39.189
Jose Romero: So? We believe the greatest leverage. Leverage comes from not just doing things faster, but from deciding to do the right things in the 1st place. So instead of guessing, just get it right the 1st time.

03:46:41.699 –> 03:46:56.889
Jose Romero: So you, you know, you can simulate. Go to market trade-offs. You can close this growth guest gap and replace your spreadsheets with an always on strategy, you know, moving beyond just automation, automating tasks. Truly engineering your growth.

03:46:57.310 –> 03:47:04.940
Jose Romero: So that’s you know what precision Gtm is all about? Is this confident, coordinated execution, using causal, not correlation.

03:47:06.280 –> 03:47:12.090
Jose Romero: And with that I will stop and answer any questions.

03:47:15.680 –> 03:47:22.210
Julia Nimchinski: Phenomenal session. Let’s get into the question. So one of the questions.

03:47:23.106 –> 03:47:30.540
Julia Nimchinski: that keeps coming up is most of our Gdm trade-offs still happen in spreadsheets and back channel threats.

03:47:30.850 –> 03:47:34.420
Julia Nimchinski: What actually changes when you move to AI simulations.

03:47:35.310 –> 03:47:35.920
Jose Romero: So.

03:47:37.557 –> 03:47:39.680
Jose Romero: The the spreadsheets.

03:47:39.680 –> 03:48:08.919
Jose Romero: you know. So there’s there’s many data sources where this information is kept. It’s it’s kept to your point of spreadsheets, on, maybe on, on conversations, on chats, on emails, on, you know, even calls that are transcribed with gong so the the main difference here is that in order to for you to be able to appropriately simulate right? Ai systems are only as smart as the data that they have access to so

03:48:08.920 –> 03:48:28.970
Jose Romero: feeding it. All the right data in order for it. To be able to simulate things properly is the only approach, but more precisely on on the question here. Right? I mean, there’s 30, some variables in go to market. So the the key thing that we’re doing here is trying to simulate the

03:48:29.000 –> 03:48:33.479
Jose Romero: correlation and then causation of these variables. So

03:48:33.600 –> 03:48:55.949
Jose Romero: the the mathematics basically behind and the science behind growth, and how how moving one lever impacts the other lever. So it’s a little bit less about where data is stored, and more about how to apply mathematics to the data in order to come up with reliable outcomes.

03:48:58.120 –> 03:49:06.939
Julia Nimchinski: They? People are asking, what’s the most surprising insight? A team uncovered running these simulations in your practice.

03:49:06.940 –> 03:49:34.369
Jose Romero: Yes, so we we had one customer in essence identify that the board number that was being provided was just not achievable with the resources they had. So they they had a number, a number a number over their head, and no matter what adjustments they could make there was no way for them to actually hit that number. So

03:49:35.061 –> 03:50:01.759
Jose Romero: it it it uncovered in essence. Gave you gave that team the ammunition to go back to the board and say, based on what we have today. We can’t achieve this, and may may not be able to achieve it in some time, because things like sales rep ramp time are are a factor that you cannot really manipulate. Right? So if it takes time in this case that company. Their ramp time was 9 months.

03:50:01.850 –> 03:50:10.450
Jose Romero: Even if they start hiring people that day, they would not see an impact to that number until the the following year.

03:50:12.700 –> 03:50:19.079
Julia Nimchinski: Another question. Here rev-ups often find itself as a cleanup crew.

03:50:19.250 –> 03:50:23.980
Julia Nimchinski: How does this help move the function upstream? Closer to decision making.

03:50:24.790 –> 03:50:25.800
Jose Romero: So

03:50:26.990 –> 03:50:37.149
Jose Romero: And and you know, I’m gonna elaborate on the cleanup crew. So things don’t go well, they are told. Go look into why they did not go well and fix it.

03:50:37.460 –> 03:50:45.180
Jose Romero: So it goes back to the whole firefighting reactive approach that I was talking about earlier. The

03:50:45.490 –> 03:50:47.199
Jose Romero: the this approach

03:50:47.380 –> 03:51:09.560
Jose Romero: changes the paradigm completely, because it does 2 things for you in real time. It’s telling you, what are these metrics look like right now? So you don’t have to do a lot of spreadsheet work to figure that out. And not only that, it’s also looking into the future and tells you, and it’s telling you already what’s going to happen before it happens. So

03:51:09.720 –> 03:51:12.537
Jose Romero: if you think about it from the standpoint of

03:51:13.030 –> 03:51:37.129
Jose Romero: today, you’re basically blind to to what’s going to happen. And and then your door gets knocked by the Cro saying, we have a problem. You don’t even know what the problem is. They’re explaining you the problem. You actually have to figure out what exactly happened, because they’re going to explain to you something at a very high level. You got to go deep. Now figure it out, and then come up with a proactive response.

03:51:37.160 –> 03:51:48.469
Jose Romero: This, on the other hand, flips the script. You actually know, even before the Cro. What is going on? You’re being proactive. Now you’re looking at things and going.

03:51:48.680 –> 03:52:08.420
Jose Romero: We’re going to have a problem here and here and here. So you go to the Cro and you tell a Cro we’re going to have these problems. What can we do about them? Cro Net will never get us prize, because they already know from the system what the next problem is. So whenever they’re having a C-suite conversation that Cro is never on the hot seat.

03:52:10.740 –> 03:52:17.070
Julia Nimchinski: Question from Diana, how ready are your typical clients with data to feed this system.

03:52:17.730 –> 03:52:28.317
Jose Romero: So yeah, that’s a typical problem that we hear from from our customers. Everybody has a data clean cleanliness problem. We do have

03:52:29.950 –> 03:52:47.291
Jose Romero: systems behind the data that are actively looking at it and trying to clean it up. For example, our territory planning component we so cut their companies that leverage geographic territory assignments and their

03:52:48.050 –> 03:53:14.169
Jose Romero: the accounts may not have all the correct information in some cases that doesn’t have any information on the geography, or where they, that account sits. So we go and clean that up. We make correct addresses, or we may actually fill addresses. So that’s just one example. But across the board we were looking at the customer’s data and try to clean up as much as we can.

03:53:14.170 –> 03:53:21.806
Jose Romero: There’s still, things that you know can’t be done by us. They would have to be done by the by the customer. But ultimately,

03:53:22.552 –> 03:53:27.089
Jose Romero: you know the the. It’s a collaborative approach to, to, you know, getting good results.

03:53:29.010 –> 03:53:36.309
Julia Nimchinski: Another question here, how do you simulate headcount shifts without overfitting to historical performance.

03:53:37.510 –> 03:54:02.829
Jose Romero: So I think the question is, you’re shifting a resource from one Geo to another. Geo. Or one role to another role, maybe from mid market to enterprise. So we do a couple of things right? So we’re looking at both that person’s performance against their peers. So how’s that person performed? So let’s say hypothetically.

03:54:02.900 –> 03:54:30.634
Jose Romero: just to pick an example. You have a mid market rep that is promoted to be an enterprise. Rep. So they normally get promoted, because, you know, they’ve done better than than their peers. So we we do take that into consideration. But we also calculated an estimated yield for reps across time. So we do. A ramp time, calculation. How long does it take for a rep to to

03:54:31.590 –> 03:54:52.610
Jose Romero: to hit a hit their stride in essence. And then what is that? Stride like? What? What is the estimate yield across different tenure bands? So we’re looking at, you know. Say, 0 to 6 months versus 6 to 12 months, etc. And we do make those calculations. And then, whenever a rep gets transition, we we consider that also in the math.

03:54:55.900 –> 03:54:57.540
Julia Nimchinski: Another question here.

03:54:57.950 –> 03:55:07.700
Julia Nimchinski: Super interesting. Given your background. Say, what? So yeah, what breaks first, st when companies try to build this kind of

03:55:08.702 –> 03:55:14.030
Julia Nimchinski: lever model in house without a purpose, purpose, built platform.

03:55:15.140 –> 03:55:36.640
Jose Romero: The 1st thing that breaks is the models themselves. So there’s 2 things that are going on behind the scenes here. So you have both a machine learning model, which is different from an Llm. I think some of the folks in previous sessions discussed that a little bit before. Think about a machine learning system

03:55:36.640 –> 03:55:46.279
Jose Romero: whenever and you use spotify right spotify is making recommendations to you. That is not an Llm. But it’s still an engine that is looking at

03:55:46.280 –> 03:56:09.249
Jose Romero: what you clicked on, and the things that you did, and trying to give you a recommendation on what you may want made like next. That’s a machine learning model versus a large language model, which is, you know, we all know, Chat, Gpt, and Demini and all these other tools. So we do. A combination of these. So the biggest, the 1st challenge that customers typically have is an accuracy challenge.

03:56:09.260 –> 03:56:13.970
Jose Romero: So they try to clean up their data. They try to come up with.

03:56:14.805 –> 03:56:41.310
Jose Romero: A. You know a scenario where they manipulate these levers and come with an outcome. But they may not have the 1st first, st the full picture of how all these things are intertwined, and the the causal effects of them, and then tying that, together with an with an Llm. That would give you recommendations. So, tying the Ml. And the Llm. Together so that they

03:56:41.510 –> 03:56:51.699
Jose Romero: can work together in concert or orchestrate the the movement of these levers that that is typically where things fall flat.

03:56:53.470 –> 03:56:59.769
Julia Nimchinski: Let’s say there’s a lot of noise and confusion these days about AI hygienic AI, especially

03:57:00.090 –> 03:57:01.980
Julia Nimchinski: being a product leader.

03:57:02.130 –> 03:57:17.880
Julia Nimchinski: What would be your recommendations to all of the executives watching Ptm leaders. Where do they start? What do they read? What do they do? How do they transition from, you know? Well established Ptm. Model of running things to this new

03:57:18.060 –> 03:57:19.810
Julia Nimchinski: type of thing that no one knows.

03:57:19.840 –> 03:57:33.606
Jose Romero: No, I I appreciate the question. I something I think, about all the time. So one of one of the things that I do with our own organization is is, I’m kind of the AI sponsor AI

03:57:34.170 –> 03:57:41.988
Jose Romero: evangelist. And then for specifically with my CEO and and some of the executives. What?

03:57:42.980 –> 03:58:06.510
Jose Romero: One of the things one quote from Sam Altman that I that I really liked, that I keep repeating is, I think, about about a month ago. He said that he analyzed the demographic patterns of people who use chat, gpt, and notice that the prompts being used are drastically different between teenagers.

03:58:06.510 –> 03:58:16.979
Jose Romero: 20 year olds and 30 year olds not he never mentioned 40, 50, 60 year olds. He just stayed with those and said that 30 year olds use chat gpt as a replacement to search.

03:58:17.410 –> 03:58:33.794
Jose Romero: Well, night, 1819 year olds are using Chat Gpt as an operating system. And then he went deeper and saying, explaining what that meant. And so in essence, they are using chat, gpt as a

03:58:34.880 –> 03:58:39.550
Jose Romero: somebody to bounce off life decisions before making them.

03:58:39.620 –> 03:59:09.600
Jose Romero: So they’re thinking about doing something, and they bounce off that idea off that system. So it’s not replacing their decision. It’s not that they’re now having the chat. Make the decision for them the chat they’re just consulting. They think they have a thought that they’re going to go in this direction, and they want to consult with AI on it. I think that’s a powerful thought that everybody should consider is as you’re transitioning into this AI world. You know.

03:59:09.780 –> 03:59:19.530
Jose Romero: We know now that the younger generation uses cell phones differently than the older generation. Same things that happen with AI, and the faster we start adopting

03:59:20.020 –> 03:59:26.579
Jose Romero: using AI like, like like younger people, the the more attuned we’re going to be, the smarter. We’re going to be.

03:59:28.280 –> 03:59:30.820
Julia Nimchinski: What do you think they’re working on with Johnny? Ive.

03:59:32.407 –> 03:59:47.829
Jose Romero: I I’m can only tell you. You know these are just. I’m just putting together the dots. But the fact that apple is no longer working as closely with Openai.

03:59:49.090 –> 03:59:54.500
Jose Romero: Tells me that they are basically want to put together an AI 1st

03:59:55.378 –> 04:00:06.559
Jose Romero: device. Honestly think that, you know, 5, 10 years from now. The way that we interface with phones is or devices just gonna call it a more generic term is different.

04:00:06.800 –> 04:00:20.109
Jose Romero: We still gonna need screens for certain things. But majority of the time, we’re just gonna be talking to the devices. And so instead of you know, booking. I don’t know. Let’s say that you want to book something in your calendar or send an email. You just tell the AI

04:00:20.330 –> 04:00:36.219
Jose Romero: so why why have to open up your phone and start clicking clicky, clicky, clicky when you can just tell it to do what you want it to do. And it’s learning your preferences, anyway. So it’s becoming better at understanding how you think so. It it eventually becomes an extension of you.

04:00:37.210 –> 04:00:41.060
Julia Nimchinski: Last question, I promise, what do you think? Chat gpt, and

04:00:42.110 –> 04:00:46.149
Julia Nimchinski: the alike will turn into a distribution platform.

04:00:46.780 –> 04:00:51.230
Jose Romero: I’m sorry I didn’t catch your whole question, so what do I think? Chat, gpt, and.

04:00:51.490 –> 04:01:01.450
Julia Nimchinski: Yeah. And you know all of the perplexity, Claudes and the like will turn into full blown social media, slash distribution platforms.

04:01:03.560 –> 04:01:08.943
Jose Romero: Not sure if I’m answering your question by saying this, but I think there there was a

04:01:10.180 –> 04:01:39.220
Jose Romero: an idea of having a super intelligence, and I think that idea is becoming slightly harder to achieve as time passes. So the Nirvana moment was, we have. We’re going to create the super intelligence that is going to do everything. And as they’re getting to having these systems be generalizable, being able to say something, it learned here and apply it over here is becoming hard to impossible. So

04:01:39.710 –> 04:02:07.399
Jose Romero: I think what’s going to end up happening is all of these companies are going to pick an area, and they’re going to excel at that area. And, and, you know, drive revenue from being good at that. But this idea of having one company that rules at all, because they have control over everything, at least, for now they’re having where I think on the the research papers that I’ve read, people are having a hard time getting to that point.