How Generative AI is Revolutionizing SaaS

Episode Description

In this insightful episode, host Frederick Vallaeys welcomes Tom O’Malley, a seasoned Silicon Valley entrepreneur and founder of the startups Current and Ascend.

With a rich background in strategy at Oracle and a passion for addressing unmet needs in B2B interactions, O’Malley shares his journey and the innovative solutions his startups offer.

They discuss:

• Startup Challenges and Successes
• Leveraging AI for Business Networking
• Impact of AI on Business Operations
• Future of AI in Business
• Personal Insights and Experiences

Episode Takeaways

Tom O’Malley’s Background and Current Startup:

• Tom O’Malley, a seasoned entrepreneur and strategy lead at Oracle, founded Current to address the gap in B2B innovation, sales, and marketing teams not listening sufficiently to market voices.
• Current operates as an online platform facilitating professional panels and focus groups, used by over a hundred Fortune 500 companies for diverse research and engagement needs.

Challenges of Scaling and Customer Acquisition:

• Despite its success, scaling Current has been challenging due to the project-based nature of the business and the high cost of customer acquisition.
• The growth strategy shifted from traditional sales approaches to leveraging the platform’s users (panelists) as potential customers, transforming the sales model into a more organic, self-sustaining cycle.

Integration of Generative AI:

• The introduction of generative AI into Current reduced the need for human intervention by automating tasks such as campaign writing, moderation, and report generation, significantly lowering operational costs and improving margins.
• This technological pivot demonstrated the potential of AI to streamline operations and opened up avenues for further innovation in customer engagement and service delivery.

Concept of Ascend - A New AI-driven Project:

• Ascend is envisioned as a platform that uses AI to extend human capacity in maintaining and expanding professional networks beyond the natural limit (Dunbar’s number), enhancing personal and professional connections.
• The platform aims to act as an “advocate,” using AI to manage and nurture relationships based on dynamically changing motivations and contexts, providing a more personalized interaction model.

Episode Transcript

Frederick Vallaeys: Hello, everyone, and welcome to another episode of PPC Town Hall. My name is Fred Vallaeys. I am your host. I’m also the CEO and co founder at Optmyzr.com, a PPC management tool. Today, I have a very special guest in studio, Tom O’Malley. Welcome, Tom. Thank you.

Tom O’Malley: Good to be here.

Frederick Vallaeys: Tom, I’ve known you for quite a long time.

You are the founder of a couple of startups, longtime Silicon Valley entrepreneur, and I’m so honored that you came into the studio to tell our viewers a little bit about what works for you and what they might be able to learn.

Tom O’Malley: The honor’s on my side. I’ve known you for like 15 years now, and for you to call me up and say, come on in for this, I’m totally excited to be here.

Thank you.

Frederick Vallaeys: But we’re also working on a new project. A couple of projects together, so you have two startups. One of them is Current, and the other one is Ascend. So, tell us a little bit about Current first. That’s the older one, right?

Tom O’Malley: Yeah, that’s right. So, you know, prior to that, I was a strategy lead at Oracle here in the Valley, and just saw an unmet need that really B2B, call it innovation, product, and sales and marketing teams really don’t listen or empathize enough to the voice of market.

I think Oracle’s a culture back then anyway, was sort of the, the prime example or the epitome of a culture that just sort of speaks and doesn’t listen. So I thought, wouldn’t it be cool if we just made it easier to bring people together, professionals, in, kind of think of it as like online focus groups or workshops, we call them panels, but yeah, it’s a, it’s a platform that’s been alive nine years now and it’s been used over 600 different topics or, you know, panels across more than a hundred Fortune 500 companies.

Frederick Vallaeys: Nice. Well, and a lot of startups died before year two. Yeah. So having made it nine years congratulations on that.

Tom O’Malley: Maybe it should have died in year two. I’m coming to think of it. It’s been tough to scale because, you know, this type of project based work as much as you make it more affordable or faster, easier.

People don’t have more problems, right? So research is sort of a tough place to enter. I’d say a scalable SaaS you know, marketplace model.

Frederick Vallaeys: Yeah.

Tom O’Malley: And you know, the nine years really just reflects just a relentless effort to find the right model to, to make it scale.

Frederick Vallaeys: Right. So there’s obviously running the business, finding the right model, pricing model keeping existing customers engaged and coming back, reducing churn.

Right. Right. But my listeners tend to focus a lot more on the B2B acquisition. And so, acquisition in general, but B2B has always been a problem because you have these long sales cycles. You have these big, heavy CRMs. Yep. And how do you connect the pieces together with something like Google Ads? So, so tell us a little bit maybe about how you found new customers.

And it doesn’t have to be PPC, right? Like whatever worked for you. Right. Yeah,

Tom O’Malley: it, it wasn’t pay PPC. In fact, because it had this nature of, You know, project based, you know, those models work really well if you charge a lot to compensate for the expensive cost of sale. So our theory was we could get cost of sale way down.

And the way we did that was let’s host what we noticed was about 20 percent of the panelists coming in to interact on the panels themselves were potential customers. So our potential customers were already on our platform. And so that was the idea is can we make a smooth conversion model to go from, hey, I’m here to be a panelist.

Wow. I kind of like this. I could use this for my business. And so that became like our whole growth strategy. And so we went from a boiler room type environment, you know, of cold calling in the beginning and getting, you know, starting conversations that way to then more of a flywheel type model, I’d say.

As lean as we got that, and it was pretty slick. We still spent a lot of energy on retaining customers. And then even in the nature of that service that we were delivering over the platform, there was a lot of handholding. So there are gross profit margin on a given project might have been 75%. It wasn’t 99.

You know, it wasn’t pure software.

Frederick Vallaeys: Right, exactly. So there was a software component to it, but it was very heavily service driven to enable it. But at the same time, it does sound like it was product led growth because you’d come onto the product. That’s right. You’d see, oh, this, whether it’s a service or a software, this is really cool.

Yeah. I’m finding new insights. I’m using those insights to maybe go and write a big white paper, maybe solve a big problem that I have within the company. That’s right. And so so now you want to go and do this for your own company.

Tom O’Malley: Yeah, that’s right. Sometimes getting people to imagine how they would use it was like a bigger leap than we expected.

Like, it’s so obvious to us, right? You use it for research, you use it for innovation, for engaging. You know, your target market in product based discussions around features or benefits that they’d like to yield or marketing for thought leadership or even sales enablement. I want to talk to those panelists because we have a high ticket item, you know, sale.

And even if we win one, you know, per panel, it was worth doing the panel.

Frederick Vallaeys: This is what I find interesting, Tom, is that the product you’re selling, it’s not even something where people are sitting around at the office and it’s like, Hey, I need this solution like current. I need to do a virtual panel. We all have problems in business, right?

But we try to solve those problems. We’re not necessarily looking for a bigger mechanism that sits on top that enables us to solve all business problems by bringing in experts. So how would you go out to market and sort of get people to understand that maybe. Hey, let’s take a step back. Let’s look at this software and this service that can help me solve across the board as opposed to like that one you know, marketing issue that we have or finance issue that we have within our organization.

Tom O’Malley: Yeah, I mean, it’s, you almost start preaching a vitamin if you’re too broad, right? Like everyone should be engaging with their target market, right? It’s, you know, grow empathy, right? Empathy is a business muscle we don’t use enough. It all is like, yes, no one disagrees. If we’re going to be attracting people for innovation, let’s ourselves host a panel on AI and blank, and how it’s going to impact innovation programs or venture, you know, programs within corporates.

Get them into a discussion that’s very specific about the problem that they live with, while demonstrating that our product is actually helping them do that. So those meta layers of having them realize that they found their solution before they were even looking for it, was the way that we would gain attention.

To close though, it still required a few conversations, and then running even like a POC, or let’s populate a panel so they can really see it come together. I think we overestimated in this project in general, just like how much hand holding and spoon feeding is necessary. If you’re not doing exactly what you’re saying, which is this is for that, you have this problem, this solves that exact problem.

Right. Yeah our tool is so versatile that it’s almost like, how do you get that across?

Frederick Vallaeys: Right, but it’s really cool because then it sounds like you did basically dogfooding. For internal projects, you brought in the experts and then those experts were probably highly curated from organizations. So you basically probably were doing account based marketing.

Tom O’Malley: Yeah, that’s pretty much it. I mean, the way we find people to go into panels is by using like a semantic crawler, like if people are on the web talking about a certain topic, we can identify then who it is, match it back to a LinkedIn, match it to an email, reach out to them and say, Hey, this might be a great panel for you to be on.

So we had that tool for recruiting. Panelists for our product. But we were then using that same tool to prospect for our own customers.

Frederick Vallaeys: Yeah. Best of both worlds. Now I mean, human connection, right? I think you’re a master networker. I you know, anytime I go to downtown Palo Alto, Tom is out and about talking to people about the solutions and just the industry in general, which is really great.

But, but how do you scale that? And let’s talk maybe a little bit about, How that happens in a Gen AI world, because that’s your new startup, right?

Tom O’Malley: Well, yeah, but to get there, we applied Gen AI to Current, the platform that we were just talking about. And it did make it a pure product play. It did transform and take a lot of the services away.

Now, it wasn’t like for like, it became instead of more white glove, more self AI bot was able to write campaigns like some of our back office did. It was able to put a probability on each applicant of how How probable it is that they’re going to be a good fit for that panel. It would then even write up the scope or discussion guide for the panel.

It then even moderated, which was a huge cost for us. It was like 3000 to have a human moderator, moderate a panel. This eliminated that again, didn’t get us to a hundred percent of what the human was doing. But sometimes 70 percent is good enough and, and then it would even write the reports at the end.

So we finally got to our 99 percent margin in the product. So then there was hope again, like, oh, this could be cool. But I think what that demonstrated for us is that, like we, like our attitude that we came into that project with, is that technology is not here to replace people, it’s here to help people connect better, faster, smarter.

And while we had demonstrated that in current, Right, by getting groups of people and bespoke groups of people together for narrow conversations and deep conversations. I kind of got inspired into the notion that, Well, AI is this whole next level thing. And one thing that I’ve realized and sort of was like the genesis of this notion for the next startup, Ascend, is that people, you and I, we can manage maybe about six really close relationships in our life.

Outside of that, there’s probably 25, a ring of 25 that we would call our friend group. Outside of that, there’s like a 50, you know, group that are a little bit, you know, we know and we have life experience with, And then after that, there’s this other sort of professional networking ring that is, maxes out at about 150.

And this is by this guy Dunbar that did the studies and stuff, so I didn’t come up with this. But it really resonates to me, like there’s, there’s a limitation to what we can do. So then we started thinking about, well, what could we use AI to do to expand that number? How do you 10x the 150? And so that became sort of the beginning of a journey that, you know, It had me talking to AI experts, academics neural network folks that could help think about that problem some more.

Frederick Vallaeys: Interesting. So when you talk about these circles of closeness, I suspected also shifts, right? So as you go to a different stage in life where you work on a different startup, Yeah, who you’re close with probably changes as a result of the realignment of interests and that can be a struggle, I think.

And that’s, I think, where your solution comes in is to say, Hey, it looks like you’re starting to think about bringing generative AI into your SaaS company. Here’s a bunch of people who’ve been doing that for other SaaS companies. It might be really relevant to talk to them, but how do you find that within?

And so LinkedIn, what is it? 5, 000 connections or. Yeah, maybe your spine

Tom O’Malley: is more like two, but yeah,

Frederick Vallaeys: but how do you go and find those right people and how do you shift within that network? And I think that’s where technology again can be super helpful.

Tom O’Malley: Yeah. Yeah. I mean, you know, what we say is to connect is human, like to you and I to sit down, you know, share a problem together.

Or, you know, share some vision for the future together. That’s a very human thing. AI will never replace that. But networking is math. It’s all about measuring probabilities. And right now, there’s a species called AI on the planet that is a probability engine. So how can we, how can we let that, I mean, as you were just saying, like, humans go in and out of those rings all the time.

And the reason why they do is because their motivations change. Right. And what we haven’t had before, we’ve had static views of each other, right? Our LinkedIn profiles, our resumes, but they’re not, they’re not motivations. They’re stats. But can you use stats and data to infer motivation and then connect people and reconnect people when necessary?

as their motivations change through life.

Frederick Vallaeys: Yeah, and I think that’s really relevant to a marketing audience because I have a motivation that hey, maybe You know, my team is growing too big. I need a better okr system And so that’s on my mind, but how do I go and find that right? And again, I would love to work with a company a vendor Where I have a connection where I know the people behind it where I know what they stand for but now that’s a lot of digging.

That’s a I got a lot of other stuff going on. So Yeah If that AI, that new species can help with the probability of at least saying, listen, here’s two or three companies that you have in common and here’s maybe a bunch of people that you can talk to who’ve gone through that same thing and who can advise you.

And I think that’s also interesting because when you talk about your system. It’s not an assistant, right? You call it, what is it? We call it an

Tom O’Malley: advocate, yeah.

Frederick Vallaeys: So, so why, talk about that difference, like why is it an advocate as opposed to an

Tom O’Malley: assistant? Yeah, and, or a digital twin, right? So we, we kind of looked at the space and we thought, okay, first of all, the frontier has to be around personalization because it needs to know me.

And I’m a human being, which means I’m a very complex thing. I have layers from cultural context to personality, skills. Yeah. Achievements, age, like priorities, where I am personally, and then you get into goals and passions. And all those layers are the combination of what makes Tom O’Malley, not one particular layer.

And so to do personalization needs to ingest all of those nodes or all those layers of nodes. And then the other thing you have to then is figure out, well, how are you going to apply that towards the betterment of the subject? So if I’m the subject, do I want a digital twin? Do I want an assistant? And we thought, what’s, the way to think of it is, this person is a, is a we, they’re in my, they’re on my team, right?

And it’s their job, almost like a liaison, to be in front of me and be the filter to the world. So, I’m not going to spend time with somebody that hasn’t passed through my advocate, because my advocate is advocating for my time and mental space. And that’s not just to meet people, That could be to consider a product.

Right. Right, because if we are, if our needs are holistically understood, we could trust the said advocate to filter for us.

Frederick Vallaeys: And that’s where it’s interesting because I think this fundamentally what digital marketing is today is machine learning is calculating all these probabilities and putting the probable right solution in front of you.

Because the ad engines, they make money when you actually click on those ads and have high conversion rates. That makes you want to invest more into those platforms. Yeah. But then there’s also. For as much connection as it makes, there’s a level of impersonality of like, Hey, here’s my thing. Here’s the world of people like, just go and figure out who to show my ad to.

And so Google now goes to the point where it says, give me a bunch of headlines, give me a bunch of call to actions. And then it puts the ads together on the fly based on what it thinks is going to resonate with Tom or with Fred or with Sandra or whoever it is. But then I think what you’re talking about is much more, okay, let’s, let’s Take that to that next human level, because ultimately now we’re talking about conversion rate optimization, landing pages, right?

Like, do we really want people to sit there and have that whole process be impersonal on the web? Or at what point does a human come in and does a human connection take over? Yeah. And I think you’ve proven it in current where. You know, people have talked to each other and that gets the ball rolling. So, I think it’s really fascinating how you’re doing this.

Tom O’Malley: Yeah, I think the biggest picture is, if I was to say the era that we’ve been in, is the push model. Where we push ideas out to people and though we try to use some metrics to get higher probabilities, we’re still spraying and praying. We’re still lumping a lot of people together into assumptions on stats, not motivations.

But if we could turn it into a motivation model. Where it becomes a pull model. So I’m going to pull the web to me via my advocate because it knows what I want.

Frederick Vallaeys: And talk a bit about that, how does the advocate know what you want, how does it know motivation?

Tom O’Malley: So this is actually really important to the project, and I think to the potential of personalization.

AI, what we found, is a good conversationalist, generally, right? It gets it going, especially if you can ground it in some form of reasoning, or make some sort of objectives for it. And the more narrow those objectives are, the better it does. So, if we teach our advocate to be curious. about my cultural context, where I’m from what belief systems, general values that I have.

Personality, that it can use a disk framework and map me to it, right? That it can be open to ingesting all of my skills from different, you know, whether it’s LinkedIn, email, calendar, what have you. It can ingest from other data sources. And then if, if we can have conversations around my goals, it can infer them and I can qualify them.

If I, if it can understand me or store me, nodes at those six layers and put that into a container and we’re using knowledge graphs. We just think knowledge graphs are a great way to sort of hard code linkages between nodes and then use that. And we found out that AI is a really good producer of knowledge graphs, meaning a digital thumbprint of somebody.

And so if we could use that then digital thumbprint to become the grounding for my agent, Now it’s personalizing my experience with that agent back to me. But what we think is entirely like next frontier thing is taking that, that thumbprint and asking it to personalize other experiences on the web. So now I’m going to log into some other website that is call it personalization compatible or has a chat bot.

I can ground that chat bot back to my, My, my personalization thumbprint.

Frederick Vallaeys: You can instantly do

Tom O’Malley: that because it’s got that profile? And I don’t have to expose that, that data to the source. I can just ask to filter it through. So don’t show me a page that I don’t want to see. Don’t take me to content. Or don’t talk to me in a way that I’m not going to understand it.

Put everything back in the context of me through my agent.

Frederick Vallaeys: And that preserves the privacy of those conversations that you’ve had that have built up who is Tom, that fingerprint, basically. Yeah. But it makes the whole web potentially more useful. So that sounds like a lot of potential, many things that this could improve.

Tom O’Malley: Well, but, but I think the net of that message around, you know, what is AI to connecting B2B marketers with their target market, it’ll evolve, it will become. It’s a frontier of personalization. Our expectations over time are becoming, I don’t have time to hear things that aren’t accustomed to my ears. I don’t have time to listen to things or meet people that I don’t have a mutual fit with.

Like, even if I want to have a meeting with somebody, if I can’t add value to them, nothing’s going to come of it. So I should stop wasting time on talking to people. That look like somebody I want to sell to, but I know probabilities are very low. They’re going to close. I’d rather spend my time using filters to make sure I’m always being matched with somebody with high personal or rather mutual fit.

And my only point is AI will prove to be the ultimate personalizer and take us into that next era.

Frederick Vallaeys: Yeah. Is there a risk for living more in an echo chamber as it sort of blocks you from maybe having conversations that it doesn’t believe are immediately useful to, in a mutual way, right? And it was very important how you said it’s not just about being useful to you, but it has to be mutually useful.

What do you, do you feel there’s a risk that it’s going to sort of narrow the vision that we all have?

Tom O’Malley: Yeah. So what’s kind of cool about this as well is your advocate is also your coach. So it can identify blind spots. It’s not just looking at you. It’s able to look at a thousand other profiles and, and fingerprints like you.

And say, hey, Tom, like, you’re crushing it in this area, but you’ve got some blind spots that could get exposed here in the future, so let me give you some content, or let me introduce you to some people that can open up some of those blind spots for us. So, I think that personalization matched with benchmarking is the answer to stay like, You know more self checked as we enter in this personalization era.

Frederick Vallaeys: All right And what’s also fascinating that is as ai takes over some of the roles like you said in current, right? It’s now writing the briefs. It’s in moderation Hopefully you can also identify like what is that blind spot in the area? That’s still going to be relevant where you do add value to whatever project you’re working on.

Because that’s ultimately the big question, right? Like Sam Altman went out there and said, hey, 95 percent of marketers, you’re going to be irrelevant within five years. So what do we do, right? There’s clearly something to be done, but what is, what is it we do?

Tom O’Malley: Yeah, I mean, you know, our vision for this, and it’s, and it’s full form, and it’ll take us some time to get there.

Your advocate is also a coach. If you can do benchmarking, which is a snapshot of where you are, you can also do a benchmark on where you want to get, and do a learning plan on how to get there. Right. And so we, we think that the future of, of professional development, it starts with networking and aligning goals, and then it moves into benchmarking for today.

How can I improve as a person? But then it clearly becomes where do I want to go or who do I want to become? And it can be that same mechanism can be applied.

Frederick Vallaeys: And it’s interesting as well to me that what you’re describing is a very intentional way of. Helping your advocate understand who you are, because today, Facebook, Google, Amazon, they all understand something about you based on your purchase profile, your social media behaviors, but they own that, right?

And you have very little insight into how they understand you. And Google does a little bit of a decent job. They show you, it looks like Fred is into cars and sneakers and these types of things. And you can uncheck the box if that’s not true, but I don’t think it goes to the level that you’re describing of what is my motivation?

Like, what is it about cars that I like? What kind of cars? And so there’s that ownership of data and that sounds really appealing that you have your advocate and you talk to it and you can sort of like deploy it in the areas where you see it to be useful. But you can also rein it back in and you can say in this area, just let me be me, no advocate necessary.

So how do you think about that privacy layer and that maybe portability of the thumbprint?

Tom O’Malley: I mean, I think you just nailed it. That thumbprint needs to be inherently portable. I think the web three era. There’s so much goodness that came from that movement, right? And data portability, and the vehicles to make it portable.

Like, all the tech is out there now to solve our problems. We just need to apply it. What I don’t want to sound like is we’re an AI blockchain based startup that’s got all the buzzwords. But I think down the road, portability is implied. It needs to happen. Whatever security mechanism is the right one to use, we’ll apply that.

But that’s our in our mind is that that thumbprint is is yours that is you And it becomes a personal asset. And I think, you know, it should be open to APIs to ingest my data as, as data portability and data ownership becomes more of a standard, we’re going to want to feed it into us and into a central repo of our own to be able to define our own.

Digital self and use that to our advantage

Frederick Vallaeys: and I suppose it’s okay to be blockchain and web 3. 0 and gen AI If you’re actually solving a problem, right? It’s not the technology trying to find a problem, but I think we could all agree We know too many people but we don’t really know what those people are up to you lose connection with folks over time Yeah and I think over the last four years, like I used to go to so many in person conferences and that’s really happening much, much less in the United States.

And so folks that I would see all the time, and I’d know what was going on in their lives, what they were working on. You know, if I’m not super active now on social media, like you lose, you lose track and we could actually really help each other. And I do find that a problem. And if this is the solution, and if, if blockchain makes it more trustworthy and more private and more portable, then great, let that be.

And if Gen AI is the one that. Pulls it out of me what my motivation is then great. Let that help me as well Because it is a real problem that you’re trying to solve.

Tom O’Malley: Yeah, there’s a myth that somebody I meet today is more relevant than somebody I met five years ago It’s not that that that person I met five years ago is still the same person that they are And today we just have this idea that we got to pile on more network is more value but everything shows that it plateaus and can even Degrade or go down.

Diminishing returns happens first. I mean, and you just get up to that 150. And that piling on effect, it’s not additional. So if we have a mechanism to, like, reconnect our network, somebody you met five years ago might be the most relevant person for you to re engage with today. And in fact, having that history, would really be a bonding agent to, to, you know, start a new relationship or go achieve something together.

The fact that you met five years ago. But we have no mechanism to track and trace each other over time. Advocates have just infinite amounts of calculable power slash memory to know when our, you know, Our mutual motivations, or our motivations are mutually aligned.

Frederick Vallaeys: Now you’re moving back to the Bay Area.

You used to live here. You moved outta state for a couple of a couple of years. Tell us a little bit about what that decision process was like.

Tom O’Malley: Yeah. You know tech was getting boring. I’d say. Palo Alto lost its vibe, I’d say about seven years ago or at least that’s when I realized that I kind of felt it was getting very ty, very lawyery.

And I really like working with entrepreneurs who are just like doing it for the creation, you know, not necessarily for the payout, though the payout’s inevitable if you create something valuable. And so that spirit, I felt like, was like, we were losing it. Yeah, Palantir,

Frederick Vallaeys: Palantir sort of sucked it out of the downtown.

Tom O’Malley: Everybody was getting priced out, you know, is what it came down to. But I just felt like the spirit was gone. I just, I needed to take a break. And I took a hiatus and went down to Coronado focused on my daughters a little bit more, got them through high school. And then I started a a Mescal brand.

In fact, I mean, to my surprise, it was probably the best thing that happened to me, just to open my eyes back to like, what is branding, what is product, what’s the relationship between, you know, leadership and, you know, how to effectively move in an environment where. We’re an industry that doesn’t change very much that isn’t tech.

Right. So all these kind of things that I think I was taking for granted in, in the tech space became much more evident in the non tech space and in particularly the luxury spaces. It’s a luxury Moscow. So anyway, it was a great little break, lived in Miami for a bit. And Miami is like a whole different culture.

Though strong in Web 3, just doesn’t have the density or, you know, the serendipity that you get walking the streets of Palo Alto and I started hearing that the vibe is back in Palo Alto, that entrepreneurs that have fresh ideas about this new technology are sort of flooding back into the valley.

Frederick Vallaeys: Let’s talk about that a little bit. Let’s, let’s double click on it. But I think that excitement is largely driven right now by the wave in generative AI. Yeah. What is your take in terms of gen AI? as a feature to existing SaaS versus like entirely new products.

Tom O’Malley: Unlike SaaS, if you’re a pretty good engineer, you could like create a pretty smart product through some automations and, and the way you store data and retrieve it and then user experience.

And so I think that was reflective in the startups that you saw. Like, I’d say on average, I don’t have the stat, but my gut tells me that product probably only owned about 25 percent of the shares of the founding shares of a typical SaaS company. In the AI movement. The projects that I believe in tend to have a much larger share set aside to real tech.

And so I think we’re back into an area of real tech. And it’s a little emphasized right now in like the NVIDIA’s, right, the infrastructure layer. And it hasn’t yet manifested in the application layer. But what I like about being an old guy that’s been around is I understand the application layer a lot because I’ve lived with business to business problems, whether that be in marketing or operations or sales or hiring, right?

And so you have to really understand the problem. To make an impact in the applications layer. If we’re just going to strap AI on top of existing software systems, you will get a benefit. But is that going to be the next generation thing? I doubt it. So I expect there’s going to be a wave of startups that have heavy tech teams that are focused on business problems that are not looking for just incremental gains, but like 10x, 100x gains on the era that we’re coming out of the SaaS era.

Frederick Vallaeys: Yeah. And so I think existing SaaS companies, for example, they must use generative AI. It’s like saying, Hey, we use electricity, right? You must have gen AI. Because it makes things faster, it’s fairly obvious how it can help you. And then you have sort of that big vision of entirely radical shifts in how business is done.

And Jenny, I can also drive that. Where I sort of am worried is that you have this bubble coming again of, hey, we just took chat GPT, we fettered a couple of prompts, we did a little bit of fine tuning, and boom, here’s this amazing kind of thing. Whatever it is, right. But, but it’s really just a different skinning on top of the existing LLM.

But it sounds like what you’re talking about is much deeper tech. There’s a lot of model building stuff that actually takes real effort. And that’s going to drive that next wave of whatever comes out of Silicon Valley. And

Tom O’Malley: it’s not just a startup problem. It’s a, it’s a funding problem as well, because I think VCs that were in the SAS era are still looking for they want to see revenue really early.

They, they’re more inclined to get excited about something that’s showing traction because that became our language, right? You had to show traction to get funded in this new, deeper tech era. The traction is going to be a little bit further down the road. There’s a little bit more engineering ahead of time if it’s going to be real and have a defensible piece of technology.

It’s going to move that, that, the stages back a little bit. And so what I find now is while I’m out there, if I’m talking to sort of a VC that started 10 years ago, they’re still looking for those older metrics. Whereas funds that have been set up recently for AI are actually looking for metrics like how much of the founders shares are set aside for tech.

How much of your model is going to be built and how much data. They want to see the data roadmap, not just the traction roadmap. And so it’s going to take a shift in ecosystem to get it right. And what you’re saying is probably true. There’s going to be a little bit of a flash bubble on the. On the prior until we get to the latter.

Frederick Vallaeys: And you want to learn from history, but it sounds like you also want to, this is probably one of the bigger changes we’ve seen in Silicon Valley. And there’s sort of this this notion that it comes in 30 year waves. Yeah. It was microprocessors in the 1960s that made computing basically marginal cost of zero.

30 years later in the 1990s, it was the internet that made the distribution, the marginal cost of that basically approached zero. And now we’re at. the stage where the marginal cost of creation of content becomes zero thanks to generative AI. And that is such a big fundamental change again. And we’ve seen all of these other In these 30 year cycles, there’s so much other stuff that happens, but that just doesn’t quite have the huge impact.

And so I think, yes, let’s look at history, the KPIs that have worked, but let’s also sort of be more forward looking and maybe acknowledge, like you said, that there’s different KPIs that will matter in this the wave that we’re currently starting.

Tom O’Malley: Yeah. I mean, you know, software ate ate the world, right.

And then if you were a software company. And then the internet came, it’s almost like you said, you don’t have to say like, oh, we’re also going to do internet. No, it just became internet. Now, the internet, if you’re an internet company, of course, you’re an AI company, like, or there’s no survival. It’s like the way you go.

So yeah, this is one of those foundational shifts. I think the jury is no longer out on that. That is just what to anticipate. Now, I do think, though, that people overestimate what today’s AI can do. Yeah. It’s limitations once you start getting into it and start asking it to do so. And when you start playing with it, it’s, it’s got limitations, right?

And, and you hear about it and like hallucinations and whatnot. But even, even like grounding or training, we’re still at the nAscend stage of this journey, which to me excites me because it just means there’s so much more innovation to come. And though we took the great leap, we’re still on an arc that is going to be very exciting to be a part of.

Yeah.

Frederick Vallaeys: Right and then so AI has obviously been building up since what was it the 1960s and then you had the AI winter But it feels like there was that fundamental shift that happens when Chad GPT and the transformers were built by the Google brain lab. And it’s just, yes, there are limitations.

It’s not doing everything perfectly, but I think if we sit today and say, Oh, well, you know, it’s, it’s still doing my, I can do my job better than it can because it hallucinates or because it messes up here and there. Like that’s very short sighted. Right. I think a hundred percent,

Tom O’Malley: like, you know I don’t want to say that it’s not.

A great loop forward. It is a great leap forward. But if you feel like you’ve missed it, you haven’t. Yeah. Right? It’s so easy to catch up right now. I mean, I recommend, for anybody that’s feeling behind, 15 hours of YouTube. There’s just so much stuff out there of people explaining it in very logical, you know, tangible ways.

But there’s a lexicon, there’s a certain, you know, there’s a language you’ve got to pick up. And then And then you’re brought up to speed, and now you can have intelligent conversations as long as you’re conceptually on board with, you know, how it works. But to

Frederick Vallaeys: what degree do you think it’s important for the average, say, like, marketer to know these concepts versus just saying, Oh, let me go and play with Claude, let me go and play with GPT.

And really, instead of those 15 hours of watching videos on YouTube, like, should they be playing 15 hours and, like, spend 15 hours figuring out how it can, help you in your current challenges in your job?

Tom O’Malley: Yeah, it’s a good question. I mean, I guess it depends on what your job is and the level of which you are hands on in innovating new solutions for your job.

I mean, what you want

Frederick Vallaeys: your job to be, right?

Tom O’Malley: I mean, I think there’s a new language. I think there’s a baseline language that you need to understand about AI. Because if you’re not, then you’re not understanding sort of what’s the magic behind it. Like, it was almost like going and just using the internet, not understanding that servers are connected, you know, through a cloud environment.

I think that you, it’d be good to spend some time getting the fundamentals, because there’s a lot you can do at the API levels. Like, messing around with cloud is interesting, it’s fun, it’s very self serving. Like, I got something, I got a report I gotta write, let me use that, you know. And it’s very helpful as a tool, but if you want to change your job or change the environment that you work in, it’s I think you got to get a little bit more savvy with understanding how it works so that you can maybe pair up with a growth hacker and, and, and become the, the sort of product manager of your future marketing stack.

Because that stack is going to take some time and, and again, I think personalization is the era that we’re in. These APIs are so malleable to apply it to your specific need. And it’s, so if you can pick up a little language around what does it mean to prompt engineer? What’s so great is prompt engineering really doesn’t take a coding degree.

Like you don’t even need to be that savvy in Python. You just need to see a couple of templates to understand what are the components in a call that I might want to have say in as a marketer. And then pair up with somebody that will actually write the templates. And do the actual engineering, but, but I think to be a part of that, there’s like this new relationship in marketing where kind of like when you built a website, you kind of had to understand and work with the person that actually coded the website to understand the limitations.

I think it’s kind of similar here.

Frederick Vallaeys: And the way I think about it is in my world, ad scripts, if you’re looking to automate or be more efficient in Google ads, you can get a piece of software like Optmyzr or any of the other ones that are out there. Right. But all of these softwares do have limitations because they are ultimately coded by a group of people who make product decisions about them.

And so where the scripts come in is that you can take that same methodology, but slightly tweak it. You can only do that if you understand the underlying. Components of you know, what does the Google ads database look like? What are the components of a keyword? What are the components of targeting?

What are the capabilities in these APIs? And I think, I think it’s much the same now with generative AI. It’s like there’s base level capabilities that all of these SaaS companies are adding, but there is more that you could do something as simple as changing the temperature on a keyword generation or an ad text generation.

And if you don’t know what that, I know you know what that means, but if you don’t know what that means, like. That’s, that’s what Tom is talking about. These are the fundamentals, like understand what the temperature can achieve for you. And then you can go and tweak it and you can get it to do exactly what you need it to do.

And now you become the leader. within the organization who’s figured it out for your company. And that’s, that’s one way to keep your job for at least a couple more years.

Tom O’Malley: I, I think you put the nail on the head there. I mean, when I talk to folks you know, across different roles and, and whether it’s marketing or operations, there’s a, we’re at a point where there’s an opportunity for leadership to emerge.

That just sounds AI, intelligent, you know, and so as long as you’re not the one afraid of it, you’re probably not going to get run over by it. And if you can actually start leading some internal conversations and finding the individuals, I promise you there are people in every organization. That are maybe closet hobbyists with AI right now, find them in your organization and start an AI, be part of that AI movement, because it’s coming.

It’s, it’s, it’s parallel was the internet. Before the internet, there were people that ran away from it, tried to hide from it, refused to use it. It’s going to be a fad, right? And then those that actually lean forward into it, they sort of progress their jobs. They progress their, their reputations as learners, as you know, growth mindset.

So I think there’s an opportunity here to go the opposite than the fear.

Frederick Vallaeys: Yeah, makes a ton of sense. Hey Tom, thank you so much for sharing all your thoughts and the companies you’re working on and what’s worked for you. And thank you for watching. If you’ve enjoyed this episode, which I’m sure you have with a great guest like Tom, hit that subscribe button at the bottom and we’ll let you know when the next episode comes out.

And then as far as getting a hold of Tom and all the cool projects he’s working on, Tom, tell us how to do that.

Tom O’Malley: Thank you. Yeah. So please. The most exciting project we’re working on right now is ascend. So you can visit us@www.my ascend.ai and you can sign up there for the stealth release.