AI Doesn't Drive Revenue Today! | Ep. 3040

So Eric and I are here in San Francisco at HubSpot's inbound event. I don't know what year they're on, but they've been doing this for a while, usually in Boston, this time in San Francisco, which is actually more convenient for Eric and I. But what was interesting at HubSpot, and Eric and I have been coming to marketing events for a while, this event typically has quite a few SMBs. And SMBs, when it comes to marketing, tend to adapt faster, which I'm assuming you agree, than enterprise because they don't have all this legacy and archaic stuff to deal with. So during HubSpot, Eric and I asked the audience, how many of you guys are using AI? Almost everyone raised their hand, you know, AI within their marketing. And then we asked how many of you guys actually saw a 5% increase in revenue because of AI and usage in marketing? Very few people raised their hand, call it like 2, 3%. Then we asked how many of you guys saw, you know, any sort of increase? And pretty much it was similar, only a few percentage points. The point, you know, I'm making here is everyone's using AI in marketing, but very few people are actually seeing results from it. And when I say results, I'm not talking about time savings. The purpose of marketing is to generate more leads, more sales, more revenue, and more profit. So the real question is, is why do we think that people aren't generating more revenue from AI?

Yeah. So to set the tone here, I think it had to be less, probably less than 5% of people that raised their hands when this question was asked.

Yeah, right.

But everybody's using AI right now.

Everyone, almost 100% raised their hand.

Yeah.

And again, this is just AI within marketing. They're probably using it in many other parts of their organization as well?

Yeah. I mean, in my mind right now, I think the way people are using it right now, it's— they're gaining a lot of efficiencies, but it's hard to track it though, like from an apples-to-apples comparison when it comes to numbers. Right. So sure, they might be creating more content, for example. That's the easiest example when it comes to marketing.

You agree?

Yeah, right. Or maybe you're creating more ad creative and things like that, but you're, you're doing things faster, but you know, you're not necessarily seeing the ROI from it out the gate right now. So yeah, people are spending a lot more on tokens. There's a lot more infrastructure investment that's happening right now. But I still think it's, in my opinion right now, it's a little early. I certainly, I think you feel faster. I feel faster as well. But are we making dollars hand over fist? Not right now. And so I kind of want to see what you think and then we can continue this conversation.

I don't think the issue is that it's early because, yes, we're in the early innings of AI, but it's still been out for years and marketers have a ton of software options that have AI integrated. I think the bigger issue is, is companies focus on the wrong stuff. They use these AI tools, they do cool things like, wow, look at this video that I created. All right, what are you going to do with the video?

Oh, we'll just throw up ads.

Well, did you think about the script? Did you think about the messaging that you're going to use to convert that visitor who's cold into someone who wants to get to know the brand or the business and then eventually end up converting? And the big issue that I see is AI put in the wrong hands just causes people to spin their wheels and do a lot of useless stuff that seems cool and flashy. But unless you actually have a KPI mindset and you know how to take your marketing and not just create more output, but instead focus on what is actually moving the needle, I think you're screwed. And most of these companies, when you really dig into it, They don't look at their marketing from, hey, what's actually causing revenue growth right now? And whether those channels are profitable, and how can I use AI to make those better? Instead, they look at AI as in, what can I use AI for marketing for? And I think it's the wrong way of thinking.

Yeah, so I think also the problem is when you have that many people, we're talking thousands of people in the audience raising their hand, when they say they've used AI, so when I interview people, say, oh yeah, how's your AI fluency? How would you rate it 1 through 10? 9 out of 10, 10 out of 10. We've talked about this last week, I think, right? Or Monday we talked about this. And when you start to dig in a little more, it's like, okay, so, oh yeah, so have you built any agentic workflows or agents? Like, oh yeah, I do all that, right? So yeah, tell me, what have you built? Oh yeah, like I just got some certifications and da da da. Oh, okay, you lied to me. So my point of saying that is that most people are using it at a very surface level. The other thing is when the examples that I share with you or that we see on X that you share with me sometimes, yeah, cool, you know, Google Nano Banana, awesome, right?

Yeah.

You can create all these things, But they're not— I think they experiment with it, but they're not actually building things, right? Like, you have to go the extra mile with this right now. And the stuff that I talk about— oh, my microphone just fell. But the stuff I like to talk about is, okay, we're— I talk about Claude Code, I talk about these agents, right? Like, even with that, yes, I'm getting reports faster. Yes, I'm telling people what to do. But if, if I'm not actually executing on these things, if I'm not actually putting these things to production, then it's as good as not doing anything at all. It's, it's, it's nice for us to talk about on this podcast and nice for me to share. But if we're not actually putting it to production, and people don't know how to use it this way, then of course, you're not seeing an ROI. That's what it is. Alright, so I wanted to take a moment to tell you about my podcast co-host Neil's agency called NP Digital. And they work with a whole host of global companies or global organizations. Also, Neil has SEO tools such as Ubersuggest and AnswerThePublic. All you have to do is go to npdigital.com to learn more and we'll see you on the other side.

Yeah. And when you look at Nano Banana, Banana or whatever, and people using it to showcase their products, it's— they think like, oh, this is cool. Yes, but how is it going to actually increase your revenue? Just using these tools isn't enough. And going back to what you're saying on AI, when I talk to people on if they use AI, they think paying for marketing software that has AI in there or talks about AI is actually using AI. No, it's not. Yeah. And I I have that same pet peeve with agents right now. Everyone's talking about agents and you see all these techie entrepreneurs being like, oh, this is the new future for agents and this is what we're doing that's cool. And I'm like, guys, you're making things too complicated for people. It's not even agents. People just want to pay for software that just does shit for them. Excuse my language. And just completes the task.

And I don't think we're there right now. I think, you know, you and I are building things that will, that will get there. A lot of people are building things that will get there. And I think it's important for people people that are listening right now and people like you and me to study these things. But I do think what we say, the one thing that sticks out to me is money amplifies your nature and then AI amplifies your mind. That's where we are right now. And a lot of the people that say they're doing AI, I think it's probably 99% or so, they're just scraping the surface of it.

Yeah.

And so you have— you do have to go deeper. You have to experiment with these things, but you also have to push these things to production to see what the real ROI is. Because I'll tell you this, like We have— yes, we have this SEO orchestrator sub-agent in Claude Code. We have this SEO A/B tester optimizer. And yes, it will map out A/B tests to run. It'll map out impact, confidence, ease. It'll map out the revenue, the estimated revenue. But if my team's not pushing that stuff to production, it's not doing anything. It's, it's, yeah, it's mental masturbation.

That's right.

Yeah.

So I think that will change in the next year or two because companies, what we're seeing is a lot of them are spending more because of AI, But they're not seeing any revenue growth. And some of them are getting fed up. I was at a large corporation a few days ago here in the Bay Area. When I say large corporation, I'm talking about like Fortune 100, literally one of the biggest 100 companies in the world, much higher than the 100 list, right? So they're quite a bit high up on that 100 list.

Top 20?

No comment. But yeah, they're really big. So Internally, I was talking with some of the executives and they were breaking down how they're using all this AI stuff within marketing. And one of the executives straight up just said, he's like, yeah, our budgets are increasing. Keep in mind, this company spends like billions and billions and billions of dollars every single week. Like they're massive, or per month, whatever you want to call it. And when we were looking at it and they're just like, yeah, we're starting to spend a lot more on all these tools. But they're just like, all this is doing is increasing our budget. Where the heck is the ROI?

Yeah, well, my take on it is this, and I think you probably agree with this. We are investing in the future, just like the early days of the internet. You know, Cisco was really big. Everyone was investing in infrastructure for the internet. I think that's what we're doing with AI right now. So we're not saying don't study this stuff. Absolutely. Like, I love nerding out on this stuff all the time. Yeah, probably a little more than you do. Yes. But, but you— so you have to keep the pressure on, but you have to be a little more patient and just understand that Yes, keep investing. Yes, keep learning. Yes, you have to push stuff to production. And also you have to push AI fluency throughout your organization too. And I was actually listening to the Airtable CEO talking about how they're pushing AI fluency throughout the organization.

Yeah.

So they— what they have is they have a fast thinking team and a slow thinking team. So things that are fast, maybe you're vibe coding things, you can, you can execute on it really quickly and push it out. Things that are slow where you're, you're making infrastructure changes or pushing it to like hundreds of millions of lines of code and things like that, maybe you need to take a little more time there. But he came up with a new phrase that I'm interested in, that's interesting to me, called the IC CEO. And you're not a—

Wait, what?

IC stands for individual contributor. Okay, so he's the individual contributor CEO, meaning that he's in the trenches, he's executing on this AI stuff. Aaron Levy from Box does that as well. I look at us as IC, probably more CMO-ish. You don't— you have the CMO title, I have the CEO title, but I think you and I I'm not really the CMO of my company. Yeah, but you have the title. You're just kind of everywhere in your company. You're like Dharmesh.

Yeah, I should probably get rid of the title.

But Dharmesh has a CTO title. See? But yeah, even though he's not really CTO. Yeah. And like, I don't really think the CEO is a fit title for me either. But like, I do a lot of the CEO's job right now, right? Yeah. But my point of saying this is that the IC CEO or the IC executive is in there in the trenches testing things. And even though like You're not reading up on the latest on AI all the time. You're still in there testing things, right?

Yes.

And you have from— we're just— the point I'm bringing this up is you are a large shareholder in your company. I'm a large shareholder in my company. Right. And it has to start from the top. If it doesn't start from the top, then you're slowly going to, I don't know, slow down and kind of eradicate yourself. Yeah.

You have to get the fluency throughout the whole organization. But at the same time, I think they all need training on KPIs and understanding that when you're using this kind of stuff, you got to figure out how can either help you save time and then what can you do with the additional time that you just saved to help the company? And then the second thing is, is what can you use it for that actually moves the needle for your main KPIs for whatever your job is?

Yeah. So by the way, it's— I actually like that you picked this topic because this is a good topic to lead off. And what I've been— what's been annoying me about our podcast on YouTube is it takes forever to grow. And so I'm like, the packaging is not great. I'm like, okay guys, here's what we're going to do with our packaging. We're going to pick the best idea to lead off with first, and that's just going to be the headline and that's going to be the thumbnail. And then like if they like the first 10, 15 minutes and we have a good conversation, they'll probably stay for the rest and then there'll be good retention. But the way we're packaging is like, oh yeah, AI innovation in 2025 or, you know, revenue ideas, trends for 2025.

It doesn't work as well anymore. It's not good.

Yeah. But if we really want to grow, like you lead off with a good idea and you should be good.

Yes.

Yeah. So I just wanted to give that a little bit.

Yeah.

Yeah.

Yeah. And when you're doing social content, you don't have to describe everything that you talk about in the video. Just pick the best thing that the video starts off with because that is in essence your hook.

Yeah, by the way, and if it's newsworthy, it's going to do well too because we record weekly. But speaking of AI, so this week I talked to you earlier this week. So, so Neil and I recorded on Monday. Today's Thursday. Neil spoke already and then we're marking schools today and I have my session tomorrow. But a lot has happened in the news cycle. So Oracle has begun laying off people. I think they cut 3,500 this week. Tenure doesn't matter. So basically what I'm seeing here, is, okay, Oracle layoffs are intense. So remote and in-office, top performers, doesn't matter. People with 7, 13, 18, 20+ years at Oracle gone. And so if you got the email for saying project updates, that's the bait. You join the call, an HR rep reads the statement, access cut within 5 minutes, laptop wipe. That's it. And so here's, here's some of the quotes here. So this morning, SVPs laid off. So senior vice presidents laid off, longtime employees ghosted by leadership. One-third of some teams are already gone. One Oracle vet— this is the punchline here— I just got traded in for a GPU.

Oh, wow.

Yeah. And then Salesforce and we know this already.

But let's go back to Oracle really quickly. Is what roles are they eliminating right now? Is it related to like customer success or not? No, it's customer support.

No. Okay. So it says cloud comms, marketing, engineering, ops, sales directors, ICs, even SVPs, all levels.

I bet you they're just cutting out the bottom performers.

They are.

I don't think it has anything to do with tenure. It's like, hey, we've gotten too fat. Let's just cut all the people that aren't performing.

But if we go to Salesforce for a second, They've cut most of the customer service support people. So he said 4,000 or 5,000 people. Marc Benioff, the quote from him is, I don't need that many heads anymore.

Yeah, but have you seen his stock?

No, I haven't. What's the deal with the stock?

It slumped again, like, I think like 7% today.

So he just said like, I need less heads. I'm reading this from your favorite— oh, sorry, this is not CNBC. This is NBC Bay Area. Salesforce confirms 4,000 job cuts because I need less heads. That's the quote for him. So I reduced it from 9,000 heads to about 5,000 because I need less heads. And, you know, Scott Budman tweeted, Salesforce confirms 4,000 layoffs because AI means they need fewer people. So I don't, I don't think you and I are necessarily celebrating this. I think we're just calling out that this is happening more and more.

Uh-huh.

And that's why you're better off being more hardcore about AI in an organization and getting your people to be more resilient, because I don't want to be the person that ends up hiring a bunch. We talked about this the other day when we're walking. I'm like, I don't want to— like, things are going decently right now, like knock on wood. But if we keep compounding, we throw people at the problem. That means we're going to have to cut people. And I'd rather not have to. And you and I know a lot of people that have cut hundreds, if not thousands of people. And we're seeing this happen right here. Yeah.

And I think it's going to continue.

Yeah.

And this is also why one of the reasons why the job market is weaker, in my opinion.

Yeah. So you saw that report today. What's the deal with it?

Yeah. So ADP reported, I think it was like 40,000 or 50,000 new jobs created, much less than they expected. This isn't the government's data. This is just ADP's data. Although I think ADP's data is more accurate because they don't do tons of revisions like the government does. Not trying to talk crap, but it's just the reality. And when you look at ADP's data, you know, the job market is just weakening and weakening. A lot of large corporations are laying people off because of things like their corporation isn't performing as well, or AI is helping replace some people, or Actually, those are the main 2 reasons that we're seeing. But in this market, I believe this is going to cause the Federal Reserve to have no choice but to reduce rates. The rest of the world is pretty much reducing rates. We're the only one that's not.

You know, it's interesting on the rates piece. So we're not going to go too much into economics. But you look at the, the, the, the ECB, European Central Bank, right? They cut 5 times, I think, in the last 12 months. But their— the, the interest rates have kind of remained— or sorry, the yields, Treasury yields have continued to go up. Right. So it's not, it's not really reacting to the interest rates, which is interesting to me because usually when you cut, it should reflect on the yields.

It depends for which yield. So the 10-year looks at what it's going to be roughly 10 years from now.

Yeah.

So it's not directly 1 to 1 when you're cutting.

Yeah. And mostly people talk about the 10-year. You got the 10-year and the 30-year.

Yeah. And then you, of course, have short-term, but If the Federal Reserve cuts, you know, the Fed rate, I think SOFR right now is— our Fed rate is 4.25 to 4.5. SOFR is around 4.4 the last I checked. If they cut it 1%, you know, I'm just going with that. If they go 1%, it doesn't mean the 10-year goes down 1%.

Yeah. So they're saying— I mean, what I've read is— and Neal's right on that. But usually what happens, though, is if they start to cut like 75 basis points, whatever, Treasury yields start to come down and people start to go further down the risk curve. They start to bet on, on more risky things. But again, if, if we're seeing a pattern here with the ECB and it's not really moving that much, it's like the market's not really moving that way. That means, that means we're— and I'd be keen to say, so George Gammon actually listens to this. George, I kind of want to know what you say here. But that means what he predicted and what you and I predicted years ago, actually, like we're going to be right at some point, right? We're like, but it seems like stagflation is going to come. Yeah.

It could be, but when the Fed reduced the rates last year, it ended up making the 10-year go down.

Yeah, we'll see what happens.

The 10-year for most people is mainly relevant for mortgages. Yeah. And the reason I know this is I— no joke— I just redid my mortgage when they ended up reducing the rates.

What is it? I think it's cheaper now to buy a newer home than an old home in some cases.

Because of maintenance or for what reason?

I think it's I think it's a combination of things. I just saw— I just saw a graph that it's just cheaper for a new home now.

Yeah, I didn't realize that.

Yeah, but okay, so but to speak on the ADP jobs report over here, because I have a— I have the internet and Neal doesn't have the internet. So ADP nonfarm payroll jobs created 54,000. The projected was 73,000. So pretty significant miss. Previously it was 104,000.

Yeah, they did 2 revisions.

Yeah. So that's, that's what it is.

Technically one revision, but yes.

Anyway, we'll catch you later, guys.