Made possible by

Rokt is the global leader in ecommerce, unlocking real-time relevance in the Transaction Moment. Rokt's AI Brain and Ecommerce Network help the world’s leading companies deliver more relevant customer experiences and unlock incremental value from every transaction. Learn more at rokt.com

[00:00:00] Marcus Johnson: When the consumer reaches checkout, they're no longer browsing, they're buying. It's a moment of peak intent, attention, and engagement. That's where Rokt comes in. Rokt helps brands reach customers at the moment that matters most, delivering relevant offers and content that feel like a natural part of the transaction experience, not an interruption.

[00:00:16] Marcus Johnson: You can learn more at rokt.com[00:00:20]

[00:00:23] Marcus Johnson: Hey, gang. It's Monday, June 22nd. Gadjo Grace and listeners, welcome to Behind the Numbers: an EMARKETER podcast, made possible by Rokt. This is the show that keeps you up to date with everything media, marketing, and technology in about 20 minutes or so. I'm Marcus, and joining me for today's conversation, we have two of our AI and tech [00:00:40] experts joining us from New York, Senior Analyst Gadjo Sevilla.

[00:00:42] Gadjo Sevilla: Hey, Marcus. Hey, Grace. Happy to be here.

[00:00:44] Marcus Johnson: Hello, sir. We're also joined in California, uh, by Grace Harmon.

[00:00:48] Grace Harmon: Hey, guys. Thanks for having me.

[00:00:50] Marcus Johnson: Yes, indeed. Uh, thank you for joining us as well. Today's fact

[00:00:57] Marcus Johnson: Okay, this one's for Gadjo and all you Knick fans out [00:01:00] there. About 70%, 7-0, of New York City's population was not born the last time the Knicks won the NBA championship back in 1973. Of course, they've just won it, so it doesn't count anymore, but until then, you'd have to be over 53 [00:01:20] years old to have been alive.

[00:01:21] Gadjo Sevilla: That's crazy. It's like looking at newsreels, right? Just-

[00:01:26] Marcus Johnson: The footage- Yeah ... is so bad. Um, yeah, congrats to the Knicks for winning. Congrats to, to Gadjo, Stuart runs the team, my dad, um, all the little Knick fans out there. I've got some other interesting Knick stats [00:01:40] for you quickly. So they had a historic comeback.

[00:01:42] Marcus Johnson: Um, it was the game four. It was the most watched game four since Michael Jordan's last championship, number one. Number two, this is the only series in NBA history, um, dating back 30 years when they started tracking this kind of thing, where every single game was within five [00:02:00] points with under five minutes remaining on the clock.

[00:02:02] Marcus Johnson: So incredibly close series by historic standards. Um, game four of the finals, um, was, uh, not only the most watched, but when Knicks pulled off the largest comeback in NBA finals history after recovering from a 29-point deficit, so Gadjo remembers that vividly. Uh, Jalen Brunson became the [00:02:20] first and only New York Knick player to score 40 or more points in an NBA finals game, and then the last one here, this one's fascinating.

[00:02:27] Marcus Johnson: History repeats itself. So when you compare the last Knick team to win, 1973, with the 2026 Knicks, through the first four games of the finals, there are just some insane parallels. So all four [00:02:40] games, the first four games, uh, were within three points in the last two minutes back in '73, uh, same in 2026. The total number of points deciding four games, the first four games, was 16, total number of points deciding them in '73, same in 2026, and in 1773...

[00:02:59] Marcus Johnson: Sorry, [00:03:00] 1773, that's too long. That's too, that's three years before America. Uh, in 1973 , the Knicks closed out the series in game five on the road, same as they did in

[00:03:11] Grace Harmon: 2026. Phenomenal.

[00:03:11] Marcus Johnson: I would talk about basketball all day if I could, but Stuart's gonna kill me for going so long on the fact of the day, but it's for him because he's a Knicks fan, so you're welcome.

[00:03:18] Marcus Johnson: Today's real topic, Apple's [00:03:20] new Siri is about to crash the AI party.

[00:03:27] Marcus Johnson: Harry McCracken of Fast Company explains that at Apple's 2024 developer conference, the smartphone maker showed off its new Siri complying with free-form requests, like, "Add this photo to the email I drafted to my [00:03:40] friend," and, "Show me my photos, uh, from Stacy, uh, i- in New York wearing her pink coat," and it will just do it.

[00:03:47] Marcus Johnson: Two years later, that promised software update, as it could be called, is expected to become a reality. Siri AI, as it's being coined, dubbed, uh, will include Google's Gemini AI model, uh, and cloud technology. Apple is [00:04:00] also releasing a dedicated Siri app to compete with chatbots like ChatGPT and Claude.

[00:04:03] Marcus Johnson: Grace, I'll start with you. What are the main reasons that Apple might be about to take over, or make some noise at least, in the AI race?

[00:04:10] Grace Harmon: Sure. I think that the two big things that are really working in Apple's favor are scale and trust. I mean, we're talking about billions of active Apple devices, like an almost [00:04:20] unimaginable distribution network for AI features.

[00:04:23] Grace Harmon: Um, it's spending more on R&D than ever before. And then, like you mentioned, that they're gonna be using foundational models developed in combination with Gemini, um, in addition to its own on-device processing. So I think they're kind of not trying to reinvent the wheel. They're trying to own this [00:04:40] user experience level on top of the best AI that is around and is available.

[00:04:44] Grace Harmon: Mm. Uh, but the scale of its devices is the biggest thing, just like I said, that distribution network is... I- it can't, can't make it otherwise. You can't make that up.

[00:04:51] Marcus Johnson: Yeah, it's an excellent point. And Gadjo, Harry McCracken of Fast Company was noting, "Siri doesn't have to be the smartest assistant on the planet.

[00:04:58] Marcus Johnson: It just has to be the one that's already put [00:05:00] in their hand, and Apple is about to put fully capable AI in reach of hundreds of millions of people, a lot who have never used or even opened, uh, ChatGPT." Um, what are your thoughts on Siri's advantages, um, in this race?

[00:05:14] Gadjo Sevilla: Yeah, absolutely. I think it's less of a takeover, but more like a [00:05:20] fork in the road.

[00:05:20] Gadjo Sevilla: Um, you know, people already use AI. They have their chosen brands, their chosen loyalties. They use things for image generation. They use another AI for helping with, with research. So Apple is not competing in that space. It's not a general free-for-all model. What it is doing, however, is [00:05:40] infusing its own brand of safe, and as we're seeing, highly personalized AI features designed around its users.

[00:05:48] Marcus Johnson: Mm-hmm.

[00:05:48] Gadjo Sevilla: So that is a small subset of the population, but it does build dependencies through, through use and, and defaults. Uh, especially, as you said, Marcus, for, for people who maybe are not [00:06:00] interested in AI as a general application. But if it shows up and makes, you know, your personal, uh, computing on your phone or your, or your devices a lot easier, then you're slowly converting-

[00:06:12] Marcus Johnson: Mm-hmm

[00:06:12] Gadjo Sevilla: uh, people who didn't think that they needed, uh, those features.

[00:06:16] Marcus Johnson: Yeah.

[00:06:17] Gadjo Sevilla: And I think- Maybe the end [00:06:20] game there is to convert some of those users into longtime Siri AI subscribers, right? Mm. Because that, that is Apple's model for all their services.

[00:06:30] Marcus Johnson: Yeah. What do we think of the features that they're gonna be rolling out?

[00:06:34] Marcus Johnson: Like, how much of an impact is this gonna have on consumers, on Apple users? [00:06:40] Because at large, other than of Tech Radar was saying, "Siri AI superpower is not that it's smarter and chattier, but it understands you through the data on your phone." Yeah. And there are a few examples of how it can help. Jennifer Jolly of USA Today saying, "You could ask, 'What podcast did my sister recommend?'

[00:06:54] Marcus Johnson: And it would pull the answer out of an old message thread. You can ask about a friend's camping email and have it list [00:07:00] those products for you to pack, and remind you to pack them at a certain time." And Eric Sullivan of Scientific American noting Apple's keynote, they'd said, uh, asked Siri to demonstrate the capabilities in terms of finding World Cup match schedule, planning a watch party, suggesting themed dishes from the countries playing, pulling out relevant information, um, from messages, and drafting a group invitation, all without opening any apps at [00:07:20] all.

[00:07:20] Marcus Johnson: Is this something you think, out of 10, how much of an impact do you think it's gonna have on consumers? Is this something that they're clamoring for, or is it something where, you know, they might use it a few times, but really it's a feature that just sits unused?

[00:07:30] Grace Harmon: I don't think it's gonna sit unused. I think, you know, there's a few things it depend on, depends on.

[00:07:34] Grace Harmon: It depends on how interested a consumer is in using AI at all, if they used Siri already to begin [00:07:40] with. Mm. If they're pretty accustomed already- Good point ... to just googling things. Gadget was saying that people already have, you know, their favorite LLMs chosen and picked out. They also have their favorite way of figuring out a dinner pa- party plan already probably.

[00:07:52] Grace Harmon: Yeah. Um, so I think for the most part, besides some of the deeper examples it gave, um, Siri just continues to be optimized for more quick device [00:08:00] tasks. It's not doing the deep, deep thinking that Claude and some of these other models are capable of. And to me, it felt like a lot of what it was showing at WWDC were things that rivals are already doing.

[00:08:10] Gadjo Sevilla: I think the, the key there is on device. A lot of what Siri AI is going to be able to do for its users securely will be on device. So the [00:08:20] personal context, awareness, understanding what's in their messages, what's in their photos, tying that all together with their calendar. Now, that's easy to do on device since it's a very controlled space.

[00:08:32] Gadjo Sevilla: But you also have to remember that it's not gonna be on all devices. I think there are hardware limitations which, [00:08:40] good for Apple, means they need, you know, it gives them a chance to, to sell newer devices that can run the whole gamut, right? Yeah. So I think i- in the beginning, you're, you're probably just gonna be able to use, uh, some of the very basic Siri AI functions.

[00:08:56] Grace Harmon: I think that, you know, like, that context awareness you mentioned is really big, just because [00:09:00] personalization is something that everyone, everyone who's into AI wants out of their AI experience. And like you said, if it's able to pull on not just what you've already said to it, but of, of your photos, of your emails, of your texts, I think that that is just a huge asset.

[00:09:13] Grace Harmon: That kind of builds into another thing that's kind of a growth engine for Apple, is its scale. Mm-hmm. And that's a scale of [00:09:20] devices in use, and that's also a scale of data information it has on its users.

[00:09:23] Gadjo Sevilla: Mm-hmm. And no talk of payments yet, right? I think-

[00:09:27] Grace Harmon: Mm ...

[00:09:27] Gadjo Sevilla: as people become more comfortable and, you know, more trusting, it's a matter of time before, uh, they, they put Apple Pay as part of the- Yeah

[00:09:36] Gadjo Sevilla: the, the stack. Because once they've earned the trust and they [00:09:40] already have those partnerships, uh, it, it makes sense to make Siri AI more agentic to proactively-

[00:09:48] Marcus Johnson: Mm-hmm ...

[00:09:48] Gadjo Sevilla: do things for, for consumers, like buying tickets or setting up reservations.

[00:09:53] Marcus Johnson: Yeah, Apple uniquely positioned as well to do that because of the privacy part of this.

[00:09:57] Marcus Johnson: Jennifer Jolly of USA Today noting [00:10:00] Apple's pitch is privacy. For Siri to be useful, it has to rummage through your most personal stuff, messages, mail, photos. Apple says it handles what it can right on your device, and anything sent to the cloud is used only to answer your request and then deleted. Do we think...

[00:10:12] Marcus Johnson: Gaz, you mentioned fork in the road, but is there another fork in the road moment here in terms of how these models respond to folks' requests? [00:10:20] Because, um, another, uh, I thought key point that came out of the, this announcement was that this model, this new Siri, uh, Siri AI, won't exhibit the kind of sycophantic behavior, the kind of sucking up, uh, behavior that other models have been accused of.

[00:10:33] Marcus Johnson: Uh, Mariella Moon of Engadget was noting, um, Apple, Apple's SVP of engineering, uh, saying the upgraded [00:10:40] Siri, uh, won't suck up to you and encourage users to reveal more about themself, and then use that information to establish a connection. Like other AIs, Siri has boundaries and just wants to help get things done.

[00:10:51] Marcus Johnson: Do you think that we're gonna lean one way or another in terms of how Siri responds to folks, or do you think there's a world where, um, both universes can [00:11:00] exist, one where the, the AI just gets things done, and one which is a bit more personal and talks to you like a person?

[00:11:04] Gadjo Sevilla: I, I don't think Apple wants to give customers an AI friend.

[00:11:08] Gadjo Sevilla: That's just a, a gray area I don't think they, they want to cross. But they are going for helpful assistant, um, proactive assistant. So basically a digital butler that, that can kind [00:11:20] of anticipate w- when, when to jump in.

[00:11:23] Grace Harmon: Yeah, I definitely think they're positioning it as more, like you said, a concierge- Yes, yeah

[00:11:27] Grace Harmon: and an actual assistant compared to an online companion.

[00:11:29] Marcus Johnson: My next question, Grace, I'll throw this at you first. Pis- uh, Peter Kafka of our sister company, Business Insider, uh, saying the new Siri is not meant to compete with the most sophisticated versions of AI engines like ChatGPT [00:11:40] or Claude. He was explaining that the term Sher- uh, this term Sherlocking, uh, saying that, um, it was when, uh, there was a search tool called Watson a long time ago.

[00:11:48] Marcus Johnson: Apple built a thing called Sherlock that destroyed it, and every year Apple does this with their operating system. They look around the ecosystem of cool third-party apps, and they just build it as a feature. And what they have done here is Sherlock, [00:12:00] he says, basically kill off, uh, in, uh, kind of, uh, for lack of a better term, uh, the free version of ChatGPT.

[00:12:06] Marcus Johnson: Fundamentally, the announcement adds up to why on earth would you use free ChatGPT when Siri is right here behind a button on your phone and can do all the same stuff? What, what's your take on Siri AI Sherlocking free ChatGPT?

[00:12:18] Grace Harmon: I think that's totally fair. [00:12:20] I think that, um, it depends on execution and if that actually pans out.

[00:12:25] Grace Harmon: Having something that you can carry around in your phone and just speak to is great. I think that free ChatGPT already kind of hinges on being able to do all of that minus the, the voice request, so part of that is just about form factor.

[00:12:39] Gadjo Sevilla: I think we're [00:12:40] also starting to realize, uh, that third party AI developers may have a tough time accessing Siri AI integrations because anything outside of Apple's control, including features from third party apps, that might take time to, to show up.

[00:12:55] Gadjo Sevilla: And so for if you're a developer, uh, or, or even another [00:13:00] AI company, that, that could discourage you from working with them from, from the get go, right?

[00:13:05] Marcus Johnson: Yeah.

[00:13:05] Grace Harmon: Mm-hmm.

[00:13:05] Gadjo Sevilla: So I, I think they're saving handful of features that they alone can control, and then maybe later on those integrations will, will start to be made available.

[00:13:16] Marcus Johnson: In terms of Apple working with others, John Rouech of [00:13:20] NPR noting that Apple is partnering with a major hardware rival to strengthen its AI efforts. In January, Apple and Google announced a multi-year deal to use Google's Gemini model as the foundation for Apple's AI features. And Harry McCracken of Fast Company, who I mentioned earlier, was questioning, is there some conflict of interest coming with Google's Gemini Spark personal AI [00:13:40] agent and Apple's new Siri AI?

[00:13:42] Marcus Johnson: Gadget, do you think the two can coexist?

[00:13:44] Gadjo Sevilla: I think they can. Uh, what, what Google offers is the cloud compute, which, you know, you, you need to run a successful AI even though it's on device. And Apple executives said that there's not a drop of, uh, Google in AI, in Siri AI, [00:14:00] meaning they might be using some of the model, um, development, but the actual product is, is really insulated, right?

[00:14:08] Gadjo Sevilla: So.

[00:14:08] Marcus Johnson: It's an interesting relationship. Gil Luria of investment company D.A. Davidson saying Siri will be built on Google tech, but Apple's going to own that relationship with the consumer. And they're paying, I think, about a billion dollars a year for this [00:14:20] at the moment, which seems quite low. Maybe Google will ask for more if this becomes popular.

[00:14:24] Marcus Johnson: Let's end by looking at what might not go right for New Siri. Grace, amongst the current slew of AI assistant competitors, what challenges is new Siri AI going to face?

[00:14:36] Grace Harmon: Well, I think for a long time Apple has been kind of over [00:14:40] promising, under delivering. Um, since most of its revenues are coming from hardware, uh, it just hasn't felt or been pressed by that same urgency that Google or Microsoft have to act and be spending really aggressively on AI.

[00:14:53] Grace Harmon: And I do think while, you know, it's s- starting to catch up a little bit with getting Siri out finally, um, that lack of [00:15:00] investment's probably gonna come back to bite it down the line on the consumer product side.

[00:15:03] Marcus Johnson: Is it fair to say Apple kind of brings everything together, or is planning to, but other companies might know you on a deeper level?

[00:15:09] Marcus Johnson: Because we have Mridul of USA Today writing, um, "Like a lot of people, I live on Apple hardware, but spend my days in other companies' apps, Gmail, WhatsApp, Google [00:15:20] Photos. Siri can read what's on my screen in any of them, but that's surface level. Until th- until those apps actually let Siri in, my real digital life has a big blind spot."

[00:15:28] Marcus Johnson: What do you think?

[00:15:29] Gadjo Sevilla: Yeah, I think that's, that's accurate. Um, and from what we've seen from the early, uh, demos, Siri AI takes its time. It's not instantaneous. So it gives you the [00:15:40] sense that it's still very much in development, and we're not gonna see that out of beta till September. But for now, yeah, you do get the sense that it's gonna operate at its own speed, um, before it, it can sort of take over operating system functions to, to, to the extent that they're, they're hoping.

[00:15:57] Gadjo Sevilla: The delay has actually helped them because [00:16:00] people now kind of know how, how they like their AI and how they use it.

[00:16:04] Grace Harmon: Mm-hmm.

[00:16:04] Gadjo Sevilla: And how they use those models, and they can even segment, uh, across different products of what works best for them.

[00:16:11] Grace Harmon: That's a great point. That's kind of the little bit of the advantage of, you know, accidentally making a credibility of marketing the future before they'd accidentally built it, is they got to see what the future [00:16:20] is gonna hold.

[00:16:21] Gadjo Sevilla: Yeah, and they didn't have to spend billions or trillions- ... to, to get there, right? Just by chance, I suppose.

[00:16:27] Marcus Johnson: I'll close with this, Peter Kafka of, of BI saying the industry is mounting the threat, and Apple needs to respond because if you're not using Siri, you might use ChatGPT. And if you use ChatGPT enough or Gemini [00:16:40] enough, then when they have the hardware, they can say, "You're not even using the rest of this phone.

[00:16:45] Marcus Johnson: Buy a point-and-shoot camera and our little pendant, and you'll be set. And you won't have social media to distract your kids." Uh, so maybe that's the play. We'll see as we- Yeah,

[00:16:56] Grace Harmon: but an O- Please ... an OpenAI, I'm gonna say an OpenAI- Please ... standalone device isn't gonna [00:17:00] unseat the iPhone. It's not.

[00:17:02] Marcus Johnson: Yeah.

[00:17:02] Grace Harmon: It's, people aren't going to throw out their iPhones in favor, or at least en masse.

[00:17:06] Grace Harmon: Like, people aren't gonna throw out their iPhones because of... I mean, look at the friend pendant

[00:17:10] Gadjo Sevilla: Yeah.

[00:17:11] Marcus Johnson: Do we think this is why Apple has said we don't need to spend so much on AI, and they're not nervous about being behind in this race because they know that [00:17:20] if they control the hardware, similar to like, um, content is king, but the ace is the pipes that provide that content.

[00:17:27] Marcus Johnson: Um, they're saying that we have the, we have the portal into tho- tho- those AI services.

[00:17:33] Grace Harmon: I think that's a good point. I think that, like I said before, like it hasn't had to be under as much, uh, extreme pressure [00:17:40] financially to like get its- Mm-hmm ... get its margins up or like- Yeah ... have a supplementary business.

[00:17:44] Grace Harmon: Also, I think that, you know- Right ... when it went into some issues around like hardware and manufacturing amid the height of the tariff craziness last year, like it was leaning more on services. It has its own digital streams and digital services that aren't AI related to lean on for extra revenue and [00:18:00] for, uh, more customer support.

[00:18:01] Marcus Johnson: That's all we've got time for for this episode. Thank you, gang, for hanging out with me today. Thank you to Gadjo.

[00:18:05] Gadjo Sevilla: Thank you very much.

[00:18:07] Marcus Johnson: And to Grace.

[00:18:07] Grace Harmon: Thank you.

[00:18:08] Marcus Johnson: And to the Knicks and the Spurs for a great NBA finals. And thank you to the production crew, Lance helping us out with this one, and to everyone for listening to Behind the Numbers: an EMARKETER podcast, made possible by Rokt.

[00:18:16] Marcus Johnson: Tune in Wednesday, where Arielle will be guest hosting this month's unofficial monthly [00:18:20] retailer awards on the Reimagining Retail show, and I'll see you on Friday. I hope.