From magic to muscle: Qualcomm's Brent Summers on why people — not models — drive enterprise AI transformation

In this episode of Humans of AI, hosts Alaura Weaver and Diego Lomanto sit down with Brent Summers, head of AI platforms and GTM innovation at Qualcomm. Brent shares his journey from a UX-obsessed kid on dial-up bulletin boards to leading AI transformation at one of the world's great engineering companies.

We explore what it means to be a "human marketer" in an organization that bends physics, why AI needs to transition from "magic to muscle" through daily practice, and why the most important part of any AI transformation is the people in the middle. Plus, Brent offers a bold contrarian take — the biggest risk of AI isn't that it makes us less human. It's that we use it as an excuse to avoid the hard, uniquely human work of building relationships.

Subscribe to Humans of AI for more stories from people navigating the intersection of business and artificial intelligence.

Watch the full video interview on the WRITER YouTube channel for bonus content and deeper insights.

Learn more about WRITER at writer.com. 

Transcript

SPEAKER_01 0:08

It's not that AI makes us less human, but that we use AI as an excuse to not do the hard and human work in the first place.

SPEAKER_02 0:18

That's Brent Summers, head of AI platforms and GTM Innovation at Qualcomm, one of the world's great engineering companies. He said it as the person leading Qualcomm's AI transformation from the inside. The first time I heard it, I almost let it go. It sounds simple, but it's actually an indictment of everyone who has reached for the AI tool when what the moment really called for was a conversation. Brent has spent his entire career asking one question, from a middle school chat room to the boardroom. What does the person on the other side of this actually need? Every chapter of his story arrives at the same answer. Technology is only as meaningful as the connection it creates between people. I'm Alora Weaver. Welcome to Humans of AI. When ChatGPT launched in late 2022, Brent had already been using generative AI for his own writing since the pandemic. Then when he was hired by Qualcomm, he hit a record scratch moment.

SPEAKER_01 1:42

And then when I joined, my personal software license was not permitted. And I just sort of accepted that and went on to initiate the ABM program that I was hired to build.

SPEAKER_02 1:52

Day one, stopped at the door by a software policy. He put the tool down. His personal tool, the one he'd been building fluency with for two years, and kept moving. Built the business case, found the right partners, got the approved tools into the room, watched the sales team compress hours of account research into minutes, watched their eyes go wide. That's what it actually looks like to lead an AI transformation inside a large organization. You arrive, adapt to the constraints, earn the trust, build the infrastructure, and then once it's working, you meet the next problem. Let's start before the job title with a basement in Daytona Beach, a 486 PC, and a modem dialing into the void.

SPEAKER_01 2:46

The first thing I did was dial up to a local bulletin board system. The paper in Daytona Beach had a BBS, uh, and there I was in middle school online chatting with strangers. I was in chat rooms, I was playing text-based uh RPGs, just having all kinds of fun building community.

SPEAKER_02 3:02

To Brent, at 12, a modem was a portal to other people. The technology was beside the point. That instinct, ask who's on the other side before asking what the tool can do runs through everything that follows.

SPEAKER_00 3:15

You've described yourself as not a product marketer, but a human marketer. And you know, you're doing that inside one of the world's great engineering companies. So can you unpack that?

SPEAKER_01 3:27

So I'm always thinking about what's the human need? Um, what is causing people to hesitate or what kind of things do they are they trying to avoid? Um, what are the the handoffs with technology? That's an interesting tension within Qualcomm, uh, because we're often like trying to bend physics to our will. And you know, I'm going, why do we need that?

SPEAKER_02 3:53

Why do we need that? In an engineering culture built to answer can we do this? Brent asks, should we, and for whom? It's the question most of us skip when we're moving fast and telling ourselves we'll optimize for the human experience later.

SPEAKER_01 4:08

I think at the application layer, there are all kinds of questions around what is someone uh interested in, willing to adopt, uh, and that is wholly separate from you know just pushing the frontier forward.

SPEAKER_02 4:24

Growing up, Brent was captivated by a short-lived piece of software called Microsoft Bob. One of Microsoft's most spectacular failures. Instead of a traditional desktop, it gave you a virtual house to open a door to reach your documents, turn on a lamp, to change settings. The bet was that a home would feel more intuitive than a file system. Brent was one of the few people who actually loved it.

SPEAKER_01 4:54

Bob really reimagined the way that you would interact with the desktop, and all I was doing was tinkering with my PC. I had a lot of free time on my hands, and I was having fun, you know, turning on the light switch and petting the dog. That felt like a game to me as a kid, right? So I was doing productivity tasks in a way that felt very approachable for me as an adolescent. You know, my parents who were learning to use the computer in their 30s, and my dad is much older, my dad would have been like 50 at the time. That reinterpretation of the way that human-computer interaction should work was too much too fast, I think, for a lot of folks.

SPEAKER_02 5:32

Brent's parents couldn't meet the technology where it was. The design didn't reach back toward them. That gap between a capable technology and the people it was built for is the question he's been carrying ever since. And it's the same one he's asking now about enterprise AI. In most organizations, AI arrives the same way Bob did, dropped on top of existing workflows, with the implicit expectation that people will adapt. Brent's instinct runs the other way. You design the technology to meet people where they are. That's the difference between pilots that stall and transformations that compound. We've been here before. The PC didn't just change how offices worked, it required a new kind of literacy. So did the Internet. The organizations that thrived weren't the ones with better tools, they were the ones that helped their people develop a new relationship with how work gets done. AI is that kind of shift, a new language, not a faster version of the old one. And fluency doesn't come from onboarding, it comes from the willingness to feel incompetent for a while before you feel capable. Good design lowers the barrier. It gets people willing to try. Brent built that belief into how QualClimb's transformation is framed at the top.

SPEAKER_01 7:01

As an organization, we're committed to something that Don McGuire, our CMO, is calling a people-led AI-powered transformation.

SPEAKER_02 7:09

And most AI transformations, including many we tell ourselves we're running, the sequences get the tools, then figure out the people. What Don McGuire is naming and what Brent has been building is the opposite. The human stays at the center of every decision. Who delegates to AI, who collaborates with it, who sets it aside. That's a structural claim. And it points directly at what gets lost when the AI leads. Judgment, taste, and the courage to take creative risks.

SPEAKER_01 7:39

Which is what Brent means when he says The difference between a bold idea and a changed organization is the humans in the middle.

SPEAKER_02 7:48

So what does becoming fluent actually look like? Not in theory, but in the workflow on day three when the magic wears off?

SPEAKER_01 7:57

AI definitely feels like magic to people. When you first enter a prompt and AI does research for you or generates a reasonable first draft, that feels really magical. For it to really be productivity gains or to help transform an organization, there are reps and sets required. You have to teach people right prompting techniques.

SPEAKER_00 8:39

I think it's all of the things that we teach people in terms of how to break a habit and form a new habit, right? You have to have the reps in before it becomes natural because people are so trained to just do things the way they've always done it.

SPEAKER_02 8:59

Magic to muscle. Magic requires the trick to work every time. When it doesn't, you lose the audience. Muscle survives the bad sessions, the days when nothing clicks, when the prompt returns garbage, when the team looks at you and says, This isn't working. The difference is whether you're willing to be bad at something long enough to get good at it. And that problem isn't the technology. It's the organizational immune system operating not just at the policy level, but at the human level too. In the person who had one bad experience with the tool and quietly went back to doing things the old way.

SPEAKER_01 9:38

There are still a few skeptics. There are people who, you know, if they have uh a negative interaction, then they have an aversion. So we're managing that. And I I think that's that's the play for for a while longer.

SPEAKER_02 9:53

Every AI transformation case study talks about the tools, the ROI, the wins. Almost none talk about the person who had a bad first experience and quietly went back to the old way. That person is on your team, maybe on mine. And our job is to earn enough trust that they try again. Not to overwhelm them with productivity stats, but to stay in relationship long enough that the next attempt lands differently. The pilot Brent eventually ran was not large. Twelve users, 16 use cases, a few weeks. His VP was skeptical. This was well outside the lane of someone hired to run ABM. He ran it anyway. Every single person in that pilot group saw the value and wanted to keep their license. Three and a half years later, every person on the Qualcomm marketing team uses generative AI every week. A hundred percent. Not because it was mandated, because the design worked. Because the tools were introduced in a way that met people where they were, because someone had spent years earning the trust to make that possible. That's the outcome of human-led AI-powered. And Brenn is proud of every person who got there. What he'll tell you though is that the pride isn't in the adoption numbers, it's in what those numbers represent. People learning, adapting, and doing their best work.

SPEAKER_00 11:25

If we're talking to the CMOs and they're still in experimentation mode, and they're sitting there with pilots and use cases, they haven't built an infrastructure, what would your advice to them be? What should they be doing differently?

SPEAKER_01 11:39

You need to nominate someone to own the business case and the outcome. If you don't assign a responsible individual, uh that initiative is not likely to truly be propelled forward if it's managed by a committee. Uh if you're three years into experimentation and not seeing meaningful results, then maybe you need to examine your assumptions uh about the way that work in your organization has to get done. Because I think AI has more value when it's designed into a workflow as opposed to simply like bolted on top.

SPEAKER_02 12:20

Someone who believes AI will reshape the nature of work and is building the infrastructure to make that happen still has to go home and sit with the question: what does this mean for the people I love working with?

SPEAKER_01 12:33

The stuff is getting crazy good, right? Uh like it is certainly more advanced than I thought it would be when I became the de facto champion for AI use inside of our organization. And so that is both significant opportunity and there's some risk there. So I have a lot of anxiety about making the right uh judgment calls and guiding people and practices towards an outcome that keeps our head count where we're at, uh, keeps the people who I love working with employed, and also keeps people fulfilled. You know, I get a lot of personal satisfaction from working, it's a big part of my identity. And, you know, there's some people espousing that the need for human work is gonna go away. And that for me would I think life without work would be a bit drab and dull.

SPEAKER_02 13:29

He's the AI champion and he's sitting with the fear that he might be building something that hurts the people he's created a solution for.

SPEAKER_00 13:38

We have to remember that those of us who are at the edge here, you know, are in impacting and influencing what happens next for people, right? Um I appreciate that you have that empathy and that thoughtfulness in regards to what it means.

SPEAKER_02 13:55

How does someone who carries both of those things keep going? I think the answer is in how Brent defines what we're actually building toward.

SPEAKER_00 14:04

What is one piece of marketing advice that you used to swear by that AI has completely changed your mind on?

SPEAKER_01 14:13

Content as the most important unit of output for a marketer. I think you know, signal is more important than volume right now.

SPEAKER_00 14:21

Yeah. Signal and differentiation.

SPEAKER_02 14:24

Large language models are by design averaging machines. They produce the most statistically probable response to any prompt, the consensus answer, the thing that sounds right because it sounds like everything else. Divergent thinking works differently. It's the capacity to break from pattern, to arrive at the idea no one was expecting because it comes from a perspective, a set of lived experiences, a point of view no model was ever trained on. That gap between averaging and diverging is where human marketers now live. It's the entire value proposition, which made Diego's final question the right one to end on.

SPEAKER_00 15:08

How do we stay human while growing a business in the age of AI?

SPEAKER_01 15:13

Well, I'm gonna push back on the framing. There's an assumption there uh worth examining, right, that we need to stay human. That implies that our default trajectory with AI is gonna pull us away from humanity. And yeah, I guess if we're not careful, uh vigilant, intentional, that we might drift. But I think the opposite risk is uh just as common uh and maybe more dangerous. And it's not that AI makes us less human, but that we use AI as an excuse to not do the hard and human work in the first place.

SPEAKER_02 15:49

His warning carries weight precisely because he believes in what he built. Don't let the tools become a way to avoid the conversations that need to happen.

SPEAKER_01 16:00

We have to have difficult conversations with each other, right? We have to talk to our customers to understand what their pains are, sitting with, contemplating, simmering. And I think a lot of that is human interaction. You know, we focus a lot on human-computer interaction, but human-to-human interaction is just as important. We're not in show business, like my agent shouldn't call your agent in every situation. Sometimes the humans should talk to each other.

SPEAKER_02 16:27

Stay human is a daily practice, not a posture. Choose the conversation, choose the call, choose the thing the tool can't do for you, because the moment requires you. That's what Brent has been building toward his entire career. And that's what he's asking the rest of us to build toward now. From a 12-year-old on a bulletin board in Daytona Beach to the person leading one of the most ambitious AI transformations in enterprise tech, every chapter of Brent's story has been a version of the same question: are we using this tool to get closer to each other or to avoid each other? The answer is the only thing that determines whether we're building towards something worth having. So, here's your Monday morning action. Think about the last five times you used AI. For each one, ask yourself: did this help me create a deeper understanding and connection with someone, or did it put distance between us? Some will pass. The research that freed you up for a deeper customer conversation, the summary that got your team aligned faster so you could spend the meeting on the hard questions. Some won't. The re-engagement email that went out when a call was what was needed. The brief that got sent when the team needed a conversation about what we're actually trying to say. The tool didn't fail. The intention did. And the only way to close that gap is to keep asking the question. Technology is only as meaningful as the connection it creates between people. I'm Alora Weaver, this is Humans of AI, and we'll see you next time. Thanks so much to Brent Summers for sharing his story. Humans of AI is presented by Writer. Our hosts are me, Alora Weaver, Director of Enterprise Content, and Diego Lamanto, Chief Marketing Officer. Our producer is Alana Nevins, our graphic designer is Mila Odor. Our social media manager is Tasha Thakar. Our video editor is Jamie Watkins. If this episode got you thinking about things in a new way, please share it with a friend, leave a review, and subscribe to hear more stories of humans working at the crossroads of technology and business.