Andres Lam joined Abnormal as an AI product builder with a clear sense of what the job would involve: complex technical work, a steep learning curve, and a lot of building from scratch. What he didn’t anticipate was how much of the most meaningful work would turn out to be about people, not infrastructure.
Building the Marketing Engine
Andres came to Abnormal from big tech with a conviction that AI would fundamentally change how companies operate. At Abnormal, he found a place already acting on it. His primary focus is building what he calls a campaign factory, a system that runs marketing campaigns with AI from brief to published asset. A campaign brings together creative, copy, and budget across channels: a blog post, an ad, an email. In a campaign factory, an internal agent handles the execution, choosing the audience, generating the creative, placing the budget, and publishing across channels.
But the system doesn’t run without the people behind it. Someone still writes the brief, sets the strategy, and makes the judgment calls about what a campaign needs to achieve. What the campaign factory changes is where human effort goes: away from the operational work AI can handle and toward the decisions that actually require expertise.
“We’re changing how we build and bring software products to market with AI,” Andres said.
The Campaign That Built Itself
The first fully AI-generated ad set was built for a campaign about hiring fraud conducted by North Korea. To make it work, the team had to do two things in parallel: build the technical infrastructure and shift how the marketing organization operated.
On the technical side, that meant building connectors to ad platforms, internal knowledge systems, and sales platforms so AI could generate on-brand, performant creative from a single brief. On the people side, it meant moving away from external agencies and creative teams and bringing that competency in-house.
“We had to up-level and up-skill our team at the same time we were building out the tech,” Andres said.
That experiment also surfaced something Andres hadn’t expected: how good image generation models had become. He’d used language models extensively before joining Abnormal but hadn’t leaned into image generation as a tool. Seeing what the models could now handle, including logo generation, text, handwriting, and human faces, changed how he thought about what was possible for creative at scale. He’d predicted the models would keep improving, and seeing it actually happen was something else.
“Even if you predicted that the models were going to get better, when it actually does get better and those images are better and more realistic, it’s just super exciting. And what that allows us to do as a business for ads, for blog posts, for any generated creative is truly amazing.”
The Part He Didn’t See Coming
Andres knew going in that this role would be challenging and hands-on. What he didn’t expect was how much he would get from the people side of transformation.
“The most exciting part of my job is actually working with other people in this company and enabling them to build AI tools or enabling them to rethink how they want to do their job with AI,” he said.
As a systems thinker focused on enacting change at scale, Andres says the motivation comes directly from the colleagues he works with. Seeing what teammates in marketing and sales had built independently of his team’s tooling reinforced something important about how Abnormal operates: transformation here is decentralized. It doesn’t require a centralized mandate or a single team holding the keys. People across the organization are building, writing prompts, and experimenting on their own. That energy, he says, has been as important to the work as anything his team has shipped.
“Their excitement to build with us, to write prompts and skills with us, has been very motivating and instrumental to the success of the various transformation projects we’ve undertaken so far.”
What Only Abnormal Makes Possible
Finding a place where the conditions were right to actually do this work was just as important to Andres as finding the right role. “AI transformation fails without focus and support from the entire leadership team of a company,” he said. “You need support from the people team just as much as you need support from your technology leaders or your product leaders.” Watching executives operate cohesively and with genuine commitment to the work gave him confidence that what he was building would actually land.
He also found it in how the company measures contribution. Watching a leader on his team build out the AI GTM function showed him something important about how Abnormal operates: good ideas get backed here regardless of where they come from. “At Abnormal, you’re only bounded by your own ambition,” he said. “Regardless of seniority, if you have a good idea here, you can take it and run with it.”
The experiment he’s most excited to run next is giving AI control over an increasingly large marketing budget, letting it make economically consequential decisions about how spend is allocated across channels. For Andres, that’s not just a performance marketing optimization. It’s a step toward what he describes as a self-improving enterprise.
“At Abnormal, you’re only bound by your own ambition.”
Building Toward Something That Lasts
When asked what he’d tell someone considering Abnormal who wants to do real AI work, Andres doesn’t pitch the role or the team. He talks about the broader moment the industry is in. Cyber attackers are using AI to run more sophisticated operations with less sophistication required on their end. AI has changed knowledge work. The gap between people who are building with it and people who aren’t is widening.
“You want to join a place that is AI-native and has a very clear thesis on how humans and AI will work together at that company,” he said. “I think Abnormal has a very mature stance on this.”
For Andres, the specifics of what he builds will keep evolving and the craft will shift, but that’s not a concern. “You still want to push yourself to be excellent and work with excellent people,” he said. “And you can find that criteria here at Abnormal.”
If you want to do real AI work at a company that’s already living it, explore open roles at abnormal.ai/careers.