The AI-native growth team: what it looks like in practice
“AI-native” has become one of those phrases everyone uses but nobody really defines. Over the past year I’ve had more than 80 conversations with app growth leaders while building the Appsforum agenda, across UA, CRM, ASO, product, monetisation and leadership. AI came up in more than half of them, and by the autumn it was coming up in most, often before I’d even asked!
We all know AI is being used everywhere now, so that wasn’t necessarily the interesting part. What stood out was how teams are using it, and how many leaders are focused on making their teams AI-native before they fall behind.
Here’s what an AI-native growth team looks like right now, based on what they told me:
1. Output grows, headcount doesn’t
The most consistent change is that teams are producing far more without hiring to match. This was clearest in creative, where UA leads described scaling creative testing and production volume with the same small team, and smaller startups described AI as the way one or two people can cover an entire growth function. That’s the first real marker of an AI-native team: it isn’t about everyone having a ChatGPT login, it’s that the team’s output no longer tracks its headcount.
2. It has moved well beyond creative
A year ago most of these conversations were about ad production, but that’s no longer the case. Teams described using AI to:
run CRM operations, using MCPs to connect campaign building to live data
automate reporting so decisions get made faster
support modelling, attribution and decisions about which markets to enter
iterate on paywalls without waiting for an app release
support ASO through what one person called “controlled automation”
In other words, growth teams aren’t just making ads faster. They’re taking manual work out of every part of the funnel.
3. At the leading edge, people supervise agents
A few teams are further ahead. One described a two-layer agent setup, where the first set of agents turns growth playbooks and market research into product specs, and the second set builds them, leaving developers to review and validate the output rather than write the code themselves. Others described building a growth “operating system” of small agents that run parts of their marketing.
Most teams aren’t at this stage yet, and even the more advanced ones described parts of their agent work as “rudimentary”. Still, the direction is clear: the job is moving from doing the work to directing it.
4. Humans stay where trust and taste matter
The teams furthest along were also the clearest about where AI shouldn’t take the lead. In one recent paid social test, real UGC voiceovers outperformed AI-generated ones. Others stressed that ASO still needs human analysis and direction, and that a badly trained support bot can damage trust very quickly. As production gets cheaper, several people said original ideas and quality control matter more than ever, because the bottleneck has shifted from making things to deciding what’s actually worth making.
5. The messy middle is real
Not every team is a neat AI-native success story, and many were open about the friction in between:
Too many tools: a growing pile of AI tools, changing faster than workflows can settle around them.
Teams out of step: engineering and creative teams using different tools and speaking different languages.
Uneven speed within one company: newer product teams already working in AI-native ways while longer-standing teams take months or even years to catch up.
Stuck at the start: using ChatGPT every day, but with very little real automation built into workflows.
For a lot of companies, becoming AI-native turns out to be less about the tools and more about change management.
6. The catch: everyone else has it too
There was one more theme, and it’s a less comfortable one. Because AI makes building apps so much easier, the market is filling up with AI-enabled competitors. Several leaders said the hard part is no longer building the product, but standing out, getting distribution and being recommended, which increasingly includes being recommended by AI assistants themselves. So an AI-native team doesn’t just use AI to move faster; it also plans for a market where everyone else can move faster too.
What this means
Across all of these conversations, being AI-native came down to four things:
output that’s decoupled from headcount
automation across the whole funnel, not just creative
people moving from doing the work to directing and checking it
deliberate human judgement wherever brand and trust are at stake
None of the teams I spoke to claimed to have it fully figured out. If anything, the ones furthest ahead were simply the most honest about what’s still messy!





Comments