The last decade of growth was about execution speed — who could produce creative faster, scale spend wider, hire more optimizers. The next decade will be about decision efficiency — who can extract judgment from signals, turn judgment into action, and compound each action into the starting point for the next.
On May 20, Meta laid off 8,000 employees — primarily from engineering and product teams — while simultaneously reassigning 7,000 people to newly formed AI teams. This wasn't a cost cut. It was an organizational hard restart — and the problem it exposed points directly at every growth team.
But Meta isn't an anomaly. According to the WRITER 2026 Enterprise AI Report, 79% of enterprises report serious challenges in AI adoption, up double digits from 2025. Gartner puts it more bluntly: only 1 in 50 AI investments produces transformative value.
Tools were purchased. Budgets were spent. 79% say they're stuck. The problem isn't the tools — it's that growth organizations are still running 2019 decision processes with 2026 technology.
Ⅰ. What Meta got right: they changed the org, not the toolstack
The 8,000 roles weren't cut for performance reasons. Meta concluded that its legacy team structure couldn't support AI-native workflows (Bloomberg, May 2026). The 7,000 reassigned employees entered new AI-centric "pod" structures — not adding AI tools to old teams, but rebuilding teams around AI capabilities (NPR, May 20 2026).
This is the difference between "using AI" and "becoming AI-native."
Using AI means adding an AI assistant to your existing workflow. Humans still make every decision. Meetings still happen. Approval chains still run sequentially.
Becoming AI-native means restructuring how the team works around AI's decision-making capability. AI doesn't just execute — it participates in judgment. The human role shifts from "making decisions with incomplete information" to "calibrating decisions with structured information."
Anthropic's recently published Founder's Playbook describes this shift precisely: in AI-native organizations, the human role moves from individual execution to orchestrating AI systems. The focus shifts from "how to do it" to "what to do and why." While written for startups, the same principle applies to growth teams: the growth lead's job is no longer to personally approve every creative and bid — it's to design decision rules, calibrate AI judgment, and handle exceptions the system can't cover.
Ⅱ. Three types of growth teams, three outcomes
Growth teams in 2026 are splitting into three distinct forms:
Type 1: Tool stacking.
Multiple AI tools purchased — creative generation, bidding optimization, analytics — but none connected to each other. Each tool is an island. Humans bridge the gaps manually. Weekly meetings persist. Excel remains the system of record. This is the majority today. Diagnostic: your team uses 3+ AI tools, but no tool's output automatically feeds into another tool's input.
Type 2: Partial automation.
The creative-to-launch pipeline is AI-connected, and speed has improved. But strategy decisions and performance reviews still rely entirely on humans. When entering a new market or facing unexpected changes, the team reverts to "let's get in a room and discuss." Diagnostic: at least one end-to-end pipeline is automated, but strategy input upstream and review processes downstream are still manual.
Type 3: Decision system.
AI participates in the full loop — signal collection, strategy generation, execution, and review. Each campaign's data automatically becomes the input for the next cycle's decisions. Humans set rules, calibrate judgment, and handle anomalies. Growth compounds through system accumulation, not individual experience. Diagnostic: review reports generate automatically, and last week's effective signals appear in this week's strategy recommendations without anyone manually pulling data.
Meta jumped from Type 1 to Type 3 through layoffs and restructuring. Most companies don't need to be that aggressive — but the direction is the same.
Ⅲ. What AI-native growth actually looks like
Consider a brand's growth team expanding into Brazil.
Before AI transformation:
Monday, the growth lead and media buyer discuss Brazil strategy. No local data — they use Southeast Asia as an analogy. Two hours, no conclusion.
Tuesday, the creative team produces three directions based on gut feel.
Wednesday, the lead is traveling and doesn't respond — creative stalls.
Thursday, campaigns go live using US market parameters.
Friday, CPM looks high but no one knows why. "Review next week."
Next Monday, the review gets pushed by new priorities. Brazil data sinks into Excel, never to surface again.
After AI transformation:
The growth lead still makes decisions. But the inputs change.
Monday, market signals are already collected and structured — competitor spend intensity, local audience preferences, channel cost benchmarks. The lead spends 20 minutes confirming direction, not two hours debating.
Tuesday, AI generates three creative sets and campaign configurations. The lead reviews and adjusts — decision authority stays with the human, but AI did 80% of the preparation.
Wednesday through Friday, performance data flows back continuously, with the system flagging anomalies and opportunities.
Friday, a review report generates automatically, with effective audience-creative combinations locked in as next week's starting point.
Meetings go from two hours to twenty minutes. The human still makes the call — but the information density and decision quality in front of them are fundamentally different.
Ⅳ. Start with the slowest decision, not the slowest task
Not every company needs to restructure 8,000 roles overnight. The starting point for AI transformation doesn't require org changes, major procurement, or new hires.
It requires one thing: identify the highest-latency decision point in your growth team. Not the slowest execution task — the decision that always waits for someone to make a call based on incomplete information. It might be budget allocation, market selection, or creative direction.
Then introduce AI judgment capability at that single point. Not to replace humans — to let AI propose, humans confirm or override. Four weeks. Look at the data.
The endgame isn't "fewer people." It's "people doing fundamentally different work." Growth teams evolve from execution machines to calibrators of decision systems.
The last decade competed on execution efficiency. The next decade competes on decision efficiency. Meta chose the most aggressive path — tearing down the old org and rebuilding around AI. Most companies don't need to go that far, but the direction is the same: AI tools aren't scarce. What's scarce is a serious organizational upgrade for how growth decisions get made.
If you're rethinking how your growth team makes decisions, we'd welcome a conversation. → get.lanbow.com
