Derya Matras, Meta’s VP for the EMEA region, poses a question she keeps hearing from business leaders: If you started from scratch knowing what AI can do today, how would you design your business? For most companies, she argues, the honest answer exposes a mindset gap, not a technology gap.
The core tension she identifies is organizational. CFOs want incremental efficiency. CEOs see a bigger strategic bet. CMOs are squeezed between long-term brand building and short-term output targets. All three are looking at the same technology and reaching different conclusions about what it is for.
Why agentic AI is a different category
Matras draws a sharp line between AI as an optimization layer and agentic AI as something structurally different. Three properties mark the distinction: proactivity, autonomous action, and multi-modal intelligence. An agent reasons and plans across multi-step strategies rather than responding to a single prompt. It stops functioning like a search engine and starts functioning like a colleague.
Her concern is that businesses treating this as a minor upgrade are leaving real opportunity on the table while AI-native competitors build from a blank slate and move faster.

The Movida number worth noting
The most concrete data point in the piece comes from Movida, one of Brazil’s largest car rental companies. Using Meta’s Business Agent, they moved their entire booking flow, from vehicle selection through pricing and payment, into a single WhatsApp conversation.
- 44% increase in daily bookings within one month
- 85% of conversations resolved entirely by the AI agent
That is the kind of before-and-after that makes the abstract argument real. An agent that handles the full transaction loop, available at any hour, removes friction at every step a human would otherwise introduce.
The equalization argument
Matras makes a specific claim worth taking seriously if you run a small operation: agentic AI does not care about the size of your brand. More than one million businesses are already using AI agents on WhatsApp and Messenger. The targeting depth, personalization capability, and market responsiveness that used to require large teams and large budgets are now accessible to a solo founder in the same way they are to a global enterprise.
Her prediction: single-person billion-dollar companies become possible because of these tools. That is her framing, not a verified data point, but the structural logic holds. The cost of scale is dropping.
The operator takeaway
The piece is a Meta VP writing in Fortune, so the Meta products are front and center. Read past that and the underlying argument is sound: using AI to run yesterday’s processes slightly faster is an expensive treadmill. The businesses building measurement and optimization disciplines around agentic output now will have a structural head start over those still running pilots.
