Why companies are becoming a series of loops | Anish Acharya (a16z)

Founder takeaway

Treat AI adoption as redesigning recurring workflows, not merely automating isolated tasks. Map each function as a loop—input, decision, action, feedback—and identify where AI can shorten the cycle or improve its output.

Why it matters

Companies gain durable leverage when learning and execution compound across repeated cycles. Improving an entire loop can increase speed and consistency more than optimizing a single step.

Relevant guides

Put it to work

Choose one recurring workflow this week. What are its input, decision, action, and feedback stages—and which stage could AI improve without weakening accountability?