External signal·Andreessen Horowitz (a16z)·Jan 8, 2026·Anish Acharya·6 min read
Notes on AI Apps in 2026
“The first is that every team + every task (marketing, legal, procurement, finance) should be software first, and all of these leaders are going to have to learn to reach for a software toolbox before the process / human systems they've traditionally relied upon.”
Summary
a16z general partner Anish Acharya argues that cheap AI-generated code has not yet diffused across enterprises, and that the binding constraint is shifting from 'how do I build it' to 'what do I build.' He predicts a new category of 'thinking' or exploration tools (citing Cursor and Google's Antigravity) to complement execution tools like IDEs and Figma. His central organizational claim is that coding agents will let software absorb traditional 'service' functions (legal, finance, HR, procurement) that were historically human-capital-heavy, so 'every team should be a software team,' using either domain products like Harvey or bare-metal agents like Codex and Claude Code. He contends AI-native apps will diverge from, not be subsumed by, foundation models, pointing to over $1 billion in new coding-startup revenue generated in 2025.
Predictions for the future of work
Acharya expects every enterprise function, including non-technical 'service' roles in legal, finance, HR, and procurement, to become 'software first,' reaching for coding agents before human or process-based systems. He envisions a near-future product manager who sets broad goals and reviews features that AI built and A/B-tested overnight, and predicts incumbent CEOs will collapse customer-facing roles (sales, support, collections) into single broad-goal functions. The timeframe is immediate (2026), the affected population is knowledge and service workers, and he frames the shift as creating broad operating leverage rather than mass displacement.
Originally published by Andreessen Horowitz (a16z) · Jan 8, 2026
Read the original at Andreessen Horowitz (a16z)