Keep pulling the thread on Eleven Labs.
Marc Andrusco predicts that by 2026, the prompt box will no longer be the primary user interface for AI applications.
AI applications are expanding the total addressable market for software by approximately 30 times, shifting focus from the $300-$400 billion annual software spend to the $13 trillion U.S. labor spend.
Stephanie Zane predicts that by 2026, the primary design paradigm for content and applications will shift from optimizing for human users to optimizing for AI agents.
In certain geographies, human call center labor remains cheaper than best-in-class AI voice agents, but this cost advantage is expected to erode as AI model costs decrease.
Olivia Moore believes the AI voice agent market is large enough to be considered an "industry," which will support successful companies at every layer of the technology stack.
The next generation of AI applications will operate proactively by observing user actions and suggesting interventions, rather than relying on explicit user prompts.
Marc Andrusco envisions a future "AI native CRM" that will proactively analyze a salesperson's historical data, such as emails from the last two years, to identify and resurface dormant sales leads.
AI Site Reliability Engineers (AISREs) are changing incident response by automatically analyzing telemetry data and delivering hypotheses directly into platforms like Slack, replacing manual dashboard analysis.
A16Z portfolio company Decagon is currently using AI to autonomously answer customer questions.
In high-liability fields such as security operations and incident resolution, AI agents are primarily used to assist humans, with humans remaining in the loop for final decisions.
Olivia Moore states that in 2025, AI voice agents transitioned from a conceptual technology to a product that enterprises are actively purchasing and deploying at scale.
High employee turnover and staffing difficulties in the healthcare industry are significant drivers for the adoption of AI voice agents.