As 2025 comes to a close, we’re taking a moment to look back at the sheer scale of what you folks have built on Oration.

Last year, the conversation around AI voice agents was largely about potential: "What could this do for my team?" This year, that question changed. The numbers suggest that 2025 was the year AI agents stopped being an experiment and started being infrastructure.

2025 by the numbers

Every single one of these digits represents a real person getting an answer or solving a problem without the wait.

• 21,832,312 conversations: Over 21 million times this year, an Oration agent handled a real-world interaction.
• 69,857 hours of conversation: That’s almost 8 years of continuous, non-stop conversation happening in a single calendar year.
• 378 unique AI agents created: From customer support to debt collection to lead qualification, our customers are building specialized agents for every possible workflow.

Moving beyond the pilot

In early 2025, we worked with teams running small-scale pilots. By December, many of those same teams have integrated Oration into the very core of their operations.

We’ve seen agents manage increasingly complex tasks—handling multi-step verifications, navigating nuanced customer emotions, and executing real-time API actions to solve problems on the spot.

The bar for "good" has moved. It’s no longer enough for an agent to just sound human; they have to be useful. They have to get the job done.

To 2026 and beyond

If 2025 was about proving scale, 2026 will be about depth. We’re already seeing agents become more versatile, more context-aware, and more integrated into the rest of the tech stack.

Thank you to every customer, developer, and partner who built with us this year. You’ve turned 21 million calls into 21 million moments of progress.

Happy New Year!