AI Agent Traffic Surges, Exposing Organizational Blind Spots
Event summary
- DataDome recorded 7.9 billion AI agent requests in January and February 2026, a 5% increase over Q4 2025.
- Agentic traffic accounted for up to 9.75% of total traffic for some customers over a 30-day period.
- Meta-externalagent was the most impersonated agent with 16.4 million spoofed requests, followed by ChatGPT-User with 7.9 million.
- E-commerce and retail industries saw ~20% of AI agent traffic volume, highlighting high-value transactional data risks.
The big picture
DataDome's report underscores the growing challenge of managing AI agent traffic, which is surging in volume and complexity. Organizations struggle to distinguish between high-volume and high-value agents, exposing them to fraud risks. This trend highlights the need for advanced bot and agent trust management solutions as AI interactions with websites become more prevalent.
What we're watching
- Identity Crisis
- How organizations will adapt to accurately classify AI agents by identity and intent, given the current visibility gap.
- Industry Concentration
- Whether high-value transactional data industries like e-commerce and real estate will prioritize agentic traffic management.
- Spoofing Risks
- The pace at which spoofed AI agent requests will evolve, challenging existing allowlisting strategies.
Related topics