State of Agentic Traffic is HUMAN’s monthly benchmark on how AI agents are showing up across the web, reporting on agent mix, sector destinations, and activity patterns observed across the Human Defense Platform. Each edition extends findings from the annual State of AI Traffic & Cyberthreat Benchmark Report and is produced with analysis from HUMAN’s Satori Threat Intelligence Team, which investigates agentic traffic and other automated traffic to identify emerging trends, developments, and threats.
Key findings
- Agentic traffic grew 2.6x (157.5%) from September and approximately 8x from August.
- Meta and OpenAI now account for about 90% of agentic traffic. Meta’s Muse alone holds 59%, after nearly quadrupling month over month.
- In two months, Meta’s Muse replaced Perplexity’s Comet as the top source of agentic traffic. Comet held 41% in August and now holds 3.6%.
- Standalone consumer agentic browsers are fading, as consumers move to agents built into platforms they already use.
- Agents still browse far more than they buy. 85% of agentic traffic goes to product and search pages, while just % reach checkout. However, absolute checkout volume grew 117% month over month.
Agentic traffic grew 2.6x since September, and over 8x since August
October’s agentic traffic was 2.6 times September’s, a 157.5% increase. Compared with August, it was about 8 times higher (697.5%), reflecting the combined growth of September and October.
After holding roughly flat through the summer, agentic traffic took off in September with the arrival of Meta’s Muse, and the growth accelerated in October.
Meta and OpenAI account for about 90% of agentic traffic
In two months, Meta’s Muse went from a new entrant to 59% of all agentic traffic. Its share rose 19 percentage points from September, and its traffic nearly quadrupled month over month. Muse now sends 3.7 times as many requests as ChatGPT Agent, the next-largest agent.
OpenAI holds second place. Three of its products made the top 10, together accounting for about 30% of agentic traffic:
- ChatGPT Agent: Its share fell from 28.5% to 15.9% as Muse grew faster, but its traffic rose 44%.
- Codex Browser: Now in third place, its share grew from 7.7% to 8.8%, and its traffic nearly tripled (2.9 times).
- Codex Chrome extension: It moved from sixth to fourth place, and its share grew from 3.3% to 5.2%.
Outside Meta and OpenAI, the Claude Chrome extension dropped one spot to sixth place. Its share halved, from 7% to 3.5%, even though its traffic grew 30%.
Anthropic also launched a new agent, Claude Duet, in research preview in October. HUMAN Sightline detects Claude Duet, but its traffic is still very low while access is limited to the research preview.
Two newcomers rounded out the top 10. Google Agent (Gemini), which Google launched on October 8, reached eighth place with 0.8% within weeks. xAI’s Grok Agent, launched on August 11, debuted in ninth place with 0.7%.
Standalone consumer agentic browsers are fading
Perplexity’s Comet browser led agentic traffic with 41% share in August. Its share fell to 9.4% in September and to 3.6% in October, as it slipped from third to fifth place. It was the only top-10 agent whose traffic shrank in October, and its request volume is down about 30% since August, while total agentic traffic grew about 8 times.
Comet is not alone. Genspark’s agentic browser, after a months-long decline, dropped out of the top 10. And in August, OpenAI sunset its standalone agentic browser, ChatGPT Atlas.
Consumers are moving to agents built into platforms they already use, from companies that already reach them at scale. Developer-focused agentic browsers are a different story: OpenAI’s Codex Browser nearly tripled its traffic in October.
Industry shares of agentic traffic
The surge in agentic traffic raised absolute traffic in every sector. Ecommerce remains the largest, holding a stable 60% share, and its traffic grew 2.5 times.
Travel took second place for the first time. Its traffic grew 2.9 times, and its share rose from 17.1% to 19.3%. Media follows closely in third: its share fell from 21.1% to 19.1%, though its traffic grew 2.3 times.
SaaS is fourth at 1.2%, but it is growing fastest. Its traffic nearly quadrupled in one month and is about 15 times its August level.
Financial services, streaming and gaming, and healthcare grew more modestly, by 57%, 62.7% and 59.5%, respectively.
What agents are doing: browsing far more than buying
Agents are still mostly researching. 85% of agentic traffic went to product and search pages in October, up from 81% in September. Content engagement rose from 3.6% to 4.3%, moving into second place.
Account and transaction pages make up a shrinking share of agentic traffic. Account management fell from 7.1% in August to 4% in September and 3.2% in October. Authentication fell from 5.5% to 3.1% and now 2.8%. Miscellaneous traffic rose from 3% in August to 5.8% in September, then fell to 2.4%.
Checkout and payment fell from 2.4% to 1.9% of agentic traffic, but its volume grew 117% month over month. Agents still browse far more than they buy, yet the number of agent requests reaching checkout is rising fast. Next month’s report will show how agents behaved over Black Friday and Cyber Monday.
Methodology
Data reflects agentic traffic detected by HUMAN Sightline Cyberfraud Defense across applications protected by HUMAN during October 2026. Agent identification combines behavioral signal analysis, user-agent attribution, and publisher-level integration where available. Sector and route categorization reflects the destination endpoint, not the agent operator’s intent.
All figures are measured in requests. A request is one action an agent takes on a site, such as loading a page or running a search. It is not a user or a session.
This report is based on HUMAN’s AgenticTrust, which identifies agents directly through dedicated technical signals, so the traffic we report reflects agents as they actually operate on the web. We do not infer agentic activity from crawler traffic. Crawlers are easier to detect, but they largely serve LLM training, search indexing and chatbots, and so are not a reliable measure of how agents are used. Counting only agent traffic gives a clearer, more accurate view of agentic trends.
See agentic traffic on your own site
The benchmarks above show agentic activity observed by HUMAN Sightline Cyberfraud Defense and AgenticTrust. On most sites, analytics tools can’t distinguish AI agents from human visitors, let alone identify specific agents or classify their intent.
AgenticTrust is HUMAN’s trust and control layer for AI agents. It identifies and verifies Muse and dozens of other AI agents using proprietary research and digital signatures that cannot be spoofed. It gives security, fraud, commerce, and marketing teams continuous visibility into every agent session, classifies intent as the session unfolds, and enforces adaptive policies based on what agents are actually doing, down to whether a given agent is allowed to add to cart or log in.
The first step is visibility. To see what Muse and other AI agents are doing across your business, request a demo today.