Microsoft Clarity has added an Agent Activity dashboard and task simulations that let businesses see which supported AI agents visit their websites, what they do there, and whether they can complete key tasks. Announced November 4, the capabilities are available globally to Clarity users whose projects receive agentic traffic; the supported operators currently include Google, OpenAI, and Anthropic. Simulations start with commerce and travel scenarios. The shift matters because AI systems are moving from recommending businesses to interacting with their websites.
The dashboard records actions such as site searches, form submissions, logins, downloads, appointment bookings, and checkout. Simulations report whether a task was completed, how long it took, and how many interactions it required. Microsoft says the goal is to “test how effectively agents can complete important tasks on your website before a real customer journey fails.” That turns agent compatibility into something a team can inspect, rather than guess about. (Microsoft Clarity’s launch announcement; Search Engine Land’s launch coverage)
AI agents are starting to use websites, not just cite them
Clarity’s related study analyzed more than 5.2 million agentic-classified sessions across 230,511 Clarity projects from August 13 to September 9, 2026. In the same cohort of sites, a checkout event appeared in 25.21% of agentic-classified sessions, compared with 14.65% of human-only sessions. Form submissions appeared in 19.85% versus 18.56%, and appointment bookings appeared in 8.42% of agentic sessions. These are recorded website events, not proof that an order was paid or an appointment kept. (Microsoft Clarity’s agent-session study)
The strategic change is broader than search visibility. A site may be mentioned by an assistant and still fail when an agent tries to compare options, find a service area, submit a form, or move through checkout. For businesses, the customer journey increasingly includes a software user acting on someone’s behalf. A broken label, missing detail, or unclear next step can become a conversion barrier even when the underlying offer is strong.
Use the numbers as a signal, not a market forecast
Microsoft says weekly agent sessions in its analysis nearly doubled over three weeks to almost 2 million; its study describes an 88% increase across the analyzed projects. That is a vendor-reported trend across Clarity’s measured sites, not a claim that every website saw similar growth or that agents now account for a large share of all web traffic. The report does not establish that the observed checkout events led to purchases.
There is another attribution limit: an agentic-classified session can include a handoff to a person, and Clarity says it cannot yet separate the agent’s activity from the human’s activity within that session. The analysis identifies activity associated with supported agent operators, not every AI assistant or every automated visit on the web. Treat the dashboard as a diagnostic lens for friction, then confirm leads, bookings, and revenue in the systems that record completed business outcomes.
Agent readiness is now part of website quality assurance
For a retailer, test whether an agent can find the right product, understand its options, and reach checkout without ambiguity. For travel businesses, check that availability and booking steps are accessible and consistent. Microsoft’s first simulation scenarios focus on commerce and travel, so businesses in other sectors should not assume every workflow is covered yet. Where simulations are unavailable, teams can still review key journeys manually and monitor supported agent activity when it appears.
Make the important information legible to both people and software: use clear labels, descriptive buttons, structured product or service details, accurate prices and availability, and forms that explain required fields. Review barriers such as login requirements, pop-ups, consent prompts, or confusing navigation on high-value pages. These are practical QA priorities, not a guarantee that an agent will behave like a human visitor or complete every task.
What businesses should do in the next 30 days
Start by confirming that Clarity is installed on priority pages and that important actions—such as quote requests, appointment bookings, or completed purchases—are measured in your analytics and CRM. Check whether agent traffic is appearing, which pages it reaches, and what actions are recorded. Choose one high-value journey and test it from discovery through completion; use the simulation if your project and journey are supported, and otherwise perform a documented manual review.
Fix the highest-impact friction first, then compare agent sessions with verified human and business outcomes. Keep a clear handoff to a person when an agent cannot proceed, and do not loosen authentication, privacy, or fraud controls just to make a journey easier for automation. The goal is not to optimize a site for bots at any cost. It is to make legitimate customer journeys clear and resilient whether a person or an agent is navigating them.
Real Internet Sales helps businesses strengthen websites, digital marketing, and visibility across evolving search experiences. Call 803-708-5514 or visit realinternetsales.com to discuss an actionable plan.