WebRider: Persona-Conditioned Agents for Live-Web Assistance

Zhi Li, Tao Zhou, Yeqing Li, Eugene Ie, Demetri Terzopoulos· August 10, 2026 View original

Key takeaways

  • Web agents often complete tasks but fail to adhere to delegated policies.
  • WebRider formalizes delegated policies as "intent contracts" for fidelity.
  • It uses a hierarchical architecture for contract management and action execution.
  • RiderBench evaluates agents on policy adherence and persona consistency on live websites.

Who benefits

Customer ServiceE-commerceBusiness Process AutomationSoftware DevelopmentAI/Tech

Summary

WebRider introduces persona-conditioned intent controllers for live-web assistance, formalizing delegated policies as "intent contracts" to ensure fidelity beyond just task completion. Its hierarchical architecture and new benchmark, RiderBench, evaluate agents on policy adherence and persona consistency across real websites.

Current live-web agents are typically evaluated solely on their ability to complete a task and provide a final answer, often overlooking whether they adhere to the specific policy constraints or preferences delegated by the user. This can lead to situations where a task is completed, but the underlying policy is violated. WebRider addresses this critical gap by formalizing the delegated policy as an "intent contract." This intent contract is an operational record encompassing goals, constraints, evidence obligations, answer format, and task-local persona controls, which must be maintained throughout the web interaction. WebRider employs a hierarchical architecture: a top-layer controller manages the contract, a middle layer translates intentions into guarded executable actions, and a tool layer executes these actions using browser, search, and maps tools. A new benchmark, RiderBench, evaluates agents on 4,096 live-web contracts across 42 public websites, auditing both internal contract state and user experience for policy preservation and persona consistency.

Why it matters

For professionals developing or deploying web automation and AI assistants, WebRider provides a framework to build more trustworthy and compliant agents that not only complete tasks but also strictly adhere to user-defined policies and personas.

How to implement this in your domain

  1. 1Adopt the "intent contract" concept for defining and evaluating web automation tasks.
  2. 2Design AI agents with hierarchical control architectures for policy adherence.
  3. 3Utilize the RiderBench dataset for rigorous testing of web agents' policy fidelity.
  4. 4Develop internal auditing mechanisms to verify agent actions against explicit policy constraints.

Original post by Zhi Li, Tao Zhou, Yeqing Li, Eugene Ie, Demetri Terzopoulos

"arXiv:2608.06704v1 Announce Type: new Abstract: Delegating a web task involves more than asking a question; it requires transferring a policy: what to verify, how to handle uncertainty, which preferences matter, and when to stop. Yet, current live-web agents are evaluated solely…"

View on X

Originally posted by Zhi Li, Tao Zhou, Yeqing Li, Eugene Ie, Demetri Terzopoulos on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses