aiqbench analysis
This research introduces a novel framework for agent proactivity by distinguishing between horizontal and vertical information pursuit, moving beyond the traditional binary of acting autonomously versus waiting for prompts. The proposed Q&D method trains agents to retrieve necessary evidence without relying on external reward models or judges, achieving significant performance gains over larger prompted models in multi-hop reasoning tasks. These findings advance the field by defining new metrics for agent behavior that focus on content selection and timing rather than mere initiative.
Asking for What Was Never Requested: Horizontal and Vertical Proactivity in Agents
arXiv:2609.37236v1 Announce Type: cross Abstract: An agent that uses tools typically responds to what the user explicitly asks, yet completing the task may require information the user never requested. Work on proactive agents mainly studies whether and when an agent should act on its own, not what information it should pursue. We study a distinct axis of proactivity: its content. Horizontal proactivity pursues unstated information that the current context already identifies, and vertical proactivity pursues needs that only earlier evidence reveals. A need graph, recovered from a benchmark's own decomposition, records which needs depend on which, so both forms, and whether the agent stops at the right time, can be scored from a transcript without a model judge. To learn this behavior, we propose Q&D (questioner and drafter), which trains a questioner to prefer the question whose continuation retrieves more of the required evidence, with no reward model or judge. On held-out splits of three multi-hop question-answering benchmarks, at equal retrieval spend, the trained questioner improves both forms of proactivity over the same model, prompted, and outperforms a prompted model $15\times$ larger in the same role on two of the three, and the gain persists after controlling for question volume and length. Without further training, we place the questioner in an interactive customer-service agent with a simulated customer, where it completes more tasks while asking fewer questions, and in retail it outperforms the $15\times$ larger model with fewer follow-up turns from the customer. These results show that proactivity depends not only on whether an agent acts without being asked, but also on what it chooses to pursue and when it stops.
Abstract:An agent that uses tools typically responds to what the user explicitly asks, yet completing the task may require information the user never requested. Work on proactive agents mainly studies whether and when an agent should act on its own, not what information it should pursue. We study a distinct axis of proactivity: its content. Horizontal proactivity pursues unstated information that the current context already identifies, and vertical proactivity pursues needs that only earlier evidence reveals. A need graph, recovered from a benchmark's own decomposition, records which needs depend on which, so both forms, and whether the agent stops at the right time, can be scored from a transcript without a model judge. To learn this behavior, we propose Q&D (questioner and drafter), which trains a questioner to prefer the question whose continuation retrieves more of the required evidence, with no reward model or judge. On held-out splits of three multi-hop question-answering benchmarks, at equal retrieval spend, the trained questioner improves both forms of proactivity over the same model, prompted, and outperforms a prompted model $15\times$ larger in the same role on two of the three, and the gain persists after controlling for question volume and length. Without further training, we place the questioner in an interactive customer-service agent with a simulated customer, where it completes more tasks while asking fewer questions, and in retail it outperforms the $15\times$ larger model with fewer follow-up turns from the customer. These results show that proactivity depends not only on whether an agent acts without being asked, but also on what it chooses to pursue and when it stops.
48 pages. Project page: this https URL Code: this https URL Model: this https URL
Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
arXiv:2609.37236 [cs.AI]
(or arXiv:2609.37236v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2609.37236 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 379 words
Related coverage
Evaluation-First AI Agents: How Zepto Scales Customer Support on Databricks and MLflow
Databricks · Sep 9, 2026
Presentation: Exploding GPUs
InfoQ · Oct 16, 2026
Why AI Needs a “Genie Coefficient”
IEEE Spectrum · Jul 21, 2026
A shared agentic platform for Wood Mackenzie, on Amazon Bedrock AgentCore
AWS · Sep 17, 2026