<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zambrano-etal-2026-way">
    <titleInfo>
        <title>Does The Way You Plan Matter? An Empirical Study of Planning Representations for LLM Web Agents</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Alejandra</namePart>
        <namePart type="family">Zambrano</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Sara</namePart>
        <namePart type="given">Vera</namePart>
        <namePart type="family">Marjanovic</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Imene</namePart>
        <namePart type="family">Kerboua</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Xing</namePart>
        <namePart type="given">Han</namePart>
        <namePart type="family">Lù</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Leila</namePart>
        <namePart type="family">Kosseim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Ehsan</namePart>
            <namePart type="family">Kamalloo</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Nicolas</namePart>
            <namePart type="family">Gontier</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Nikolai</namePart>
            <namePart type="family">Rozanov</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Maryam</namePart>
            <namePart type="family">Hashemzadeh</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Akhil</namePart>
            <namePart type="family">Arora</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-485-9</identifier>
    </relatedItem>
    <abstract>Despite recent advances, LLM-based web agents still struggle with limited exploration, omission of critical steps, and sensitivity to task constraints. Prior work suggests that many of these failures stem from weaknesses in planning, yet the impact of alternative natural language plan representation remains unexplored. To address this, we introduce , a static planner-executor framework that evaluates the impact of plan representation in agent performance. We first automatically categorize WebArena tasks into 3 difficulty levels, enabling consistent difficulty grading without human annotation. Then we systematically evaluate 4 different plan representations on the tasks categorized as hard: sequential subgoals, narrative, pseudocode, and checklist; across different families of multimodal LLM powered agents (OpenAI, Alibaba, and Google). To account for stochastic variability, we introduce a novel decomposition of agent Success Rate (SR) into 2 interpretable evaluation metrics: Achievement Rate (AR) and Solved-Task Consistency (STC). Our results show that both, the plan formulation and the underlying LLM generating the plan, significantly influence web-agent robustness and task success. Code and dataset are available [here](https://github.com/alzambranolu13/plan-ahead).</abstract>
    <identifier type="citekey">zambrano-etal-2026-way</identifier>
    <location>
        <url>https://aclanthology.org/2026.realm-1.96/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>1375</start>
            <end>1388</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Does The Way You Plan Matter? An Empirical Study of Planning Representations for LLM Web Agents
%A Zambrano, Alejandra
%A Marjanovic, Sara Vera
%A Kerboua, Imene
%A Lù, Xing Han
%A Kosseim, Leila
%Y Kamalloo, Ehsan
%Y Gontier, Nicolas
%Y Rozanov, Nikolai
%Y Hashemzadeh, Maryam
%Y Arora, Akhil
%S Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026)
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-485-9
%F zambrano-etal-2026-way
%X Despite recent advances, LLM-based web agents still struggle with limited exploration, omission of critical steps, and sensitivity to task constraints. Prior work suggests that many of these failures stem from weaknesses in planning, yet the impact of alternative natural language plan representation remains unexplored. To address this, we introduce , a static planner-executor framework that evaluates the impact of plan representation in agent performance. We first automatically categorize WebArena tasks into 3 difficulty levels, enabling consistent difficulty grading without human annotation. Then we systematically evaluate 4 different plan representations on the tasks categorized as hard: sequential subgoals, narrative, pseudocode, and checklist; across different families of multimodal LLM powered agents (OpenAI, Alibaba, and Google). To account for stochastic variability, we introduce a novel decomposition of agent Success Rate (SR) into 2 interpretable evaluation metrics: Achievement Rate (AR) and Solved-Task Consistency (STC). Our results show that both, the plan formulation and the underlying LLM generating the plan, significantly influence web-agent robustness and task success. Code and dataset are available [here](https://github.com/alzambranolu13/plan-ahead).
%U https://aclanthology.org/2026.realm-1.96/
%P 1375-1388

Markdown (Informal)

[Does The Way You Plan Matter? An Empirical Study of Planning Representations for LLM Web Agents](https://aclanthology.org/2026.realm-1.96/) (Zambrano et al., REALM 2026)

• Does The Way You Plan Matter? An Empirical Study of Planning Representations for LLM Web Agents (Zambrano et al., REALM 2026)

ACL

• Alejandra Zambrano, Sara Vera Marjanovic, Imene Kerboua, Xing Han Lù, and Leila Kosseim. 2026. Does The Way You Plan Matter? An Empirical Study of Planning Representations for LLM Web Agents. In Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026), pages 1375–1388, Budapest, Hungary. Association for Computational Linguistics.