@inproceedings{harsha-etal-2026-synthesizing,
title = "Synthesizing question answering data from financial documents: An End-to-End Multi-Agent Approach",
author = "Harsha, Chetan and
Phogat, Karmvir Singh and
Dasaratha, Sridhar and
Ramakrishna, Shashishekar",
editor = {Matusevych, Yevgen and
Eryi{\u{g}}it, G{\"u}l{\c{s}}en and
Aletras, Nikolaos},
booktitle = "Proceedings of the 19th Conference of the {E}uropean Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 5: Industry Track)",
month = mar,
year = "2026",
address = "Rabat, Morocco",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.eacl-industry.51/",
pages = "669--687",
ISBN = "979-8-89176-384-5",
abstract = "Answering complex questions that require numerical reasoning over financial documents is challenging due to the diverse and scatterednature of relevant information. While large language models (LLMs) excel at financial reasoning, their enterprise deployment is often limited by cost and latency. Small language models (SLMs) present a cost-effective alternative but need to be fine-tuned with high-quality, domain-specific question-answer (QA) data. Acquiring such data requires manual expert annotation, presenting a bottleneck to the wider application of SLMs.This work introduces a modular, scalable end-to-end agentic pipeline that extracts and selects relevant content from unstructured financial documents and then generates QA pairs from the selected content for SLM fine-tuning. Compared to the same models trained on previous manually generated data for the task, one of the models trained on our pipeline-produced synthetic data achieved competitive in-distribution performance, and all tested models demonstrated superior generalization. The framework thus demonstrates considerable potential to accelerate the deployment of smaller, cost-effective models by reducing manual data creation efforts."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="harsha-etal-2026-synthesizing">
<titleInfo>
<title>Synthesizing question answering data from financial documents: An End-to-End Multi-Agent Approach</title>
</titleInfo>
<name type="personal">
<namePart type="given">Chetan</namePart>
<namePart type="family">Harsha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Karmvir</namePart>
<namePart type="given">Singh</namePart>
<namePart type="family">Phogat</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sridhar</namePart>
<namePart type="family">Dasaratha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shashishekar</namePart>
<namePart type="family">Ramakrishna</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-03</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yevgen</namePart>
<namePart type="family">Matusevych</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gülşen</namePart>
<namePart type="family">Eryiğit</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nikolaos</namePart>
<namePart type="family">Aletras</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Rabat, Morocco</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-384-5</identifier>
</relatedItem>
<abstract>Answering complex questions that require numerical reasoning over financial documents is challenging due to the diverse and scatterednature of relevant information. While large language models (LLMs) excel at financial reasoning, their enterprise deployment is often limited by cost and latency. Small language models (SLMs) present a cost-effective alternative but need to be fine-tuned with high-quality, domain-specific question-answer (QA) data. Acquiring such data requires manual expert annotation, presenting a bottleneck to the wider application of SLMs.This work introduces a modular, scalable end-to-end agentic pipeline that extracts and selects relevant content from unstructured financial documents and then generates QA pairs from the selected content for SLM fine-tuning. Compared to the same models trained on previous manually generated data for the task, one of the models trained on our pipeline-produced synthetic data achieved competitive in-distribution performance, and all tested models demonstrated superior generalization. The framework thus demonstrates considerable potential to accelerate the deployment of smaller, cost-effective models by reducing manual data creation efforts.</abstract>
<identifier type="citekey">harsha-etal-2026-synthesizing</identifier>
<location>
<url>https://aclanthology.org/2026.eacl-industry.51/</url>
</location>
<part>
<date>2026-03</date>
<extent unit="page">
<start>669</start>
<end>687</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Synthesizing question answering data from financial documents: An End-to-End Multi-Agent Approach
%A Harsha, Chetan
%A Phogat, Karmvir Singh
%A Dasaratha, Sridhar
%A Ramakrishna, Shashishekar
%Y Matusevych, Yevgen
%Y Eryiğit, Gülşen
%Y Aletras, Nikolaos
%S Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track)
%D 2026
%8 March
%I Association for Computational Linguistics
%C Rabat, Morocco
%@ 979-8-89176-384-5
%F harsha-etal-2026-synthesizing
%X Answering complex questions that require numerical reasoning over financial documents is challenging due to the diverse and scatterednature of relevant information. While large language models (LLMs) excel at financial reasoning, their enterprise deployment is often limited by cost and latency. Small language models (SLMs) present a cost-effective alternative but need to be fine-tuned with high-quality, domain-specific question-answer (QA) data. Acquiring such data requires manual expert annotation, presenting a bottleneck to the wider application of SLMs.This work introduces a modular, scalable end-to-end agentic pipeline that extracts and selects relevant content from unstructured financial documents and then generates QA pairs from the selected content for SLM fine-tuning. Compared to the same models trained on previous manually generated data for the task, one of the models trained on our pipeline-produced synthetic data achieved competitive in-distribution performance, and all tested models demonstrated superior generalization. The framework thus demonstrates considerable potential to accelerate the deployment of smaller, cost-effective models by reducing manual data creation efforts.
%U https://aclanthology.org/2026.eacl-industry.51/
%P 669-687
Markdown (Informal)
[Synthesizing question answering data from financial documents: An End-to-End Multi-Agent Approach](https://aclanthology.org/2026.eacl-industry.51/) (Harsha et al., EACL 2026)
ACL