LLaMandement sorts and summarizes amendments for French officials

The Cour des comptes found that the DGFiP's LLaMandement frees staff from sorting and summarizing amendments, so teams absorbed 85,345 amendments without extra staff.

  • 1 an organization

    Staff were freed from sorting and summarizing amendments, and teams handled 85,345 amendments without extra staff.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Reported benefit.
Evidence

foundstudiedreportedclaimed

An independent audit report supports the benefit.

The system’s part
Not isolated. The Cour credits the system with the time saved but gives no measurement and no comparison of work with and without it. The Cour does not describe the model. In their own paper, the developers describe a large language model fine-tuned by the French government.
Limits
Not yet checked.
Costs and harms
Not yet checked.
Later
Not yet checked.

Download data Cite Suggest a correction

Benefit type
Repetitive Task Automation
System
LLaMendements (LLaMandement), the DGFiP's tool for sorting and summarizing parliamentary amendments
Made by
Direction générale des Finances publiques (DGFiP)
Used by
DGFiP teams handling parliamentary amendments for the economy and finance ministry
Where
France (national)
Stage
In routine use
Date
, when the source was published
Kind
Aggregate result
Broad benefit
Processing Information, Interpreting the Meaning of Information for Others
Collections
Scrutiny and findings
Last revised
31 October 2026

Sources

  1. Audit report, by authors with no tie to the system.

    Cour des comptes. L’intelligence artificielle dans les politiques publiques : l’exemple du ministère de l’économie et des finances. Observations définitives S2024-1165, adopted 18 July 2024 and published 22 October 2024, p. 35.

    LLaMendements permet donc aux équipes d’absorber la charge de travail (85 345 amendements déposés entre juin 2022 et janvier 2024) sans recourir à des moyens humains supplémentaires.

  2. Preprint, by the makers. Context or contrary evidence.

    Gesnouin J, Tannier Y, Gomes Da Silva C, et al. LLaMandement: Large Language Models for Summarization of French Legislative Proposals. arXiv:2401.16182, 29 January 2024.

    This report introduces LLaMandement, a state-of-the-art Large Language Model, fine-tuned by the French government and designed to enhance the efficiency and efficacy of processing parliamentary sessions (including the production of bench memoranda and documents required for interministerial meetings) by generating neutral summaries of legislative proposals.

Details

AI system

Workflow

The system took over the sorting and summarizing of parliamentary amendments. Staff put the time saved into the quality and speed of the ministry's answers.

Follow-up

Conditions for replication

Teams must absorb a growing volume of amendments at constant staffing.

Entry details

Sector

Public administration (legislative work).

Purpose

The DGFiP built LLaMandement to sort parliamentary amendments and write neutral summaries of them for ministry staff.

Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). LLaMandement sorts and summarizes amendments for French officials (AIBD-0680). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0680.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)

BibTeX Data Field definitions