Exploring the possibilities of enhancing the traditional measurement and reporting Scope 3 upstream emissions by utilizing artificial intelligence in supplier networks within electronics industry

Loading...
Thumbnail Image

URL

Journal Title

Journal ISSN

Volume Title

School of Science | Master's thesis

Department

Mcode

Language

en

Pages

103

Series

Abstract

Legislative frameworks play a vital role in combating global warming and climate change. The Greenhouse Gas (GHG) Protocol is one of the most widely adopted frameworks, offering strategic benefits for companies and categorizing emissions into three main scopes. Among these, Scope 3 emissions typically represent the largest share of total emissions and pose significant challenges for accurate meas-urement and reporting. This thesis focuses on data collection from supplier networks as one of the key chal-lenges in Scope 3 emissions measurement and reporting. An Action Design Re-search (ADR) project was conducted in collaboration with a Finland-based electron-ics manufacturing company, with two primary objectives: to enhance supplier data collection processes and to improve data estimation and gap-filling in Scope 3 re-porting. Over a nine-month period, two interventions were designed: a mind map–based framework to improve survey design, and an AI agent for data gap filling and estimation. The AI agent was then evaluated in terms of validity, reliability, and completeness using real-world data from the company’s supplier network. This thesis has main theoretical contributions, including identifying new challenges in Scope 3-related data collection, developing a novel use case for AI agents in the sustainability context, and highlighting the role of modularity in designing AI agents for data estimation and gap-filling purposes. For sustainability managers, this thesis offers key takeaways: designing more effective surveys for supplier Scope 3 data collection, leveraging AI agents to support Scope 3-related data estimation and gap filling, and providing improvement suggestions and design insights for using AI agents in sustainability or other organizational contexts.

Description

Supervisor

Holmström, Jan

Thesis advisor

Jaribion, Alireza
Rantonen, Sari

Other note

Citation