As the demand for accurate carbon accounting soars, the application of Artificial Intelligence (AI) in this field is becoming increasingly vital. On November 28, 2024, Carbon+Alt+Delete hosted a webinar discussing how AI can help experts to save time in carbon accounting. In this blogpost, we summarize the key insights.
AI: hype versus value
AI has become a buzzword in recent years, and its significance in various industries is undeniable. Despite the common criticisms—AI’s susceptibility to errors and its considerable energy consumption—Large Language Models (LLMs) represent a technological milestone. These models can dramatically streamline experts’ workflows without replacing the invaluable human element.
The truth is that carbon accounting professionals need the time savings that AI can deliver, in order to follow market demand. According to Carbon+Alt+Delete’s analysis, the market for carbon accounting in Europe is projected to grow by a factor of four, from €1.4 billion today to €4.9 billion by 2028. Half of this growth must come from increasing the capacity of carbon experts, by refocusing existing experts and training new carbon experts. The other half of the growth needs to come from efficiency gains. In particular decreasing the time spent on data collection and data cleaning. This last aspect is where AI is very strong.
What sustainability consultants think
During the webinar, the audience of sustainability experts where asked how they look at the role of AI in sustainability reporting. It turned out that a large majority is optimistic that AI can deliver value for them (see figure below).
Figure 1. Results from a poll taken among 139 sustainability consultants on the value of AI in sustainability reporting
In the rest of this blogpost, we zoom in on 2 concrete use cases where AI can really deliver value in carbon accounting.
Use case 1: finding the best available emission factor
Selecting the right emission factor can be a challenge. The descriptions of activity data can sometimes be too vague or overly specific, making it difficult to match them with the correct emission factor. With over 100,000 emission factors available, navigating through the options can be overwhelming. Additionally, there is often no single “best” factor, especially in Scope 3 emissions, particularly for goods and services.
This is where AI can help. By analysing language patterns, AI can identify factors with statistically similar descriptions, making it easier to find relevant matches. AI can also search across different languages, broadening the scope of available data. Moreover, AI-driven insights can suggest potential emission factors that might not have been initially considered.
Use case 2: reading activity data from invoices
Beyond emission factor selection, data collection remains one of the biggest pain points in carbon accounting. Many companies struggle with centralized data availability, inefficient data formats, long lead times, multiple iterations, and last-minute changes. These issues slow down reporting and decision-making. Even when data is collected, quality concerns persist. Companies often face missing or proxy data, unclear data sources, incomplete records, and inconsistent terminology. Without reliable data, carbon footprint calculations lack precision, making it difficult to track progress accurately.
AI-powered solutions can address these challenges by automating data collection and improving accuracy. Machine learning algorithms can process large volumes of financial and operational data, ensuring completeness and consistency.
A promising AI-driven approach is invoice screening, where financial data serves as a centralized, exhaustive source of truth. By analysing invoices, AI can extract relevant carbon data, allowing for activity-based footprinting rather than broad, less actionable spend-based estimations.
As technology improves, large language models are becoming more effective in recognizing and interpreting carbon-related data in invoices and other business documents. The long-term goal we have at Carbon+Alt+Delete is to automate invoice processing overnight, providing carbon accountants with an 80% completed footprint, significantly reducing manual workload and improving accuracy.
By integrating AI into carbon accounting, consultants can enhance efficiency. Feel free to contact us at info@carbonaltdelete or book a demo if you want more information on how Carbon+Alt+Delete leverages AI to save time from carbon consultants.
