Gains:
- Ability to distinguish between single materiality and double materiality and evaluate an issue in terms of both impact and financial materiality
- Ability to separate hundreds of stakeholder responses into themes with artificial intelligence while keeping low-frequency but high-importance signals separate
- Ability to test the list of topics produced by artificial intelligence with industry-specific reality and link the final prioritization and methodology document to humans
No company can report on every sustainability issue in the same depth; While water consumption is a secondary issue for a software company, it is the lifeline for a beverage manufacturer. This is the question "which issues are really important to us?" The answer to the question is given by materiality analysis (importance analysis). Materiality is the process of determining how “significant/significant” a sustainability issue is to an organization and forms the basis of the entire ESG report. False materiality renders even hundreds of pages of reports worthless; because the report focuses on the wrong issues. In this unit, you will learn where to use artificial intelligence (AI) as an accelerator and where as a risk in this process.
Let's clarify two concepts first. Single materiality looks only at how an issue affects a company's financial health: "How will climate change affect our revenue/costs?" Double materiality adds a second aspect: the company's impact on the world. So the question is twofold: (1) How does this issue affect the company financially? (2) How does the company impact the environment and society in this regard? CSRD and ESRS enforce this second, double look. This distinction is at the heart of both the exams and the practice.
Two aspects of double significance
Double materiality looks through two lenses:
- Impact materiality — "inside out": The positive/negative impact of the company's activities on people and the environment. Example: wastewater released into the river by a textile factory.
- Financial materiality — “outside-in”: The impact of a sustainability issue on a company's cash flows, costs, and value. Example: increasing the carbon price increases the cost of production.
If a topic is material in either of these two lenses, it is considered reportable in the dual materiality logic. AI has a hard time distinguishing between these two aspects and often drifts only towards the financial side; Therefore, it is essential to control the output with this distinction in mind.
Subject
Impact materiality (to the world)
Financial materiality (to the company)
greenhouse gas emissions
Contribution to climate change
Carbon tax, CBAM cost
Worker health and safety
Employee life and health
Accident compensation, production stoppage
water usage
Regional water stress
Water shortage, production risk
Data privacy
User rights
Penalty, reputation and customer loss
Tip: When considering an issue, "does this affect us or do we affect the world?" Divide it into two. Much greenwashing and underreporting occurs when companies report only the financial side that affects them, hiding the impact they have on the world.
Step by step: AI-powered materiality process
1. Create the subject universe. Make a long list of possible sustainability topics relevant to your industry. AI is a good brainstorming partner here; A candidate list is produced by scanning ESRS subject headings, GRI industry standards and peer company reports.
2. Identify stakeholders. Stakeholders are parties that are affected by or affected by the company's activities: employees, customers, suppliers, investors, local people, regulators, civil society. For every important issue, “who is affected?” question is asked.
3. Gather stakeholder input. Find out what issues stakeholders find important through surveys, interviews and workshops. AI is very powerful at sorting hundreds of open-ended survey responses into themes (this is called thematic analysis).
4. Score and prioritize. Score each topic on a scale on two axes (impact and financial). Topics with high scores are considered "important/material" and become the focus of the report.
5. Verify and document. A management committee approves prioritization; How the process is done (methodology) is documented. CSRD also demands that the process of materiality itself be explained.
Weak prompt / Strong prompt
Weak prompt:
Tell us about the sustainability issues that are important to you.
It does not include industry, business model, stakeholder and duality aspects; AI produces a generic list and often skips the direction of impact.
Powerful prompt:
Your role: double materiality expert.Company: A medium-sized food company producing in Turkey.Task: list possible sustainability issues and make TWO SEPARATE assessments for each:- Impact materiality (the company's impact on the environment/society) - Financial materiality (the issue's financial impact on the company)For each issue, also write down the affected stakeholders.Rule: This is a DRAFT candidate; Scoring and final decision belongs to us. Mark sector data that you are not sure about as fabricated, "must be verified". Output: table (Topic | Impact direction | Financial aspect | Stakeholders | Note).
Sort stakeholder responses into themes
It takes days to hand-read hundreds of free-text responses from materiality surveys. AI can break these down into themes; but with a caveat: AI may dismiss a minority but critical opinion (for example, a serious security issue raised by a single employee) as “unimportant.” Therefore, the thematic summary is not a substitute for raw responses; It is a gateway to them.
Below are 200 free text responses from the materiality survey.Task:1) Divide the responses into no more than 10 themes, giving each theme a short name.2) Count how many responses mention each theme.3) ALSO list, eliminate, a small number of but serious (safety, legal, environmental risk) opinions under the heading "low frequency / high importance".4) Do not make up or combine or distort any responses.Answers: [ANSWERS]
Caution: The thematic summary of AI is not a substitute for raw answers. In particular, “low frequency but high importance” opinions (a security breach, a suspicion of corruption) can be overshadowed by AI because they are few in number. Be sure to check these opinions with a separate eye.
three mini cases
Case 1 — Omitting the direction of action. A chemical company had AI prepare a materiality list. The list focused on financial issues such as the carbon price and the cost of energy; but its impact on the water supply of the village near the factory was not on the list. The expert added the “impact” aspect as a matter of double importance; This issue later turned out to be the most critical stakeholder issue.
Case 2 — Weight in numbers. A retailer analyzed 1,200 customer surveys with AI. YZ found that 41% of the responses touched on the theme of "packaging waste" and 23% touched on the theme of "working conditions". The team declared these two themes as priority material topics; However, a supply chain that appeared under the "low-frequency" heading also brought the suspicion of child labor under separate scrutiny.
Case 3 — Fake importance. AI flagged "water consumption" as a high priority material issue for a software company; whereas the company's only water use was the office kitchen. The expert downplayed this and focused on the really material "data privacy and energy consumption". The report would have lost focus if the incorrect material had not been corrected.
Prompt to set up the materiality matrix
The materiality matrix is a visual prioritization tool that places topics along two axes (impact and financial). The topics in the upper right corner are the most important.
Arrange the following topics and scores in the logic of a materiality matrix. Input: Impact score (1-5) and Financial score (1-5) for each topic. Task: 1) Divide the topics into four zones: high-high (priority), high impact-low finance, low impact-high finance, low-low. 2) Suggest the "priority" region as the main focus of the report. 3) Define the borderline (around 3-3) topics as "human judgment required" mark.Points: [TOPIC AND SCORES]
The following prompt produces an outline for writing the materiality methodology into the report.
Your role: ESG report writer. Task: write 1 auditable paragraph explaining our dual materiality process. Include: which subject universe was scanned, which stakeholders were involved, by what method it was scored, who approved it. Rule: write understated, verifiable and process-oriented; adding result assertion.Input: [PROCESS NOTES]
Common mistakes
- Being satisfied with one important thing. Looking only at the financial impact and ignoring the company's impact on the world is against CSRD.
- Keeping the stakeholder narrow. Listening only to investors and excluding employees, local people and suppliers distorts materiality.
- Accepting the AI's list as is. Critical issues specific to the sector may be missing or given the wrong weight in AI.
- Eliminating low-frequency serious opinions. Signals that are few in number but high risk should be examined separately.
- Not documenting the methodology. If it is not written how the process is done, the auditor will not accept the materiality.
In summary
Material analysis is the compass of the entire ESG report; A report that focuses on the wrong issues is worthless, no matter how quality it is written. The double-importance is both “how does this issue affect us financially?” as well as "how do we influence the world on this issue?" asks questions together. AI is a powerful accelerator in generating the universe of topics and sorting hundreds of stakeholder responses into themes; but it tends to miss the direction of impact, create false importance, and eliminate low-frequency critical views. Final prioritization and approval always rests with the human.
Application task
Choose an industry (e.g. food retail) and ask the AI for a list of 12-15 candidate topics with the powerful prompt above. Then: (1) check each topic by dividing it into two in terms of impact and financial aspect, (2) correct only those that have shifted to the financial side, (3) select the 5 topics you think are the most material with justification and draft a small materiality matrix.
checklist
- [ ] I evaluated each issue in terms of both impact and financial materiality.
- [ ] I expanded the list of stakeholders beyond the investor (employee, public, supplier).
- [ ] I tested the list of topics generated by AI against industry-specific reality.
- [ ] I examined low frequency but high importance opinions separately.
- [ ] I approved the prioritization by a human/committee decision.
- [ ] I have documented the materiality methodology in accordance with the audit.
- [ ] I dropped issues of false importance (that do not really affect us).