DMA FAQ

Key concepts

Term	Meaning
IRO	Impact, Risk or Opportunity. The unit of analysis of a DMA. Four types: positive impact, negative impact, financial risk, financial opportunity.
Topic and subtopic	The ESRS structure: 10 topics (E1 to E5, S1 to S4, G1) and their subtopics. Karomia follows the ESRS 1 list of sustainability matters, updated after the Omnibus.
AI pre-assessment	Karomia's top-down scoring of each subtopic, before stakeholder engagement.
AI IRO scoring	Karomia's scoring of each generated IRO against the ESRS 1 criteria (scale, scope, irremediability, magnitude, likelihood).
Stakeholder results	The scores produced by your surveys, interviews and focus groups. These decide materiality.
Materiality threshold	The score at or above which a subtopic or topic is material, set per axis. It is the average of your topic scores on that axis per IRO type, and never lower than 3.
Undecided	A result within the margin below the threshold, 0.5 by default, which you decide on and document.
Recurring IRO	In a refresh, an IRO that continues a matter already present in your previous DMA.

Methodology and how the AI works

Which frameworks does the AI use to identify IROs? The IRO list is anchored on ESRS, because that is the structure your CSRD report is built on. The AI cross-checks it against other standards: SASB, the GRI Sector Standards, NACE codes, the EU Taxonomy and SBTi. Each activity in your value chain is linked to the relevant sectors, including the matching GRI sector. On top of that, the AI uses sector benchmarks, scientific and sector studies and web research, including recent news about your company, your suppliers and your customers. Your own documents stay the primary source.

Can I show my auditor where an IRO comes from? Yes. Every IRO carries its sources, each labelled with a source type: company document, sector or framework evidence, web source or inference. Claims that rely on a sector baseline are marked as such, so you can tell what was verified for your company and what is a well-founded sector expectation. See the Sources section of the IRO passport.

Why do some topics have far more IROs than others, for example water or pollution compared with climate change? Topics have different numbers of subtopics. Pollution has many, climate change few. The AI generates IROs per subtopic, so a topic with more subtopics gets more IROs. Per subtopic the numbers are much closer.

Are IROs linked to each other? Indirectly, through the subtopics they belong to. The AI does link related IROs internally to keep the analysis free of overlaps and gaps, but these links are not shown in the platform yet. If your auditor asks for interdependencies between IROs, tell us.

How does the AI decide whether an IRO is actual or potential, and which time horizon applies? From your documents and public sources. An impact that is described as already happening is actual; one that may happen is potential. The time horizon (short under 1 year, medium 1 to 5 years, long over 5 years) follows from the drivers of the IRO. Both labels are shown on every IRO and you can change them if you disagree.

How does the AI score scale, scope, irremediability, magnitude and likelihood? With the ESRS 1 criteria, based on the drivers and evidence it found for each IRO. The method and the formulas are described in How Karomia generates your IROs. AI scores are a baseline; they are never an input to your final results.

How is the value chain taken into account? Before generating IROs, the AI maps your activities upstream, in your own operations and downstream. Each IRO is labelled with its position in that value chain. See Review your value chain map.

Are silent stakeholders such as nature taken into account? Nature and future generations cannot answer a survey. They are represented in two ways: the AI analysis covers environmental impacts across the value chain from scientific and sector sources, and you can engage proxies such as environmental NGOs or internal environmental experts. Document which proxies you chose in your report.

Is there a minimum number of IROs per subtopic? No. There is no minimum, no maximum and no ratio to respect. Some subtopics will have more IROs than others. You do not have to select a risk or an opportunity in every subtopic either. If a subtopic is not relevant, for example animal welfare for a software company, you can leave it out. Write down why.

Should I prefer negative impacts over positive ones? We stress negative impacts because companies tend to focus on the positive. If a positive impact is genuinely relevant, select it. If both apply, select both.

What is this number, for example 151115, in my Excel export? It is the internal ID of a question or IRO. Search for it in the export to find the title, or ask us through the chat. In the new IRO table every IRO is shown with its title.

Setting up the assessment

How does the AI use the business activities and their size? Your documents are the main input and drive the precision of the IROs. Activity descriptions are a safeguard: they help the AI correct itself if it misread a document. The size score is a secondary signal. Make the spread meaningful, for example 10 for your core activity and 5 for a peripheral one, rather than 9 and 8.

Sales is missing from the default internal activities. Is it an external activity? No, sales is an internal activity. Add it.

Can we do a DMA for one entity or one country only? Yes. Create a separate assessment with the documents and the activities of that entity. You can reuse most of the group documents and add the entity-specific ones.

We operate several sites. How do site-specific IROs come out? Describe your sites and what happens at each in your documents or activities. The AI then generates IROs per site where it matters, for example air emissions at a French plant and not at the Belgian office. If you need results per site, create stakeholder groups per site.

After selecting IROs

I selected my IROs. What is the next step? Stakeholder engagement: add your stakeholder groups, choose a method per group (survey, interview, focus group) and assign the questions each group should answer. Start with Plan your stakeholder engagement.

I still have more than 100 IROs. Is that too many? For most companies, yes. Every selected IRO is a question for someone. Three ways to reduce the burden:

1. Select fewer IROs.
2. Do not assign every question to every group; see Assign questions and weights.
3. Cap the number of questions per survey respondent.

A staged approach also works: interview a few internal experts first, drop the IROs they score lowest, then engage the other groups on the rest.

How do I decide which questions to leave out for a group such as the board? Ask yourself whether the group can judge the matter and has a stake in it. The board can judge strategy, finance and governance; it usually cannot judge working conditions at a tier-2 supplier. This is a judgment call. Ask us through the chat for a session if you want to go through it together.

Should I meet my auditor before sending surveys? Yes, if your DMA will be audited. A one-hour online meeting after IRO selection and before data collection is enough. Prepare how you selected the IROs, which subtopics you consider not material by definition and why, and which stakeholders you will engage with which methods. Karomia joins on request to answer technical questions. Start data collection once the auditor gives the green light.

I am not confident in my IRO list. How do I get it right before I engage stakeholders? No IRO list is 100% relevant before the DMA: deciding what is relevant is the point of the exercise. Two things help:

• Share your feedback with us per IRO or per topic. We fine-tune the list with you.
• Run two or three test interviews with colleagues first. Tell them they are testing the questions. Their reactions show you which IROs and questions need work before you send them to a wider group.

How do I know when my IROs are ready? You receive a notification by email when the IROs are generated, typically within a few days of submitting.

Editing, moving and translating IROs

Do I need to delete the IROs I disagree with? No. Just do not select them. Only selected IROs go to your stakeholders. Delete an IRO only if you want it out of the list entirely.

Can I move an IRO to another subtopic? Not directly. Create the IRO in the right subtopic and delete the original. For many IROs, ask us for an Excel export, make the changes there and send it back.

Can I correct a label, for example direct instead of indirect? Yes. Open the IRO, click Edit and change the label.

Do I have to translate a new IRO myself in every language? Write it in your default language and use Fill all with translations, or ask us to translate the whole assessment once your list is final. See Review questions and translations.

Can I change the wording respondents see? The IRO title, question and description are yours to edit. The fixed parts of the survey, such as the scoring tooltips and the guiding questions per criterion, are the same for all assessments. Send us your suggestions. If you need your own scoring definitions, we can host a scoring guidance page for your respondents and link it from the survey introduction.

Can I use quantitative values, for example euro bands, for financial magnitude? Not as a built-in option. Respondents score from 1 to 5. Share your own bands with them in the invitation so they map their answer to the scale. Talk to us before your data collection if this matters for your audit.

Can I add an IRO after the surveys are out? Avoid it. The DMA runs in order: you validate the IRO list, then you engage stakeholders. An IRO added during data collection is assigned to groups that have already answered, so their responses no longer match the list; the respondents who finished are never asked the new question, so it ends up scored by a smaller population than the rest; and it has to be translated into every language of the assessment before anyone sees it. If a matter surfaces during an interview or a focus group, note it, finish the round and add it in your next cycle. If the change is unavoidable, contact us through the chat first so we make it with you and record what changed.

Our stakeholders need the survey in another language. Add the language to the assessment and translate the questions before you share the links. If the language is not in the list, ask us.

Exports

Can I export all IROs, including the ones I did not select? Yes. Use Export in the IRO table. The export contains every IRO with its type, title, question, description, labels and whether it is selected. If you need extra columns or another language, ask us.

Can I share a list of our selected IROs with an investor or a board before the DMA is finished? Yes. Export the IRO list and filter on selected IROs.

Can I download the materiality matrix in an editable format? The matrix data is in the Excel export (sheet Materiality summary), so you can rebuild the chart in Excel or PowerPoint. If you need it ready-made, ask us.

Can I see the results per stakeholder group? Not as a filter in the matrix yet. The Excel export has every answer with its stakeholder group (sheet All results) and the score per group per IRO (sheet Stakeholder aggregation). Ask us if you need a summary per group.

The subtopic matrix has overlapping labels. What do I use in my report? Use the topic-level matrix and the subtopic table, or the visuals delivered with your DMA report.

Scoring, weights, results and matrix

Can I calculate results before all answers are in? Yes. Click Calculate results whenever you want an interim view, and again after new answers arrive. See Monitor and close.

I changed a topic weight and nothing moved. Why? Topic weights are relative between stakeholder groups. If only one group answered a question, its weight has nothing to be compared with and the result stays the same. A change also does not move a topic if the subtopic you affected is not the highest-scoring one in that topic. To reduce the influence of specific answers within a group, lower the expertise level of those respondents.

What do the yellow and blue colours in the results mean? Yellow is a negative impact or a risk, blue a positive impact or an opportunity. The colour says nothing about materiality. The scores and thresholds do.

How do I see which IRO type made a subtopic material? The results table shows the score per subtopic for each type: positive impact, negative impact, risk and opportunity. The type above the threshold is the one that makes the subtopic material.

Why can a respondent only rank three topics as high in the survey? In personalized surveys, respondents first rank topics as high, medium or low. The limit forces them to differentiate, so they get questions on the topics they know best. If you prefer every respondent to see all their questions, use randomized selection.

How is the materiality threshold set, and why is it exactly 3? The threshold is the average of your topic scores on that axis per IRO type - each topic contributes its highest positive and its highest negative impact score separately on the impact axis, its highest risk and its highest opportunity score on the financial axis - with a floor: if the average falls below 3, the threshold is 3. So a threshold of exactly 3.00 means your average came out lower and the floor applied. EFRAG treats the midpoint of a 1 to 5 scale as the lowest reasonable bar for calling a topic material, which keeps a conservatively scored assessment from ending up with a bar so low that everything passes it. The comparison uses the normalized score, and a score exactly at the threshold is material. See How results are calculated.

What can I configure in the results? Topic weights per stakeholder group and the expertise level of respondents, yourself. Thresholds, the undecided margin and the materiality conclusion per subtopic, through us once your data collection is finished. Always refresh the results before you export.

Do I decide undecided results at topic or subtopic level? At subtopic level. A topic is material as soon as one of its subtopics is. It stays undecided while at least one of its subtopics is undecided and none is material. See Decide on undecided topics.

Can I delete test answers from a survey? Yes. Use Delete answers for the group. The stakeholder group and its setup stay.

Can someone answer through both a survey and an interview? They can, but they do not need to. One answer through one method per person is enough.

Many respondents dropped out halfway through the survey. What can I do? Shorten the survey, use the easy complexity level for external groups, and recommend the overall score instead of the breakdown to people with little time. See How to send surveys.

What should I say in the welcome video? Keep it short and personal: who you are, why you ask, how long it takes. For example: "Hello, I'm ... I'd like to ask ten minutes of your time for a short survey. Sustainability matters a lot to us, and your feedback helps us focus on the right topics. Thank you!"

Assurance and the link with ESRS

What will the auditor ask about our DMA? The questions we see most: Did you map your value chain? How are IROs linked to their position in it? How did you determine actual versus potential and the time horizon? How were IROs scored? Were silent stakeholders taken into account? Why did you leave out certain subtopics or IROs? Is your stakeholder sample representative? Who signed off? How to prepare for the audit lists each one with the evidence that answers it.

Can the auditor access our documents in Karomia? Give the auditor access to the platform as a member, or share the DMA report, which lists every document used, and the Excel export.

Can I get the list of ESRS disclosure requirements that follow from our DMA (ESRS 2 IRO-2)? Yes. When you link your DMA to an ESRS report in Karomia, the material topics define the report scope. Ask us if you need the list of disclosure requirements as a separate file.

We also want a VSME report. Can we start it in parallel? Yes. Create a report from the Reports module. If you do not see the option, ask us to open it for you. The DMA results can be included in your report once they are final.

When will I receive the DMA report? After you close data collection and confirm your decisions on undecided subtopics. Count a few days to two weeks. The quantitative part can come first and the qualitative analysis per topic after.

Can I still add a document after data collection? For the IROs, no. For the qualitative analysis in your report, possibly: ask us before the report is generated. Otherwise use it for your sustainability report.

Will the report reflect IROs I deselected after it was generated? No. The report reflects the assessment at the moment it was generated. Ask us to regenerate it after changes.

Troubleshooting

An interview shows a blank screen. A session that stays open for about two hours is disconnected for security. Log out, log back in and start a new session for the remaining questions. Ask us to merge the two sessions.

I cannot save an IRO or its description. Refresh the page and try again. If it still fails, send us the IRO title and the language through the chat. Saving errors have been linked to a specific language version.

My results disappeared from the screen. Refresh the page. Your results are stored and not lost. If they are still not visible, contact us.

Support

Where do I find this guide in the platform? Through the question mark in the platform, which also opens a chat with the Karomia team. For audit conversations, ask us to join.