25 Questions to Ask a CSRD Consultant — and the Answers That Should Worry You

Every CSRD consultant you speak to will tell you they have deep ESRS expertise and a proven methodology. They will all say it. That's the problem with the selection conversation most companies run: the questions are general enough that a generalist who read the EFRAG guidance last month answers them exactly as well as a practitioner who has filed nine reports.

The fix is not more questions. It's questions where a weak answer is visibly weak — specific enough that someone who hasn't done the work can't improvise their way through.

Below are 25 of them, grouped by the conversation they belong in, each with the answer you're hoping for and the answer that should make you keep looking. Take the ones that fit your situation; nobody needs all 25 in one call.

How to Structure the Conversations

Three conversations, not one.

1. The screening call (30 minutes). Two or three people, early questions only. You're filtering, not evaluating. Half your shortlist should not survive this.

2. The technical deep dive (90 minutes). Bring your finance lead and whoever owns the data. Insist that the consultant who would actually run your engagement is on this call — not just the partner selling it. This is where questions 6 through 20 belong.

3. References (30 minutes each, at least two). Do not skip this. It is the single highest-signal step and the one companies most often drop when they're behind schedule.

A note on how to listen: you're not grading the polish of the answer. You're listening for whether it contains specifics that could only come from having done the work — a client's actual constraint, a decision that went badly, an argument they lost with an auditor.

ESRS Depth (Q1–5)

1. "How many ESRS-aligned reports have you taken all the way to filing?"

Strong: A number, with context — sectors, company sizes, which parts they owned versus advised on. "Four filed, two more in draft; I led materiality on all six and drafted disclosures on three."

Worrying: Firm-level credentials substituted for personal ones — "our practice has supported over 200 sustainability clients." Ask again about the individual.

2. "Which ESRS topical standards do you personally find companies get wrong most often, and why?"

Strong: An opinionated, specific answer. Something like ESRS E1 transition plans being aspirational rather than costed, or S1 own-workforce metrics being reported without the required coverage.

Worrying: A recitation of the standards list. Anyone can name E1 through G1.

3. "Walk me through a datapoint you concluded was not material and had to defend."

Strong: A real example with the reasoning and the evidence trail — who challenged it, what documentation supported the conclusion.

Worrying: "We generally recommend disclosing everything to be safe." That's not a methodology, it's an absence of one, and it will cost you a great deal in data collection.

4. "How do you handle a disclosure requirement where the guidance is genuinely unclear?"

Strong: A described process — check EFRAG Q&A, look at peer practice, document the judgement and its basis, flag it to the auditor early rather than late.

Worrying: Certainty. Anyone who says nothing about the ESRS is ambiguous has not spent enough time in the detail.

5. "What has changed in the last six months that affects our project?"

Strong: Current, specific awareness — the Omnibus changes to scope and timing, and what they mean for your wave specifically.

Worrying: Vagueness, or worse, advice that's clearly built on a superseded timeline.

Double Materiality Method (Q6–10)

6. "Walk me through your double materiality methodology, step by step."

Strong: A described sequence with thresholds — how impacts and financial effects are scored, what scale is used, where the materiality line sits and how it was set.

Worrying: "We run a workshop and agree the material topics." Workshops are an input, not a method.

7. "How do you set the materiality threshold, and who signs it off?"

Strong: Clear that the threshold is a management decision, documented, with the consultant facilitating rather than deciding.

Worrying: The consultant owns the threshold. That's your board's judgement to make and defend.

8. "How do you identify IROs in the value chain when we have limited supplier data?"

Strong: Proxy methods, sector benchmarks, spend-based screening first and primary data later, with an explicit plan to improve.

Worrying: Either "we'll survey all your suppliers" (unrealistic) or "we'll estimate it" without saying how.

9. "Which stakeholders do you insist on engaging, and what happens if we say no?"

Strong: A defended minimum — typically employee representatives and material value chain partners — and a willingness to push back on you.

Worrying: Total flexibility. A consultant who agrees to everything you want will not protect you from the auditor.

10. "How does the materiality output connect to what actually gets disclosed?"

Strong: A traceable chain from IRO to datapoint, usually a register or matrix that maps each material topic to specific disclosure requirements.

Worrying: The materiality assessment is described as a standalone deliverable. Then it's a slide deck, not a foundation.

Data, Systems and Scope 3 (Q11–14)

11. "How do you run a data gap analysis, and what do we get at the end?"

Strong: Disclosure-level, not topic-level: every required datapoint, current source, owner, quality, and the gap. A populated register you keep.

Worrying: A RAG-status summary by topic. Too coarse to act on.

12. "What's your approach to Scope 3 when our supplier data is poor?"

Strong: Start with spend-based estimation for screening, prioritise the categories that dominate the footprint, replace with primary data over several cycles. Realistic about how long that takes.

Worrying: Promises of accurate primary-data Scope 3 in year one across all fifteen categories.

13. "Do you work with our existing systems or do you require a platform?"

Strong: System-agnostic, comfortable starting in spreadsheets with proper version control, and honest about when software genuinely pays for itself.

Worrying: A software licence bundled into the proposal without discussion. Ask directly whether they receive commission.

14. "How do you handle XBRL tagging?"

Strong: Either in-house capability, described concretely, or a named partner and a clear handoff.

Worrying: "We'll deal with that at the end." Tagging is a scope line with real cost; if it's not in the proposal, you'll pay for it later.

Assurance and the Auditor (Q15–17)

15. "What will our auditor ask for that we won't have?"

Strong: A concrete list from experience — evidence for estimates, methodology documentation, calculation audit trails, sign-off records for judgements.

Worrying: "We'll get you audit-ready." Meaningless unless they can say what ready means.

16. "Can you also provide our assurance?"

Strong: An immediate no, with an explanation of the independence rules. The firm advising on your CSRD implementation cannot also assure it.

Worrying: Any answer other than a clear no. This is a basic professional boundary, and not knowing it tells you a lot.

17. "How do you handle it when the auditor disagrees with a judgement you recommended?"

Strong: A real example. Sometimes they held the position with documentation, sometimes they conceded.

Worrying: "That hasn't happened." It happens to everyone who has done this more than twice.

Who Actually Does the Work (Q18–20)

18. "Who exactly will be on this engagement, and what percentage of their time do we get?"

Strong: Named people, seniority mix, and honest allocation. "I'm on it two days a week, my analyst three."

Worrying: "We'll confirm the team at kickoff." This is where bait-and-switch lives.

19. "Will the people on this call be doing the work?"

Strong: Yes, or a straight explanation of who else and why.

Worrying: Hesitation. Ask for CVs of the delivery team, not the pitch team.

20. "What happens if your lead consultant leaves mid-engagement?"

Strong: Documented handover, a named backup who's already close to the work, continuity commitment in the contract.

Worrying: No answer. For an independent consultant this is a genuine risk, and the good ones acknowledge it rather than waving it away.

Scope, Change Control and Exit (Q21–25)

21. "What is explicitly excluded from this price?"

Strong: A clear exclusions list offered without prompting — tagging, assurance support, translations, restatements, additional entities.

Worrying: "Everything's included." Nothing is ever everything. Get the list in writing.

22. "What triggers a change order, and how are they priced?"

Strong: A defined threshold, an agreed day rate for additional work, written approval required before work starts.

Worrying: Informality. This is the mechanism by which a €60,000 engagement becomes €110,000.

23. "What do we own at the end?"

Strong: You own the deliverables, the data, the populated templates and the documented methodology, with no licence needed to reuse them.

Worrying: Proprietary tooling you lose access to when the engagement ends. You've rented your own materiality assessment.

24. "What does knowledge transfer look like, specifically?"

Strong: Named deliverables — documented methodology, training sessions, a maintained disclosure register, your team leading a workstream by the end.

Worrying: "We work closely alongside your team." That's a working style, not a deliverable, and it isn't contractible.

25. "What does year two look like if we don't re-engage you?"

Strong: A straight answer about what your team would need to do, and an acknowledgement that self-sufficiency is the goal.

Worrying: An answer that assumes you'll obviously re-engage. Some consultants build dependency deliberately; the good ones are slightly working themselves out of a job.

Three Technical Screens

Questions can be prepared for. Scenarios are harder to fake, and fifteen minutes of these will tell you more than an hour of credentials. Give the consultant a short, real situation from your own company and watch how they reason — you're grading the questions they ask back, not the answer they land on.

Screen 1 — the contested topic. "Our operations team says water is not a material topic for us. Two of our three largest customers ask about it in their supplier questionnaires. What do you do?"

A practitioner separates the two things immediately: whether water meets the impact or financial materiality threshold under ESRS is a different question from whether customers are asking about it, though customer pressure may itself be evidence of a financial effect. They'll ask about your sites, water stress in those locations, and what the customers actually do with the answers. A weak answer picks a side straight away.

Screen 2 — the missing data. "We have no reliable data for a required datapoint, and our filing is in five months. What now?"

You want to hear about estimation with documented methodology, transparency about the basis, a stated plan to improve, and an early conversation with the auditor — not silence, and not a fabricated number. If they suggest simply omitting it without discussing the disclosure requirements around omission, that's a real problem.

Screen 3 — the uncomfortable finding. "Halfway through, your analysis shows our published emissions target is not credible against our actual capex plan. How do you handle that?"

The answer you want involves telling you, early and privately, with options — restate the target, revise the transition plan, or disclose the gap honestly. A consultant who talks about finding a presentational fix is telling you exactly how they'll behave when it matters.

None of these have a single correct answer. That's the point — you're testing whether they think like someone who has been in the room when it went wrong.

The Reference Call

Ask the consultant for two or three clients from comparable engagements. Then ask those clients:

1. "What did they do that you didn't expect — good or bad?" The open question first, before you anchor them.
2. "Did the people who pitched actually do the work?"
3. "Did the final invoice match the original proposal? If not, why?"
4. "When something went wrong, how did they handle it?" If the answer is "nothing went wrong," gently push — every project has something.
5. "Could your team run the next cycle without them?"
6. "Would you hire them again for the same scope?" Listen to the pause before the answer, not just the answer.

If a consultant can't produce references from comparable work, that isn't automatically disqualifying for someone newer — but it should change the shape of the deal. Start with a paid pilot such as a gap analysis, and decide from the output.

A One-Page Scorecard

Score each shortlisted consultant 1–5 across these, with a written note per line. Score independently before anyone in the room discusses candidates.

Criterion	Weight
Demonstrated ESRS depth (Q1–5)	25%
Materiality methodology (Q6–10)	20%
Data and systems realism (Q11–14)	15%
Assurance experience (Q15–17)	15%
Named team and continuity (Q18–20)	15%
Commercial clarity and exit (Q21–25)	10%

Deliberately, price isn't on this list. Price belongs in a separate comparison — see how to compare CSRD consulting proposals — because mixing capability and cost into one score is how companies talk themselves into the cheapest bid.

Two rules for the scoring meeting. Any candidate scoring 2 or below on ESRS depth or named team is out regardless of total, because those two can't be compensated for elsewhere. And if two candidates finish within a few points of each other, the reference calls break the tie, not the pitch.

Find Consultants Worth Interviewing

None of this helps if your shortlist is three firms that came up first in a search. Build a wider one.

Browse the CSRD Experts directory to compare consultants and firms by country, expertise and industry, and put four or five in front of these questions rather than two.

For the wider process around these conversations, see how to hire a CSRD consultant; if your engagement is large enough to warrant a formal process, the CSRD consultant RFP template covers how to structure it.