CSRD Software vs Consultant: What a Platform Can't Do
A vendor-neutral look at what CSRD software handles, where you still need people, and which setup fits your company: software, a fractional lead, or both.
João Aguiam
· 7 min read
A vendor demo went well. The platform collects the data, runs the workflow, tags the report. Then your CFO asks the obvious question: if the software does the whole report, why are we paying a consultant?
It is a fair question, and almost every answer you will find online comes from someone who sells software. This one comes without a product attached. Software and people do different jobs in a CSRD report. Where one stops and the other starts decides how much you spend and how well the report holds up when the assurer reads it.
What CSRD software actually does
A reporting platform is good at the work that is repetitive, structured and easy to lose track of.
- Collection. Role-based data entry for sites, business units and subsidiaries, with units and conversions handled for you. This is where manual spreadsheets break first once you have more than a handful of contributors.
- Workflow. Who owes what by when, who reviewed it, who approved it. Reminders, sign-offs and deadlines in one place instead of an inbox.
- Audit trail. A record of who entered which number, from which source, and what changed after. This is the part assurers care about most, and it is hard to reproduce by hand.
- Reporting output. Populating ESRS disclosure templates, cross-checking datapoints, and producing the tagged file the regulator expects. For how digital tagging works and what to look for in tooling, see CSRD XBRL and digital tagging.
For companies with several entities, multiple sectors and hundreds of datapoints, this is real value. The data collection gap analysis shows how fast manual collection becomes unmanageable.
Notice what is on the list: moving, storing and checking data that already exists, or that someone has already decided to collect.
What it cannot do
A platform has no opinion about your business. Four things stay with people.
Double materiality judgement. Software can send the survey and tally the scores. It cannot choose where the threshold sits, decide whether an impact is material, or explain why a topic was left out. Those decisions define your whole report, and the assurer will ask about them. See double materiality assessment for what a defensible one looks like.
Stakeholder interviews. Real conversations with employees, suppliers, customers and communities surface things no questionnaire does. Someone has to run them, listen, and turn them into findings.
The assurance conversation. When the assurer challenges a judgement, a person has to answer. "The platform calculated it" is not an answer. Someone from your team, or someone you brought in, needs to explain the method, the assumptions and the evidence.
The narrative. Policies, actions, targets and the story that connects them are written, not generated. The text has to match what the company actually does, and it has to be consistent with the numbers.
None of this is a flaw in the software. It is the difference between a system that holds the report and a team that makes it.
Three setups and who each fits
Most companies end up in one of three arrangements. The right one depends on your profile, not on how good the demo was.
| Setup | What it is | Who it fits |
|---|---|---|
| Software only | Platform plus your internal team, no outside advisor | A VSME reporter, or a small in-scope company with clean data, a simple structure, and someone internal who already knows the standards |
| Software plus a fractional lead | Platform, with an outside expert a day or two a week who owns method and judgement | Most mid-sized groups. A company doing a second cycle. Teams with data capacity but thin ESRS experience |
| Consultant-led, with software | An advisory team runs the programme and implements the platform as part of it | A first double materiality assessment, a messy group structure, a first assured report, or a tight deadline with little internal capacity |
Read the table from your situation, not from the budget. A company that picks software only when it needs a first DMA ends up paying for the gap later, usually in the weeks before the assurer arrives.
The fractional model is covered in CSRD consultant engagement models. The wider question of outside expertise versus a permanent hire is in CSRD consultant vs in-house. This post is only about tool versus people.
The cost comparison vendors leave out
The licence is the visible line. It is rarely the biggest.
- Licence. Platform subscriptions are commonly quoted anywhere from about €5,000 to €100,000+ per year, depending on size and modules. Treat that as an indicative range, not a price.
- Implementation. Configuring the data model, mapping your entities, integrating sources and training users. Some vendors charge for it, some bundle it, some leave it to a partner.
- Internal hours. Often the largest line and the one nobody budgets. Someone on your team has to configure, populate and maintain the platform, and chase contributors.
- Advisory days. Senior independent CSRD experts typically bill €1,200–€2,500 per day. A fractional lead at one day a week typically costs €4,000–€12,000 per month. A materiality assessment for a smaller company runs roughly €10,000–€25,000.
The comparison that matters is not licence against consultant. It is the total of licence, implementation and your own hours, against what advisory days would have cost to do the same job, and what each leaves undone. Software removes manual work. It does not remove the need for decisions, and unmade decisions are expensive when found late. For ranges by company size, see CSRD consultant costs.
A practical test: ask the vendor to list, in writing, what your team must still provide and decide. If the list is short, ask why.
Choosing in the right order
Buying the tool first is the most common sequencing mistake. A platform embeds a data model. If you choose it before you know what is material, you lock in the structure for questions you may not need to answer, and you leave out ones you do.
A better order:
- Scope. Which entities, which reporting boundary, and which rules apply to you after the Omnibus. In scope now means more than 1,000 employees and more than €450M turnover, with limited assurance only.
- Double materiality. Decide which topics and datapoints are actually required. This is the output that drives everything else.
- Data model. Map those datapoints to owners, sources and systems. This is where gap analysis earns its time.
- Tool. Now compare platforms against a concrete requirement list. You will find the shortlist is shorter and the demos are easier to judge.
Companies that start with a spreadsheet for year one and move to a platform once they know what works are not behind. They often have a better data model than those who bought first.
Finding someone who has implemented your tool
If you already have a platform, or have narrowed to one, the most useful advisor is someone who has deployed it before. They know where the configuration is awkward, what the vendor's template misses, and which datapoints need workarounds.
When browsing the expert directory, filter for tooling and data management experience, then ask each shortlisted expert three questions: which platforms have you implemented, in what scope, and what did you change from the default? A consultant who has done it will answer with specifics. One who has only attended demos will not.
Independence matters here too. An advisor who is paid by a vendor has a reason to recommend that vendor. Ask about commercial relationships before you accept a recommendation.
The short answer for your CFO
Software replaces work, not decisions. If your reporting is simple and the judgements are settled, a platform may be most of what you need. If it is your first DMA, a complex group, or a first assured report, the consultant is paying for the part software cannot do, and for making the software worth what you paid for it.
