The use of artificial intelligence (AI) and ESG-software in the Sustainability management is a topic that is becoming increasingly important - especially for medium-sized companies that are struggling with the challenge of meeting the sometimes complex requirements of the Sustainability reporting (e.g. according to CSRD/ESRS and VSME) and ESG compliance. But how useful is the use of AI in practical implementation? What real potential does it offer and where is it possibly overrated?
To answer these questions, we interviewed four providers of ESG software tools: aitark, COBACK, Dina and ESGbot.
They shared their experiences and perspectives with us on the advantages and disadvantages of using AI in sustainability management. The topics covered included Data analysis, Automated reporting, the division of labour between man and machine and the Data security and the ecological footprint from KI.
Read on to find out what practical insights ESG tool providers have on the use of AI in sustainability management - and what companies should consider when implementing it.
Contents
- Brief presentation of the ESG software providers surveyed
- 1. potential and overvaluation of AI in sustainability management
- 2. ideal division of labour between AI and humans in sustainability management
- 3. time savings through AI in sustainability management
- 4. need for training and expertise for the use of AI in sustainability management
- 5. information security and legal requirements in the context of AI use
- 6 The ecological footprint of AI
- 7 AI as a strategic tool in sustainability management
- Summary
Brief presentation of the ESG software providers surveyed
Brief description:
„We make sustainability management simple. aitark helps companies and consultancies to collect and structure ESG data and create reports in minutes instead of weeks. With an interface that everyone understands, clear templates and automatic emissions calculation, sustainability work becomes intuitive - not complex. One system for everything you need. Upload, review, report creation - and soon: answers and automation via AI chat. We are building the future of sustainability management: simple, accessible and ready for any new challenge.“
Website: aitark.io
Brief description:
„COBACK is a platform for sustainability reporting in SMEs. We help companies to efficiently fulfil complex regulatory requirements such as CSRD/VSME through a unique combination of training, automation and data input. While traditional software offers little training and alternatives are often costly, COBACK empowers employees to become sustainability experts themselves. This ensures compliance with regulations while safeguarding long-term competitiveness. With the COBACK Finder, we have also developed the first B2B comparison platform for sustainability start-ups, creating a broader ecosystem for Sustainability and compliance with regulatory requirements in the SME sector.“
Website: COBACK.ai
Brief description:
„Structured ESG management - understandable, comprehensible, module-based: Dina creates a digital workspace for the sustainability management of companies. The Saas software solution guides you simply and in a structured way through a well-founded materiality analysis, supports the derivation of your Sustainability strategy, Dina is a tool that creates your corporate carbon footprint and makes it easier for you to manage and control your ESG data. In this way, Dina supports the strategic anchoring of sustainability and preparatory reporting in the Deutscher Nachhaltigkeitskodex (DNK) according to VSME or ESRS.“
Website: Dina-Tool.de
Brief description:
„The ESGbot is an AI-based solution that provides comprehensive knowledge on sustainability reporting and other ESG laws and documents. By asking specific questions, the ESGbot helps users to understand and learn the legal requirements as well as to implement and process them in practice.“
Website: ESGbot.com
1. potential and overvaluation of AI in sustainability management
We asked the ESG software providers the following question:
„From your observations: Where is the greatest potential for the use of AI in the sustainability transformation of SMEs? And vice versa: Where is AI overrated or used too enthusiastically by companies?“
The answers briefly summarised:
ESG software providers see the greatest potential of AI in sustainability management in data analysis and structuring, especially when company data such as supply chain or product information is incorporated. But be careful not to overestimate the potential: AI is no substitute for experts. While it can analyse data quickly, it still requires human expertise to draw the right conclusions and make strategic decisions.
The answers in full:
„We see the greatest potential in context-related use: in other words, when the AI knows what the task is and when it does not rely on the knowledge of the AI, but is provided with data or content from the company - i.e. context - with which the AI can complete the tasks set. In the context of sustainability: data and content from the company, the products and the supply chain are known and are analysed by the AI. This is how the best results are achieved and how companies can best move forward.
AI is overrated when it comes to questions on a wide range of topics where no context is given or no specialist knowledge is available. The answers are often correct, but not always. This means that AI does not replace experts.“
„To be honest, we see the greatest potential of AI for SMEs primarily in relieving all this bureaucratic frustration. SMEs often simply don't have the resources to trudge through hundreds of pages of legal texts. AI is a great „chaos sorter“ that automatically sorts unstructured data from invoices or Excel lists into the right logic for standards such as VSME. It is also an ingenious error checker: it immediately recognises when data is illogical before you embarrass yourself with the auditor. In essence, AI democratises expert knowledge so that even people without an ESG background can produce sensible sustainability reports, for example.
On the other hand, AI is often overrated as a kind of „magic“. Many people think that you press a button and the finished report comes out, but this is more of an AI placebo. An AI cannot conjure up data that simply does not exist. For example, if you don't know your Scope 3 values in the carbon footprint, the AI will only give you nicely calculated estimates that will immediately fail a real audit. Responsibility cannot simply be passed on either. In the end, it's still the human, not the machine, who is liable. We also see far too much hype about little chatbots that answer questions nicely but don't solve the actual problem.“
„We see great potential for AI in sustainability management, especially when it comes to analysing data. A good example is the evaluation of transport information for the annual carbon footprint: this is often only based on random samples because the detailed information is available but cannot be analysed systematically - especially in the case of business models with high transport volumes. AI can automatically analyse the data here, reducing the workload and at the same time significantly improving the data quality and informative value of the carbon footprint. This principle can be applied to all areas with large, unstructured volumes of data.
But we have to be careful: Many medium-sized companies are not yet ready. In many cases, generative AI is not yet used as an assistant. This means that comprehensive AI-based data analysis is still a long way off.“
„In our view, the greatest potential of AI currently lies in the structuring and translation of regulatory complexity. Companies are faced with a wide range of ESG requirements, from CSRD and ESRS to EU taxonomy, CSDDD and EUDR through to industry-specific requirements. These regulations are extensive, fragmented and sometimes difficult to access. This is precisely where AI can create real added value: it accelerates the identification of relevant standards, summarises content in a comprehensible way and helps to systematically categorise requirements.
AI is overrated where it is seen as a substitute for strategic sustainability management. AI can process information, but it does not assume responsibility. It does not replace governance structures, risk assessment or corporate prioritisation. Anyone who sees AI as a panacea fails to realise that sustainable transformation will always remain a leadership and cultural issue.“
Are you planning the next steps towards sustainability?
Ask me for a free information meeting.
I am ready with advice and pleasure.
Matthias Damert
Sustainability, climate and energy management
2. ideal division of labour between AI and humans in sustainability management
We asked the ESG software providers the following question:
„In your opinion, what should the ideal division of labour between AI and human expertise in sustainability management look like? Where do you see clear „red lines“ - i.e. areas where AI should not be used?“
The answers briefly summarised:
ESG software providers largely agree on the division of labour between AI and human expertise in sustainability management: AI should be seen as an assistance system. It takes over the collection and sorting of data, while human expertise remains responsible for interpretation and strategic decision-making. There are red lines, particularly when it comes to responsibility for decision-relevant issues. AI cannot make strategic decisions and must not be used as a substitute for human judgement. Its task is to provide support, but not to assume responsibility
The answers in full:
„In our opinion, the perfect division of labour is that the AI prepares data and provides insights based on the information that is made available to it (again, this has already been done by experts or with the help of tools so that nothing is left out) - but these should then be checked again by experts for plausibility. So we see the work of the experts more in the interpretation, i.e. the „acceptance“ of the results of the AI.
We see red lines with „only rely on the AI“. AI must also be operated correctly.“
„We believe the ideal division of labour looks like this: The AI does the dirty work. In other words, collecting, sorting and checking huge amounts of data so that everything fits together logically in the end. The human is then more of a strategist and decision-maker. AI prepares the playing field and builds the framework for the sustainability report, for example, but the actual assessment of which goals really make sense for the company and how to implement the transformation in practice must come from humans. AI is great at oversimplifying complexity, but it has no sense of corporate culture or long-term vision.
We draw clear red lines where real responsibility and strategic judgement are involved. An AI should never make the final decision on the Materiality of topics or even assume legal liability for a sustainability report. That's the end of it. Even when it comes to sensitive topics such as ethics or direct communication with stakeholders, AI should not be blindly left in the driving seat. Compliance is ultimately a matter of trust between people. Technology can support and automate this, but it must never be the sole sender of the message.“
„We see AI as an assistant, not a decision-maker. In sustainability management in particular, AI can be used to analyse data and recognise patterns. It can take on monitoring functions or regulatory screening, but benchmarking and market analyses can also be generated with the help of AI.
We believe that the strategic categorisation, weighing up conflicting objectives or deciding on materiality is a human responsibility. In our view, the interpretation and evaluation of AI-generated results should be carried out by human expertise - if only because this is where the responsibility for decisions lies.“
„AI should be seen as an assistance system. It can simplify research, structure content, generate suggestions and make regulatory requirements transparent. For me, there is a clear principle here: AI supports MI - artificial intelligence supports human intelligence. The decision as to how information is evaluated, prioritised and strategically used must remain with humans.
We see red lines particularly in automated assessments with significant implications, such as human rights supply chain risks or statements with compliance relevance. ESG management requires comprehensible decisions. AI should provide support here, but not autonomous assessments.
A key point is transparency: companies must be able to understand the sources on which statements are based. Generic AI models often provide plausible answers, but without reliable sources. This is problematic in the ESG context because it involves regulatory requirements that need to be interpreted precisely.“
3. time savings through AI in sustainability management
We asked the ESG software providers the following question:
„In which specific ESG processes can AI demonstrably save time today - and where are the limits? Which promises of AI efficiency gains can be kept, and which are more like marketing?“
The answers briefly summarised:
ESG software providers agree that AI saves time, especially in data-intensive processes such as data preparation and reporting. AI can achieve major time savings, particularly when identifying standards or compiling sustainability reports. But the limits of the efficiency gains are clear: AI is a tool for data processing, but it cannot guarantee the quality of the subsequent decisions. Where primary data is missing or qualitative assessments are required, it reaches its limits. The desire for a „one-click“ report often remains a marketing promise that does not cover the complex requirements of a sustainability report.
The answers in full:
„We see the greatest potential for time savings above all in data collection, analysis and visualisation - but ideally in combination with software or an appropriately designed agent workflow, so that everything remains reconstructable/traceable.
When collecting data, however, you have to be careful that you can't (yet) throw everything in and it comes out correctly allocated. This requires an appropriately designed workflow in the background and should still be treated with caution.“
„In terms of saving time, the biggest lever today is definitely all the data mapping and categorisation. It used to take weeks to manually assign hundreds of invoices or Excel lists to the correct ESG key figures. Today, an AI does this in minutes and easily saves over 50 % of preparation time. It can also be enormously helpful when analysing gaps. The AI scans the documents and tells you immediately where evidence is still missing for a certain criterion, instead of you having to tediously tick off checklists. And, of course, it can be very helpful when drafting the first sustainability report itself. The basic framework for the report is created in no time at all, which takes away the fear of the blank page and translates the regulatory jargon directly into readable texts.
But you also have to be honest: a lot of it is currently pure marketing. This „one-click“ report, where you supposedly just press a button and everything is ready, is unfortunately still a fairy tale. ESG compliance is not just a maths problem, it always needs a human context. The fully automated Scope 3 calculation also promises a lot. Without real primary data from suppliers, this is often more like „guessing at a high level“. For us, the limit is always where liability and interpretation are involved. An AI can make data consistent, but it cannot take responsibility for it. If you trust blindly, you're more likely to create a compliance risk instead of solving it.“
„We see the greatest benefits of AI in data-intensive processes, such as collecting and processing data, recognising patterns or preparing reports. This definitely saves time and increases data quality.
Where are the limits? AI cannot decide which aspects of sustainability are really important or which measures are ethically justifiable - that always requires human expertise. AI is simply an assistant, not a decision-maker.“
„AI saves time, especially in operational, structured processes: Identification of relevant laws and standards, pre-structuring of sustainability reports, initial assessment in the materiality analysis, consolidation of regulatory requirements.
The limits lie where data is missing or qualitative assessments are required. AI can bring content together, but it is no substitute for in-depth analyses of the business model or the Stakeholders-perspective.
In my view, promises that AI can „automatically make you CSRD-compliant“ are not tenable. AI can speed up processes and create clarity, but it remains the company's responsibility to analyse the content of the strategy, risks and effects.“
Methods and best practice for sustainability in your mailbox
4. need for training and expertise for the use of AI in sustainability management
We asked the ESG software providers the following question:
„Do employees need specific expertise or training to reap the full benefits of AI in sustainability management?“
The answers briefly summarised:
The providers agree that training for employees is essential. It is not about teaching technical AI knowledge, but about understanding data quality, process logic and responsibilities. AI is a helpful tool, but the added value only comes when employees understand the data properly and know how to make corrections and take responsibility for the results. Training on how to use AI correctly and how to critically review the results is the key to successful use in sustainability management
The answers in full:
„If employees use AI chat bots such as ChatGPT, Gemini or similar, they should know how the AI (in this case LLMs) works. There are several factors to consider, such as What happens to the data? Where is it stored? Is it used for training purposes? So data protection is a big issue, but also how the AI works, what it can do well, what it can do less well. Then they should also know how to build their prompts in order to achieve the best results. And which work steps the AI can do or where it makes more sense and is more efficient to do it yourself. If they want to realise the full potential for increasing the efficiency of repetitive processes, they should be trained in how to use and set up agents and workflows.
Special AI-supported software takes over precisely these parts - it provides a user-friendly interface so that work becomes more efficient and the AI provides support in the background without having to be specially set up, because everything was already thought through when the software was created so that untrained people can handle it without any problems.“
„In our view, less technical AI expertise is needed and more structural understanding of processes and data. The greatest leverage lies not in training employees to become prompt engineers, but in training them in logic, data quality and responsibilities. The use of AI only works if the underlying data is correct. Employees need to understand how emissions data is created, which regulatory requirements are relevant and the consequences of incorrect data. AI can structure, check and automate, but it is no substitute for professional judgement.
Our approach at COBACK is therefore: the software takes on complexity in the background. Users do not need in-depth AI knowledge, but rather a basic understanding of their own company data. This is why the COBACK software also focuses on „learning by doing“ through our various modules. The knowledge is imparted in parallel with the data input. This is the only training that is still required.“
„Definitely - for three reasons. Firstly, a basic legal understanding is needed, for example with regard to data protection and AI regulation, so that AI is used responsibly and in compliance with the law.
Secondly, the quality of the results depends heavily on the competence of the users - in other words, on how precisely they can formulate requirements and categorise results.
And thirdly, we need the ability to critically scrutinise AI results. In sustainability management in particular, we must not accept untested analyses or reports. AI is a powerful tool - but the added value only comes from trained, reflective users.“
„Yes - especially application expertise. Companies need to understand how AI works, what the underlying database is and where the limits are.
A common mistake is either blind trust or complete rejection. Both are problematic. ESG managers should be able to critically scrutinise AI results and compare them with regulatory expertise.
The combination of ESG expertise and digital understanding will be crucial in the future.“
5. information security and legal requirements in the context of AI use
We asked the ESG software providers the following question:
„Tools with AI are sometimes viewed critically by SMEs for reasons of information security. What should SMEs look out for in order to implement AI in compliance with the law (e.g. GDPR, NIS 2)?“
The answers briefly summarised:
Information security is a key issue when introducing AI, especially for SMEs. The ESG tool providers surveyed emphasise the importance of data being processed in compliance with the GDPR and transparency regarding data flows. Legally compliant implementation is particularly crucial for sensitive company data such as supply chain information or emissions data. Responsibility and control over data processing must be clearly regulated and the traceability of AI results must be ensured in order to minimise risks
The answers in full:
„SMEs should not simply „throw in“ all the data with AI and activate all access (especially for agents), as this harbours massive security risks. AI should be treated like an IT service provider: only use necessary (preferably pseudonymised) data, secure data flows and third country transfers (ideally through EU/EEA hosting) and contractually regulate that data is not used for training. On the GDPR side, clean AV contracts, deletion concepts and technical protective measures (encryption, role rights, logging) are required. If NIS2 is relevant: also integrate AI into cyber risk management, including supplier checks and incident response/reporting chain.“
„The scepticism towards AI solutions is absolutely understandable, especially among SMEs. As soon as sensitive company data is processed, information security and legal compliance take centre stage. It is therefore crucial that companies check exactly where their data is stored and processed. Is a clean order processing contract in accordance with the GDPR in place? Transparency regarding data flows and involved subcontractors is therefore also a very essential topic for us.
It should also always be checked with various providers that the data entered, especially in the compliance sector, is not used for the general training of language models. In my opinion, these are the most important things to look out for.“
„From our point of view, the most important thing is to empower employees. They should have a basic understanding of which legal framework conditions are relevant and how they can comply with them in practice when using AI. This means designing processes in such a way that employees know which data they are allowed to use and how. Not everyone needs to have legal expertise, but awareness and the ability to act are crucial.“
„Especially in the ESG sector, companies work with sensitive information - for example on supply chains, governance structures or emissions data.
Important points are therefore: GDPR-compliant data processing, transparent source basis, traceability of AI statements, clear responsibilities within the company and control over where and how data is processed.
Many general AI models are not designed for regulatory precision. Companies should therefore check whether the systems used are based on curated, traceable sources and whether they are specialised for the ESG context.“
6 The ecological footprint of AI
We asked the ESG software providers the following question:
„AI itself has a considerable ecological footprint - energy and resource consumption are real. How should companies deal with this paradox? AI for sustainability vs. the sustainability of AI itself?“
The answers briefly summarised:
The ecological footprint of AI remains a major paradox. Although it saves energy and resources through automated processes, the technology itself is energy-intensive. The crucial point for ESG software providers is that AI is only useful if it improves metrics or makes processes more efficient. AI must be used in a targeted and efficient manner so that the benefits justify the consumption of resources. However, energy consumption should not be disregarded and unnecessary computing effort should be avoided.
The answers in full:
„The efficiency gains from AI and therefore the faster processing of tasks - which also leads to lower energy consumption from the laptop/PC/monitor per task - make up for this. In addition, there are now also server centres where their waste heat is used for district heating. Nevertheless, AI should be used sensibly.“
„Quite frankly: AI consumes energy. Full stop. Large models run in data centres and need GPUs, cooling and electricity. Anyone using AI in the context of sustainability cannot ignore this. But that also applies to every other area.
The crucial question is therefore not whether AI consumes energy, but whether its use has a positive effect. If AI merely writes marketing texts faster, the ecological added value is questionable. However, if it replaces thousands of manual Excel processes, creates emissions transparency and enables better investment decisions, the impact can be significantly greater than the company's own consumption. If you ask yourself this question in general, you will usually be able to make the right decision.
We also regularly address this issue. This is why AI does not run permanently during the utilisation processes at COBACK. It runs specifically where it is needed and makes sense.“
„This is actually an exciting paradox. AI can help companies to achieve their sustainability goals faster - but at the same time it itself causes energy and resource consumption.
Our approach is therefore two-pronged: on the one hand, AI should be used specifically where it brings a real efficiency gain, i.e. where it makes processes significantly faster or improves data quality. Secondly, it is important that employees are competent in dealing with generative AI. They must learn to prompt precisely so that results are generated accurately without producing unnecessary computing effort.“
„AI consumes energy, that should not be ignored. However, the decisive factor is the benefit in relation to the use of resources.
If AI helps to avoid regulatory misinterpretations, make emissions more transparent or speed up decision-making processes, the positive effect can clearly outweigh the negative. It becomes problematic where AI is used in an inflationary manner without clear added value.
In my view, the sustainable use of AI means: targeted, efficient and with a clearly defined purpose.“
7 AI as a strategic tool in sustainability management
We asked the ESG software providers the following question:
„AI itself has a considerable ecological footprint - energy and resource consumption are real. How should companies deal with this paradox? AI for sustainability vs. the sustainability of AI itself?“
The answers briefly summarised:
In the long term, AI could at least indirectly play a strategic role in sustainability management by taking over operational tasks such as reporting and data analysis, so that sustainability managers gain more capacity for strategic tasks such as scenario analyses and risk management. However, AI will probably not replace the work of sustainability managers, but rather strengthen their strategic position as a leveraged tool, e.g. by uncovering important patterns in large amounts of data and using them for strategic insights. However, the responsibility for strategic alignment is still seen as lying with humans. AI helps with data processing, but it will perhaps never replace the vision and leadership of a human being.
The answers in full:
„This is where we see the greatest potential in the area of sustainability. If AI has access to all the data along the Value chain and this is also primary data, then it develops its full potential. Insights can then be generated, compliance can be achieved at the touch of a button and measures can be derived. However, user-friendly software is needed again for data collection.“
„AI will certainly be able to make its own specifications in the future. It can already do that today. Whether these are good and correct is more of a question. But if these problems are as good as forgotten, then we have a lot of potential for sustainability management.
Sustainability managers will deal less with tables and more with scenarios. What does a CO₂ price increase mean for our business model? How will new regulation change our investment planning? Where do systemic risks arise in the supply chain before they become visible?
AI will recognise patterns, make dependencies visible and play through options, but it does not replace a sense of responsibility. But we think that should be clear to everyone.“
„In the coming years, it can make the role of sustainability managers much more strategic. Especially in processes with large amounts of data - such as operational documentation - AI can automatically analyse, recognise correlations and simulate scenarios. This frees up capacities because sustainability managers are less operationally involved and can instead strategically shape the sustainable transformation of the company.“
„AI has the potential to significantly reduce administrative tasks - especially in reporting and regulatory research. This creates space for strategic work: scenario analyses, transformation planning and business model development.
We do not see AI as a replacement, but as a lever. Sustainability managers can concentrate more on content management if routine tasks are automated.
In the long term, the role will evolve from documenting to providing strategic impetus, provided that AI is used in a conscious and considered manner.“
Summary
In this article, we have looked at the Opportunities and Risks of the Use of AI in sustainability management and the market. Four providers of ESG software tools answered our questions and shared their perspectives on the practical applications of AI. AI can already be used today, particularly in the Data analysis, the Creation of reports and the Fulfilment of regulatory requirements offer clear added value.
While AI has many Operational processes in sustainability management, however, the ESG software providers also see clear opportunities for Boundaries, especially when it comes to qualitative evaluations and Strategic decisions where Human expertise remains indispensable.
The Data security and the ecological footprint of AI are discussed, with ESG software providers emphasising that targeted and responsible use is necessary to protect the minimise our ecological footprint. At the same time, the potential gain in efficiency may well justify the footprint of AI in some cases.
Ultimately, AI is seen as a strong Assistance system for operational tasks in particular, to support the company in the Sustainability transformation support, but is unlikely to be able to do so in the foreseeable future. strategic Responsibility and leadership of a company can be replaced by people.
[...] and from feedback from sustainability managers in joint projects. To get to know the perspective of AI experts from software providers, we recommend our other [...]