Your organisation’s reputation is being shaped right now by an audience that never sleeps and doesn’t read beyond what it needs. AI tools – the large language models (LLMs) powering search, chat, and research tools – are synthesising everything said about your business for your customers, investors, regulators, and employees. They are your newest communications stakeholder.
How can sustainability communicators connect to this new, non-human audience, and what do we need to do differently to ensure our key messages cut through?
The AI genie is out of the bottle
The use of AI tools is growing in Australia and around the world. The 2025 Australian Digital Inclusion Index found that almost half of all Australians recently used generative AI tools, with usage highest amongst students and people aged 18 to 34.
With so many people using AI, your customers, investors and regulators are likely learning about your organisation this way. The summaries these tools provide are influencing stakeholder perceptions of your values, performance, and future plans. AI tools are therefore influencing your corporate reputation. This means communicators now need to consider AI as an audience, one that is reading your website, scanning the Internet for context, and forming summaries that influence human judgement.
Learning how AI learns about you
To influence AI, we need to understand how AI forms its summaries. RepTrak has begun measuring AI stakeholder perception by asking selected tools the same reputation questions it asks human stakeholders. Its recent webinar surfaced an important challenge. RepTrak found that the same LLM drew on different sources to form its view of a client company and each of its peers. This means consistency of message matters across owned and earned channels. Gaps or contradictions in that picture will affect the summary your stakeholders receive.
Just like any stakeholder group, AI tools don’t have one homogeneous viewpoint. Prashant Saxena from iSentia recently described AI tools as a ‘council of stakeholders’, gathering knowledge and developing perspectives in different ways. Dr Nici Sweaney, AI expert and Founder of AI Her Way, described them more like an ‘accidental narrator’: not conscious but still capable of flattening carefully crafted commitments, misrepresenting performance, or amplifying vague language about impact through the logic of summarisation.
A particular challenge for sustainability
The stakes of AI summarisation are especially high for sustainability professionals. AI tools are increasingly being used by investors, analysts, and ESG rating agencies to read, score, and compare corporate sustainability performance. Organisations such as MSCI, Sustainalytics, and others are using AI-assisted analysis as part of their ratings processes, meaning that how your sustainability commitments are expressed and then summarised is influencing ratings scores.
This creates two distinct risks. The first is that vague or aspirational language – “we are committed to a sustainable future” – will be further generalised by an AI tool, stripping it of whatever intent it carried. The second is the mirror image: that genuine, specific commitments will be misread or misrepresented because they are embedded in dense or inconsistent disclosure language. Both risks have consequences for credibility, reputation, and increasingly, for regulatory compliance as greenwashing scrutiny continues.
Surviving the summary: Writing for humans and machines
I asked four AI tools – ChatGPT, Claude, Gemini and Perplexity – what human writers should know about writing to make it easy for an LLM to synthesise and summarise their key messages. Three themes stood out.
-
Be direct, clear and specific
The AI tools recommend leading with your main point, and find clever wordplay, metaphors and slow reveals confusing. Starting with the ‘why’ and the ‘what’ works well for these tools, to make it easy for them to identify the key message. Our human audiences also benefit from this approach, as we are increasingly time-poor and impatient with clever communications that build to the main point.
The challenge for communicators is striking the right balance between flourish and precision. AI tools are looking for “quotable units” of fact-driven information; human audiences often want to connect with the essence of the story to feel its meaning. A human perspective – the odd combination of insight and personal experience that produces something original – is increasingly appealing to read. Personally, I wouldn’t enjoy communicating without a little flourish, and I wouldn’t enjoy content that didn’t create those moments of connection. Our opportunity is to find ways to incorporate both approaches into our work – to seek to inform and to engage – so that all stakeholders can understand it.
Phrases like “making sure teams work better together” are too ambiguous for AI tools, as they have trouble interpreting what the impact is. For both AI and human readers, phrases like this should be elevated to include the benefits or outcomes: “making sure teams work better together to improve engagement and productivity” would be clearer.
-
Use both precise and plain language
The AI tools had mixed views on whether technical or complex language makes summarising easier. Three tools recommended avoiding generalised framing or generic statements like “the results speak for themselves”. Such phrasing tends to get flattened further by the summary, losing whatever meaning it carried. They want us to say exactly what we mean, once, so they can understand.
Gemini recommended a different approach, preferring the use of industry-standard terms and disliking plain English writing. For example, Gemini recommended using the term “interoperability” rather than describing it as “systems that can talk to each other”. The clarity of the technical term removes the risk of misinterpretation for the AI tool.
Your human audiences will feel differently. They would prefer the simpler language approach. If we think back to Prashant Saxena’s ‘council of stakeholders’ idea, the most useful approach may be deliberate precision. Use simple, clear language most of the time, and reach for well-established technical terms where misinterpretation matters.
This is particularly important in sustainability communications, where the terminology is not just technical but contested and consequential. Terms like “net zero”, “carbon neutral”, “nature positive”, and “double materiality” carry specific, internationally recognised meanings. Using them precisely (and consistently) reduces the risk that an AI tool will misinterpret a term or smooth over a commitment that should be on the record.
-
Structure works for everyone
Both humans and AI tools value structure to support information processing. What the tools are looking for is a lot like what we do for digital communications and website copy: short sentences; clear, descriptive headings and subheadings (multiple levels); bullets and numbered lists that help to prioritise information. Tools are scanning content, looking for the key points without having to read every word. Humans also read this way, so structure serves both audiences.
Thinking about AI as a communications audience has changed the way I write, pushing me to focus more deliberately on clarity, structure and precision, while still leaving room for the human connection that good communication requires.
For communicators, our task now is to write for both the people we want to reach, and the systems that increasingly mediate their first impressions. Ask the tools what they think of your organisation, and test your most important content through an AI lens. You may be surprised, and informed, by what you find.
We’d love to hear your thoughts – email susan.dyster@bwdstrategic.com or message her on LinkedIn if you’d like to continue the conversation.
About the Author
Susan Dyster is Senior Strategy Manager and Reporting Lead at sustainability strategy consultancy BWD Strategic, and an expert in strategic communications.
