The use of generative artificial intelligence in public relations carries the risk of diluting a brand's specific communication tone (brand voice). This was highlighted by Susan Oguche, former Executive Vice President and Chief Communications Officer of Cavaliers Operating Company, at the AI Horizons conference organized by Ragan. According to her, communication experts should not approach AI tools as a “blank slate” but as a system that requires specific context and authentic data.
A fundamental mistake when working with large language models, according to Oguche, is “blind prompting,” i.e., giving commands without providing relevant background information. If a user expects an output that corresponds to the brand's identity, they must provide the artificial intelligence with enough source material from which the model can draw. Uploading proprietary documents, communication manuals, and examples of previous texts is key to ensuring AI generates content that does not sound generic.
Oguche emphasizes that the quality of the result is directly proportional to the quality and depth of the input data. Instead of merely describing what the text should look like, she recommends actively using the file upload function.
“Please, upload documents. Don't just use empty prompts. Utilize source data that will help you achieve the best results,” she stated during her presentation.
As part of a broader strategy to protect brand identity in the AI era, it is essential for organizations to define firm boundaries for the use of these tools. Control of source data serves not only to maintain consistent style but also as a prevention against factual errors that can arise from free generation without data. Oguche also suggested that the role of communicators is shifting from merely writing to being editors and data curators who oversee whether AI outputs correspond to the company's long-term strategy.
The issue of preserving brand voice is one of the main topics currently addressed by communication experts. With the increasing availability of AI tools, there is growing pressure for speed in content creation, which often leads to neglecting stylistic consistency checks. The AI Horizons conference showed that successful integration of technologies into PR primarily requires a change in work assignment methodology and an emphasis on precise preparation of input information.