The 5 Pillars of an Agency AI Policy: Protecting Data and Driving GEO Results

Created with Google Gemini

If you haven’t created an AI policy for your employees, now is probably the time. As marketing and communications agency professionals adapt to the changing information ecosystem, your clients should understand how you’re using AI tools in your work. Whether employing artificial intelligence to work more efficiently and/or tapping into Generative Engine Optimization (or Answer Engine Optimization) to make sure your clients show up in the results, clear guidelines are key to protecting information and reputation. 

At RoseComm, we recently created an AI policy that all employees and partners will agree to before working with our clients. We’ve also given our clients the choice to opt out of our use of AI while working on their behalf. 

Why Create an AI Policy?

This is new terrain. We did a lot of research and consulted with some of our peers to come up with an approach that:

  • Ensures our AI use aligns with agency and industry values of innovation, integrity and transparency
  • Protects confidential information
  • Maintains work quality 
  • Mitigates legal and reputational risk

What Goes into an AI Policy?

To follow are the 5 pillars of our policy:

  1. North star statement, i.e., how we balance efficiency with human expertise
  2. List of approved AI tools for different applications: editing, proofreading, research, transcription, design, etc.
  3. Approved uses of AI: specific tasks and related guidelines
  4. Prohibited uses of AI, aka, your non-negotiables
  5. Uses requiring disclosure that AI was used

What Does this Mean for the PR Profession?

It’s one thing to create a policy, but the bigger question is how the proliferation of AI influences the practice of public relations. Amid all the excitement, there’s actually a renewed interest in the fundamentals of strategic communications. We’ve all seen plenty of AI “slop” in recent months, making credible, human storytelling as important as ever. In fact, according to The Wall Street Journal, the number of LinkedIn job postings that included the word “storyteller” doubled last year. The words “storyteller” and “storytelling” were also mentioned 469 times on earnings calls and during investor days in 2025, compared to 359 in 2024. Audiences are clearly hungry for authenticity and originality – two attributes AI does not possess.

If you’re just starting to think about your AI policy, we’re happy to share ours. Just reach out to rostmann@rosecomm.com.

Leave a Reply

Your email address will not be published. Required fields are marked *