Q&A: AI and its impact on PR with our Head of PR – Toby Brown

Our head of PR, Toby Brown, recently took part in the inaugural Meltwater Agency Grind – a forum for agency leaders to share experiences and ideas to enhance how we all work. The session focused on AI and its implications for agencies. The session wasn’t recorded, but we asked Toby to share his thoughts to the questions raised in the debate and have shared the responses below:

What tools are agencies using? Is it all about ChatGPT? What about Midjourney or other image generators? Any more exotic tools outside of the well known ones?

We’re currently doing a lot of experimentation with ChatGPT as it’s by far the most powerful generative AI out there today. Personally, I’m also a Google enthusiast and have been experimenting with Bard, which is useful for being internet-connected but doesn’t quite match the thinking power of ChatGPT.

Some teams are also exploring the application of image generation AIs like Midjourney and aggregator platforms such as Nightcafe, which provides access to DALL-E and Stable Diffusion and a number of other image AIs with different ‘skills’.

Given the importance of client confidentiality in how these tools are used, we’re excited that Publicis have recently launched its own private ChatGPT and DALL-E sandbox for agencies to experiment with. 

Finally, I’m really excited to see how both traditional and new PR tech vendors are starting to apply the power of generative AI to areas like media analysis, understanding sentiment and influencers and measuring the impact of PR. Being able to digest and derive insight at scale will supercharge this sector and, as a result, the value we can bring to clients.

Do agencies need to adopt policies for how their staff use AI? Perhaps even guidelines from the industry bodies?

Absolutely, this is a key priority for us. We’ve already published initial editorial guidelines for AI given this is a big area of focus, and will be providing updates and other comms to be really transparent with our clients. While we encourage our team to experiment with these tools, having clear guardrails is so important to ensure we’re producing the best and most creative work in the most effective way. Likewise, teams need to know what is and isn’t OK with regards to client information to ensure this is always kept safe and secure. 

What are your feelings on agency transparency when it comes to using these tools? How much do clients need to know? Does it matter whether or not agencies use AI so long as the results are good?

We’re committed to being transparent at every turn. Having already published our editorial guidelines, we’ll continue to provide updates on where we’re using AI (and where we aren’t!) to provide clients the impact and value they need. 

What are you hearing from clients about AI? Do they view AI as a positive, or are they worried about agencies cutting corners?

The response is varied. While I haven’t encountered any overt concerns, several early adopters are actively exploring AI’s potential. I think it’s inevitable that questions will come up as marketers and comms leaders recognise the power and potential of AI. The critical thing for me is that it’s never used to create content faster at the expense of quality. This is a race to the bottom and completely at odds to why clients need agencies in the first place. As long as we’re always transparent in how we’re working and continue to prove our value day in, day out, then clients will keep needing and buying what we’re selling. 

Is there a risk that in-house teams will use AI to replace agencies for some kinds of work?

There’s a perpetual swing back and forth between client teams growing and taking on work in house and outsourcing more to agencies. I think this will continue, even if AI helps us all work smarter and faster. We’re seeing some early adopter clients with in house content experts experimenting employing AI to generate what I call ‘derivative content’. This includes things like transforming long form content into blogs, social posts, listicles and so on. But this is by no means widespread. We’re not expecting a wholesale move here as it always depends on client teams and the skill sets they have versus what they need from their agencies. 

For me agencies have, and will continue to have, a crucial role in extracting compelling insights from client experts and producing engaging content that resonates with audiences. Creating content with personality that tells brand stories in a compelling way is critical and it’s easy for this to be lost with vanilla AI-generated content.  Most clients I speak to share my view that the real value they want from agencies is creativity rather than just ‘hands to the pump’. Regardless of how AI changes content production, we’ll always be laser focused on adding creative value to every relationship.  

What does this all mean for the agency model? How significant is the long term impact going to be?

Predicting the exact outcome is challenging, but it’s clear that AI is going to reshape knowledge work itself. Undoubtedly, it will alter how agencies operate and the nature of our work. I believe it will force agencies to become even more agile and tech-enabled to fully harness the benefits of AI. As mentioned above, our primary focus remains firmly on delivering what clients truly value: creativity, campaigns, content and connections that deliver real business value. Wherever AI enables us to do this more effectively, that’s where we’ll be focusing.