Three Big Questions: How Should Comms Agencies Use AI?

From crisis response to content creation, large language models (LLMs) are already reshaping how communications teams work. The opportunities are huge, from faster insight-gathering to synthetic audience testing, but so are the risks, including misinformation and the erosion of human oversight. For comms leaders, this isn’t a distant debate; it’s an urgent conversation about the future of our craft.

It’s exactly these kinds of challenges that inspired us to create Three Big Questions, a new series where we explore the debates no comms leader can afford to ignore.

We’re starting with AI, on the back of completing our own in‑house AI training, co‑written and delivered by CEO Debbie Zaman with Stephen Waddington over 5 months. Stephen, a former tech agency lead, comms strategist, and long‑time advocate for responsible AI, joins Debbie in this first instalment.

In three short videos, Debbie and Stephen tackle:

1. Should comms agencies be using LLMs?
2. Can an LLM write good content?
3. What is the best thing AI can help a comms agency with?

Should comms agencies be using LLMs?

Stephen and Debbie discuss how language models can augment the way teams work, from surfacing fresh insights to strengthening planning and ideation. But there’s a clear tension between the potential of AI and the need for authentic, credible communications.

“In an industry built on trust, authentic voice is everything,” Stephen notes. “If a practitioner sends over a piece of content that’s obviously been run with AI, it undermines trust. We absolutely never do that.”

The takeaway: AI has its place, but it can’t replace the human judgment, tone, and nuance that good comms demands.

Can an LLM write good content?

Great comms isn’t just about stringing words together. It’s about context, audience insight, and strategic intent, things an LLM can’t fully grasp.

As Stephen puts it, you can always tell when an LLM has been left to its own devices: “It uses lazy words, like ‘I’m proud’ or ‘I’m delighted’, the kind of generic phrasing that drains impact.”

That’s why our rule is simple: human‑first, every time. AI can support the process, but it never replaces the people who know how to make messages resonate.

What is the best thing AI can help a comms agency with?

For Stephen, the most valuable applications of AI in comms aren’t about replacing humans or churning out content. They’re about amplifying cognitive ability and improving decision‑making.

“The first big use case is improving cognitive ability,” Stephen explains. “AI can take a large amount of content – text or numbers – and extract themes, insights, and messaging. It’s like an additional brain, helping practitioners and teams make sense of complex information.”

The second is testing content from alternative perspectives. “We all come to practice with a particular bias. AI allows us to put ourselves in the frame of mind of other stakeholders to challenge and refine what we’ve produced,” he adds.

These two applications, sense‑making and perspective‑testing, have far greater strategic value than using AI simply to generate content. As Stephen puts it, “Focusing on content creation alone commoditises what you do. Why would you do that?”

Want to find out more about how we’re applying AI thinking to client work? Chat to us at hello@welcometowith.com.