What PR Gets Wrong About AI – and How to Get It Right, 3 Lessons From The With Team

With the onus on leaders across all functions of the business to show how they’re incorporating AI into their teams, comms and PR are no exception. 

And as the first stories of AI related comms redundancies in brands and agencies roll in, comms teams are scrambling to prove the value of their efforts, with every agency and its dog commissioning research tying PR efforts to Generative Engine Optimization (GEO).

But as Edward Targett, Founder of The Stack, recently pointed out on LinkedIn, AI is not just a search headwind – it’s also a huge opportunity. 

That’s why we’ve been working with Stephen Waddington, former tech agency lead and expert in AI in PR, to design and deliver an intense four-month training programme for the With team – focused on how we bring AI in at the right point, and when we leave it out.  This isn’t just about capability. It’s about raising the bar for how AI is used: effectively, safely, and strategically in the world of comms and marketing.

Here are three lessons from our journey so far:

1. AI Must Be Effective, Not Just Efficient

AI has the power to accelerate tasks – but speed alone isn’t the goal. Effectiveness comes from knowing where AI adds value, and making deliberate choices about how it’s used.

AI can streamline media monitoring, process complex data, and even map flashpoints in crisis preparedness

What it’s not going to do is spot the subtle politics playing out in a client’s boardroom, lead a creative brainstorm that shifts strategy, and challenge a client’s thinking with nuance. AI is a tool, not a conscience. 

That’s why our AI Governance Framework puts humans firmly in the driving seat – ensuring that every output still meets the high editorial standard we’ve set for over 15 years.

At With, we’re focused on the high-impact applications that genuinely enhance the strategic and creative side of PR. That includes:

• Stress-testing messaging or press releases against synthetic audiences to identify blind spots or improve resonance.
• Running competitive analysis to sharpen brand distinction and narrative cut-through.
• Mapping topic trends and media coverage to guide timing, tone, and story selection.
• Supporting messaging development and the creation of compelling narratives, positioning, and storytelling
• Stress-testing messaging and press materials against synthetic audience segments to uncover blind spots, test resonance, and refine effectiveness.
• Spotting topic trends and competitor analysis to guide timing, tone, and story selection.

Used well, AI helps us challenge assumptions, broaden thinking and identify smarter angles – not just create faster drafts.

2. AI Must Be Safe and Secure

From company secrets to watercooler gossip,  the risks of sensitive information making its way into AI tools have been widely reported. When data is used to train public models or resurfaces unexpectedly, the consequences can range from reputational damage to regulatory breaches.

For communications and PR professionals, the nature of our work and sensitive information we handle (unreleased financial results, embargoed announcements etc) means the stakes are especially high.

Because of this, incorporating AI into communications work demands a high level of trust. That comes from 

• Using enterprise-grade tools that guarantee data privacy—ensuring prompts and outputs aren’t used to train public models or retained beyond their immediate use.
  Respecting client-specific AI policies, with flexibility to accommodate full opt-outs or limited use cases based on organisational needs.
• Designating AI leads or ‘guardians’ to monitor developments, guide responsible usage, and maintain alignment with legal and ethical standards.
• Being transparent about how AI is used, objectives for its use, and where human oversight is applied – ideally through a clear and accessible AI governance framework (we have shared ours with all clients and update it quarterly).

3. AI Must Be Human-Led and Totally Tailored

AI isn’t one-size-fits-all, to be set up for success, AI tools need proper context. That means being fed with the right inputs: brand tone of voice, key messages, audience insights, past content, and strategic priorities. 

Without this, outputs risk being inconsistent, misaligned with the brand, or generic (something journalists will routinely take to LinkedIn to name and shame) 

Every organisation is different, so no two AI setups should look the same. Whether it’s building a shaping content plans, simulating audience response, or developing an ideal customer profile, the tool needs to reflect the unique environment it’s working in. When properly configured and guided, AI can significantly enhance output quality, consistency and speed -while keeping communications clear, credible, and on-brand.

AI can and should elevate your comms – but only when used responsibly, securely, and creatively. At With, we’re combining training, governance, and editorial expertise to deliver just that.Questions on our AI Governance Framework? Email hello@welcometowith.com