Generative AI is now widely used across the strategic communications industry. Large language models are embedded in research, monitoring, and drafting workflows at firms of every size. Despite repeated predictions that AI would replace the industry altogether, this has not happened. The greater risk, based on my research with King’s College London, is the opposite: strategic communications firms are more likely to go out of business not because they fail to adopt AI, but because they adopt it too carelessly. Drawing on interviews I conducted with senior consultants and decision-makers from more than ten strategic communications agencies, ranging from global firms to small bespoke public affairs consultancies, the evidence points in one clear direction. AI accelerates work, but it destroys value when it replaces human judgment, human relationships, and human accountability. Firms that confuse efficiency with effectiveness, or automation with expertise, place client trust, reputations, and business models at risk. Across the industry, there is broad agreement that generative AI dramatically reduces time spent on routine tasks. Research, summarisation, monitoring, and first-draft ideation can all be completed more quickly. Global firms report significant efficiency gains in content production and analysis, while mid-sized and bespoke consultancies report similar benefits, particularly in research and the processing of large volumes of information. But efficiency is not value. Many interviewees referred to the “20–60–20” model of AI adoption, in which only the first and final 20 per cent of work remains human while the middle 60 per cent is delegated to AI. This approach risks creating a race to the bottom. When most of the work is done by machines, it becomes increasingly difficult to justify retainers for what clients believe is human expertise. While this effect is not yet fully visible, over-reliance on AI threatens pricing, perceived expertise, and the trust on which strategic communications depend on. AI’s limitations are most visible in content generation. Interviewees across global, mid-sized, and bespoke firms were unequivocal: AI alone produces bad press releases. They are generic, lack authentic angles, miss context, and are immediately identifiable to journalists. Several firms stated that AI-generated content is unusable for high-level outlets such as the Financial Times or the BBC. This matters because strategic communications depen on credibility with journalists, editors, and policymakers. Repeated exposure to AI-generated material does not simply fail to secure coverage; it actively damages relationships. Interviewees also noted that MPs’ offices and parliamentary assistants will block public affairs firms that rely on AI-generated emails. Human judgment, relationships, and direct contact cannot be automated. The wider communications environment already reflects this problem. AI-generated content has flooded platforms such as LinkedIn, saturating the digital space and undermining trust. Rather than increasing influence, the proliferation of low-value automated content has made genuine human interaction more valuable. Another structural risk identified in the research is the impact of AI on early-career roles. Generative AI is absorbing tasks traditionally performed by junior consultants, weakening the industry’s talent pipeline. As fewer graduates are trained, fewer professionals develop judgment and experience.  Put simply, there will be a lack of talent in the strategic communications industry, and the quality of communications will go down. While AI is widely used for research, stress-testing ideas, and processing information, interviewees consistently emphasised one rule: AI must be questioned. Large language models hallucinate, lack evaluative judgment, and reproduce errors at scale. This makes AI particularly dangerous in high-stakes advisory contexts involving sensitive or market-moving information. Interviewees stressed that AI output cannot be trusted without human review, contextual understanding, and critical assessment. At its core, this is a trust issue. Clients do not hire strategic communications firms for work that can be done by AI; they hire human experts. Strategic communications firms are not technology companies. The central threat facing the industry is therefore not technological disruption but trust erosion. Overuse and misuse of AI lowers content quality, damages media relationships, weakens talent development, and undermines the justification for retainers. Clients will not pay premium fees for work they believe is automated. Several interviewees warned that firms overly reliant on AI risk reputational collapse. If clients perceive advice as generic, machine-generated, or disconnected from real-world relationships, they will leave. In that scenario, firms do not decline gradually; they fail abruptly. Generative AI has changed how strategic communications work is ...