There is a conversation happening in almost every communications agency and in-house team across the APAC region right now. It usually starts with a demonstration where someone shows what a large language model (LLM) can do with a brief, a media list, a first draft, and the room goes one of two ways. Either there is relief (someone else can do the admin) or there is unease (someone else can do the job).
Both reactions are understandable, and both miss the more important question sitting just underneath the surface - what does this shift actually change about what we should be developing in people, and how fast are we moving to do it?
The truth is most of us are not moving fast enough.
We need to know what it is to even know how to use it or argue against it. AI is excellent at pattern-matching, synthesis, and production at speed. Feed it a media landscape, a set of messages, a company filing, and a target audience, and it will generate a reasonable first draft faster than any junior account executive, without complaint, without an off day, and without any of the friction that comes with managing people.
This is genuinely disruptive for the lower rungs of agency work, where the traditional apprenticeship model has always involved a great deal of unglamorous production work such as drafting, formatting, monitoring, summarising, or translating complex information into plain language. Much of that is being absorbed, and the question worth sitting with is what we leave in its place.
The capabilities AI struggles to replicate are precisely the ones that separate good communicators from great ones. Reading a room. Identifying the actual problem to be solved. Sensing when a client is asking one question but actually needs an answer to a different one. Holding a relationship through an uncomfortable conversation, or making a call under pressure when the facts are incomplete and the optics are volatile. These are judgment calls, grounded in experience and pattern recognition built through repeated exposure to ambiguity. They are also, not coincidentally, the capabilities that are hardest to develop deliberately and easiest to let atrophy when agencies are busy and client demands are constant.
The traditional communications pipeline had a logic to it even when no one named it explicitly. Juniors did a lot of volume work and that volume encoded intuition. You learned to write by writing thousands of press releases. You learned to pitch by reaching out to hundreds of journalists and absorbing the patterns of what landed and what did not. The craft accumulated through repetition, and over time, something that looked like judgment started to emerge. Seniority, in that model, was in large part a function of accumulated reps.
If AI is taking on more of that volume work, the intuition-building loop is disrupted. The reps still need to happen, but they have to be designed more carefully rather than assumed.
Developing people in this environment requires more deliberate architecture than most leadership teams are currently applying. The conditions under which judgment actually develops - early and repeated exposure to client conversations, structured reflection on what went well and why, room to make low-stakes mistakes in high-learning environments, and feedback cultures that treat those mistakes as information rather than evidence of inadequacy - do not emerge on their own. They have to be built.
The professionals who will be most valuable in three to five years are those who can sit with a client in a difficult moment, understand what the organisation is actually trying to protect, and give honest counsel that accounts for the full political and human complexity of the situation. That is a capability that requires experience, and experience now has to be designed.
Across communications teams, the approaches currently in play tend to cluster at two unhelpful poles. One group has deployed AI tools widely without changing anything about how they develop people, treating it as a productivity gain and moving on. The other has become preoccupied with upskilling in the tools themselves, running workshops on prompting and building internal playbooks, while spending less time than ever thinking alongside their teams about judgment, instinct, and professional maturity. Both miss what actually needs to shift.
The leaders getting this right are asking different questions in talent conversations. Whether someone can produce good work is a starting point, but the more important question is whether they understand why it is good and can defend it under pressure. Technical proficiency matters less than the capacity to lead without a script. These leaders are also being more honest about the trade-offs involved. There is a version of AI integration in communications that makes teams faster and more efficient while quietly hollowing out the depth of thinking that makes the work matter to clients.
Good communications counsel is valuable because it is grounded in experience, trust, and a long-view perspective that comes from having navigated complex situations more than once. Efficiency gains that come at the cost of that depth tend to show up, eventually, in the quality of relationships and the durability of the work.
If I were designing a talent development programme from scratch for a communications team today, I would build it around five things: strategic judgment, client counsel, storytelling rigour, resilience under pressure, and the ability to give and receive honest feedback. None of these are new. What has changed is how explicitly they need to be cultivated since the traditional mechanisms for building them are no longer doing the work automatically.
Leaders have to name what they are trying to develop and create structures that actually develop it, rather than assuming the work will teach people what the work no longer teaches. Think of AI less as a threat to communications careers and more as a clarifying force. It is surfacing which parts of the work have always been commodities and which have always required something distinctly human, and it is doing so faster than most organisations had anticipated.
The professionals who thrive will be those who have done the work to understand the difference. The organisations that thrive will be those whose leaders were willing to redesign how they develop people before that gap became visible to their clients.
Patricia Malay is General Manager of Bud Communications, an independent PR and content agency operating across Asia Pacific, and founder of Candour Leads, an advisory practice for leaders navigating transitions and organisational change.
With over 20 years of experience across agency and in-house roles - including positions at FleishmanHillard, Havas, Burson, and Ogilvy - she has built a reputation for exploring the human dimensions of communications work: how talent develops, how leaders communicate honestly, and how organisations build cultures where capable people choose to stay and grow.
Patricia is based in Singapore and writes regularly on AI's impact on communications careers, regional talent development, and the gap between what leaders know and what they say.
Read more from our columnists in The Earned View