"I don't know what I don't know."
I've heard some version of that from nearly every communications leader I've worked with over the past year, and it usually arrives with a slightly rueful shrug. I've come to understand that the shrug is doing a lot of work - because when leaders don't have enough working knowledge of AI to picture what their team actually needs, they can't commission that work, can't evaluate what comes back, and often end up delegating it to someone who's navigating the same uncertainty.
That's where I suspect things can stall. And this article is a reflection on this very challenge.
The delegation problem
An anecdotal digression: I'm kicking off a pilot coaching program at the moment with a communications leader in Asia. She's sharp, genuinely enthusiastic about AI, and has been trying to get her team moving on adoption for the better part of a year.
At some point she did what a lot of leaders do: she identified someone on the team who was interested, appointed them as the AI champion, and handed the programme over to them to drive forward.
It’s sputtering. Not from any lack of effort - he's a solid, hardworking person. But he's also someone who needs more strategic direction than he was getting, and without close engagement from his manager, things drifted.
After a few conversations between us, the leader arrived at a realisation I've come to see as something of a turning point: she needed to get more directly involved herself, stop treating it as something she'd delegated, and actually champion it. That's what prompted the coaching work we're now doing together, which is, in a way, the start of the course correction.
What she also admitted, with some honesty, is that the reason she'd stepped back in the first place was uncertainty. She wasn't sure what AI could realistically do in her specific context. She hadn't built enough working knowledge to know what to ask for, what to prioritise, or what good AI integration even looks like for a team like hers.
When AI usage starts to drift
Most of the communications teams I'm working with are still early enough in their AI journeys that this particular scenario hasn't fully played out yet. They're working through initial projects with me and starting to explore what deeper integration might look like.
But I can already see the conditions forming. Leaders feel the pressure to enable AI, have given their teams access to a tool, pointed someone at it, and maybe arranged some training. Yet they haven't built enough working familiarity with what AI can actually do, in practice and in their specific context, to evaluate what comes back, spot which friction points are worth attacking first, or make a credible internal case for it when the moment comes.
The broader pattern here is well-established. BCG's 2024 research on large-scale technology programmes found that more than two-thirds miss their targets on time, budget, and scope. McKinsey's work on transformations more broadly puts the success rate consistently below 30 per cent, with leadership engagement (or the absence of it) repeatedly identified as a primary factor.
This isn't unique to AI or to communications. Change programmes that don't have genuine, knowledgeable engagement from the person at the top tend to drift, regardless of the technology involved.
What makes AI harder to lead than most technology rollouts is the knowledge dimension. A leader can engage meaningfully with a new CRM or a collaboration tool without needing to understand how it works under the hood. The process logic is familiar enough. With AI, the gap between what the tool can actually do and what most leaders can currently picture it doing is often significant enough to shape the whole programme. If the leader can't see it working in their context, they won't know what to ask for, and they won't be able to judge whether what they're getting is any good.
My hunch is that this is where a lot of AI programmes in communications will quietly get stuck in the year ahead. Not because of resistance or lack of goodwill - there's plenty of both - but because the conditions for the stall are already present: a leader bought in at the level of the idea but not yet at the level of the work, and a team waiting, consciously or not, for a clearer signal of what they're supposed to be building toward.
Knowledge from working experience
What I've come to believe, from a year of working closely with communications teams on this, is that working knowledge at the leader level isn't a nice-to-have, it's the condition most other things depend on.
A leader who understands what AI can realistically do in their context will identify the right friction points, ask better questions, set clearer expectations internally, and create the environment where adoption can actually build on itself rather than plateau after the initial wave of interest.
The coaching work I'm doing with the leader in Asia is one attempt to address this directly - building real working knowledge from the ground up, starting with her context, her team's friction points, and her own day-to-day. I'm testing it as a format with a small group of leaders. Early days, but the direction feels right.
With UnMute having turned three, the year that's felt most significant, though, is this one, with AI exploding across the industry. It feels like the right time to be testing things.
Perspectives' is a Telum Media submitted article series, where diverse viewpoints spark thought-provoking conversations about the role of PR and communications in today's world. This Perspectives piece was submitted by Darren Boey, Founder and CEO of UnMute.
Unmute is a Hong Kong-based consultancy that helps marketing and communications teams integrate and scale their use of AI across their workflows. Its services include training, designing shared structures including prompt libraries and brand-trained AI assistants, and advising on governance frameworks that enable the safe and secure use of AI.
With a career spanning more than 25 years, Darren has a background in journalism, including more than18 years reporting on financial news at Bloomberg, first in Australia and then Hong Kong, where he led coverage of regional markets. He's also led communications teams in the blockchain, gaming technology, and AI sectors.