I keep noticing how much effort agencies put into every client interaction.
Everything is treated as high stakes, high touch, and deeply personal.
It comes from a good place. No team wants to feel like they are giving clients a thin or automated experience.
But over time I’ve started to question whether all of that effort is actually valuable.
Some interactions genuinely benefit from care and judgement. Others mainly need to be accurate and delivered when promised.
When those two types of interaction get treated the same way, it usually causes problems. Senior people get buried in admin, and the moments that really need judgement get squeezed.
I see this most clearly in delivery work.
There are interactions where clients are looking for thinking. Decisions, trade-offs, context, reassurance. These are the points where clients start to trust your judgement and look to you for direction. They benefit from experience and presence.
Then there are interactions where clients just want information. Status, confirmation, reminders, summaries. In those moments, I have rarely seen a client value the fact that a senior person typed the words. Or even that they come from a human. In practice, clients tend to care more about whether the information is clear, dependable and arrives when expected.
Once I started seeing that difference more clearly, it changed how I thought about where AI could actually help.
Not to replace relationships, but to take the lead on the work that never really needed a human at the keyboard in the first place.
Status updates are a good example. Most follow a familiar structure. What moved, what did not, what needs attention. AI can draft those from project data far faster than a human. I still review them, adjust tone, and decide what needs calling out, but I am no longer starting from a blank page.
The same applies to meeting notes, action lists, and risk logs. These are important artefacts, but the value is in keeping them current, not in how lovingly they are written. AI does the first pass. Humans step in where judgement is required.
I see the same pattern show up across different pieces of work.
AI has been most useful for me on tasks that are fairly predictable and involve a lot of information. When the work involves uncertainty, difficult trade offs, or asking someone to commit, human judgement still matters.
I expected this to risk making things feel colder and more transactional, but in practice it has had the opposite effect on the client experience. Responses are quicker. Updates are more consistent. And the human time that remains is spent in places clients actually notice.
For me, that’s the real opportunity. Not to do less for clients, or to automate everything, but to stop spending human energy on work that was never really valued as human in the first place.