Is there a human in the room?

Have agencies and clients stopped trusting each other… and themselves?

I’m hearing about the emergence of a weird loop where nobody wants to be the one actually making a decision. Client uses AI to brief agency. Agency uses AI to develop the work. Client uses AI to evaluate it. Agency uses AI to respond to their concerns. Client uses AI to assess the response. Round and round until someone just accepts whatever the machines negotiated.

The pretence is efficiency. The reality is a collapse in confidence.

Clients and agencies both have lived experience and hard-won judgement. But we’ve apparently decided that LLMs should referee our professional interactions.

Not because it’s better. Because it feels safer to hide behind.

We’re teaching ourselves that human judgement is unreliable. That our instincts need algorithmic validation before we can trust them. That hitting send requires AI approval.

And the work? The work becomes whatever survives the AI-to-AI negotiation. Usually the path of least resistance. The safest option. The thing most likely to make it through both LLMs without friction, oblivious to nuance and cultural context.

We’re all using it. But someone still needs to make an actual decision.