The AI Score in Your Review You Cannot See
About half of marketing performance reviews this cycle have an AI component.
Almost none of them have clear criteria for what good looks like.
The pattern is showing up across in-house teams, agencies, and AI-first companies, and it is doing more to shape marketing careers right now than most of the conversations happening on top of it.
Three Things to Know
1. AI showed up in performance reviews before anyone defined what’s being measured
Roughly half of marketing performance reviews this year have an AI dimension.
The specifics are vague enough that marketers are fudging their own self-scoring because there was no clarity behind the rollout in the first place.
What this means in practice for you:
Your review now depends partly on a metric your manager cannot define, that HR cannot calibrate, and that your peers are scoring inconsistently.
The reviews still happen.
The bonuses still get allocated.
The promotion decisions still get made.
2. Shadow AI is now the actual AI strategy at most companies
Most marketing teams have officially sanctioned access to one tool, usually Claude, Gemini, ChatGPT, or Copilot.
A meaningful portion of marketers are paying out of pocket for the tools they actually use: a second model the company did not approve, plus specialty tools like Gamma, Granola, Tactiq, or domain-specific assistants.
They keep it off the record because governance has not caught up to usage.
This creates a strange dynamic.
The marketers doing the most interesting AI work are often the ones least likely to talk about it openly, because their stack is technically unauthorized.
The marketers most loudly advocating for AI internally are sometimes the ones using the least of it.
The mismatch shows up in performance reviews as inconsistent visibility for the people producing the most actual value.
3. The prompt-gatekeeping behavior is rational, and it is reshaping how marketing teams operate
Marketers who have figured out useful prompts and workflows are increasingly not sharing them.
For self-preservation reasons.
When your edge is knowing how to use a tool better than your colleagues, sharing that edge feels like trading job security for team goodwill.
This is the quiet version of every AI conversation happening in marketing right now.
Marketers are watching headcount conversations, watching what gets automated, and concluding that their best protection is their own depth, not the team's collective depth.
The cooperative culture that defined a lot of marketing teams over the last decade is thinning out in a way nobody is openly discussing.
Two Moves to Make
1. Build your own AI scorecard before your manager builds one for you
If your review is going to include AI and your company has not defined the criteria, define them yourself and put them in front of your manager.
The marketers who get out ahead of this conversation control the framing. The ones who wait get scored on whatever ad-hoc criteria emerge in the room.
A simple version covers four things:
What AI tools you have used this quarter,
What specific outputs they helped you produce,
What time or cost they saved (with a real estimate, not a vague claim),
And what you are not yet using them for and why.
That document, written and shared with your manager before the review window opens, becomes the criteria you get evaluated against.
Without it, you get evaluated against your manager’s gut feel.
2. Pick one workflow and become the team’s authority on it
The gatekeeping dynamic is real but it works against you in the long run.
To build a durable career capital around AI:
Pick one workflow (lifecycle copy production, campaign brief drafting, attribution modeling, brand voice consistency at scale)
Become the person your team comes to for that specific application.
This is different from being broadly “good at AI.” Broad fluency is now table stakes.
The position that actually moves your career is depth in one applied workflow, demonstrated in a way that makes you the obvious owner when that workflow scales.
Pick the workflow that maps to where your discipline is going rather than where it currently sits.
One Question to Sit With
If your performance review this cycle came with a single line that said “AI proficiency: meets / exceeds / below expectations,” and your manager had to back it up with specific evidence, what evidence would they actually have on you?
Hakan | Founder, Marketers Remote