AI decision making is the use of models to choose or recommend an action, such as which ad to show, which audience to target or where to move budget.
Also known as Automated decision making, algorithmic decisions
AI decision making covers any situation where a model, rather than a person, selects the next action or ranks the options a person will choose from. In advertising that includes bid adjustments, creative selection, audience expansion and budget shifts across channels.
The key distinction is how much authority the model has. Some systems only recommend, some act within limits, and some act without a human in the loop. Knowing which one you are running is half the job.
Agencies increasingly sit between a client and a stack of automated choices.
Accountability stays human. When a model makes a bad call, the client blames the agency. Decision rights and approval steps should be written down.
Explainability matters. If a platform moves spend and cannot say why, the account team still needs a defensible story for the client.
Limits belong in the setup. Spend caps, brand exclusions and escalation triggers keep automated decisions inside acceptable bounds.
An agency uses a tool that shifts weekly budget between Meta, TikTok and YouTube based on predicted return. The team sets a rule that moves above 15 percent need a strategist’s approval and that every shift is logged with the model’s stated reason. In week six the tool tries to move 30 percent into a channel with a brand safety flag. The approval step catches it, and the client sees the log as proof that automation is supervised.
The automations and agents module of the workshop teaches you how to build AI workflows that compress the busywork without taking the craft out of the studio.