The collection of tasks that represent ad ops will be one of the first jobs our industry assigns to AI. Menial tasks like inventory QA, campaign builds, and pacing management are perfectly suited for agents to execute in pursuit of improved advertising outcomes.

Like many roles ripe for early AI disruption, ad ops roles across agencies and brands were personified by fresh-out-of-college new grads and created a burgeoning farm system for scaled media organizations to identify future rising stars. Many companies also chose to offshore these roles to countries with cheaper labour, creating a multi-time-zone international ad ops assembly line. Now those same companies are left wondering where new pools of talent will come from, and what role their offshore teams play in an agent-first world.

This reality of job contraction has created second-order effects in work accountability. As the ranks of lower-level employees thin, the tasks they owned begin migrating up to managers with several years of experience. These managers are left to re-educate themselves on a new set of tools (i.e. agents) while simultaneously becoming accountable for work they graduated from years ago. Even though agents are now doing the work directly, these managers remain the ultimate authors of their agents' actions.

Fewer roles are required, but AI has effectively transformed what used to be multiple jobs into one. A recent HBR study captured this well:

In an eight-month study of how generative AI changed work habits at a U.S.-based technology company with about 200 employees, we found that employees worked at a faster pace, took on a broader scope of tasks, and extended work.

They went on to say:

Once the excitement of experimenting fades, workers can find that their workload has quietly grown and feel stretched from juggling everything that's suddenly on their plate. That workload creep can in turn lead to cognitive fatigue, burnout, and weakened decision-making.

We see this firsthand. Agents now allow any media organization to pursue a standard of advertising excellence through always-on ad ops execution that was never even contemplated under the constraints of human labour. The bar for advertising performance is heightened because everyone is becoming increasingly aware of the tools we have at our disposal. Team members become infinitely more productive, yes, but at the tradeoff of a single human becoming the fulcrum point of a team of decisions and actions previously held by many.

That tradeoff is the central design problem we are now trying to solve. And is one that Gigi’s customers are already asking us to address. If we rely solely on assisted AI, we are merely compressing multiple human roles into one, asking that single team member to bear the weight of decisions and actions previously distributed across a team. The solution is a bridge to guided authoritative AI, where the human sets intent and constraints, and the agent translates that intent into a plan, executes low-level operational work automatically, and surfaces only the decisions that materially affect outcomes.

The shift is toward a completion model: the agent runs autonomously to a finished state, executes what falls within its authorized confidence threshold, and returns a single reviewable artifact. The work collapses into an evaluate-and-decide moment rather than a manage-and-execute one. This requires a permission hierarchy that defines what the agent can do unilaterally versus what requires human judgment, and an output layer that gives humans enough context to evaluate the result without reconstructing the process. The human role moves from guiding every step to defining the goal upfront and setting the criteria by which the agent's output gets accepted or sent back.

We have already seen this play out in coding agents. Early tools acted as autocomplete engines, generating snippets that developers stitched together manually. Today's more advanced coding agents operate at the level of intent. A developer describes a feature, and the agent explores the codebase, writes and modifies files, runs tests, debugs errors, and presents a working implementation. The developer does not approve every line before execution. They define constraints, authorize categories of changes, and intervene only when tradeoffs are meaningful. The nature of the work has fundamentally changed.

Developers have become crafters of problems to solve rather than executors of solutions. Advertising agents are at an earlier point on that same curve. They have successfully removed low-level ad ops work from humans, but the next step is eliminating the burden of ad ops review entirely. When that happens, the people running media campaigns will stop being operators and start being strategists. That is when the work genuinely improves.

Note: This originally appeared in our Cherry Picked newsletter. Subscribe here: https://cherry-picked-gigi.beehiiv.com/subscribe 🍒