Ten years ago, a media plan was a static document: a spreadsheet of channels, budgets and flight dates, locked at launch and revisited monthly at best. Today, a well-run programmatic campaign reallocates spend across channels, audiences and creative variants multiple times per hour, guided by algorithms processing far more signal than any human planner could track manually. The mechanics of media buying have been substantially automated. What hasn’t been automated is deciding what the algorithm should be optimizing for in the first place.
What AI Has Taken Over
The manual, repetitive layers of media buying have moved to machines faster than almost any other part of the agency function:
- Real-time bid optimization across programmatic exchanges, adjusting bids by the millisecond based on predicted conversion likelihood.
- Audience targeting refinement, continuously narrowing or expanding segments based on live performance rather than quarterly audience research.
- Cross-channel budget reallocation, shifting spend toward whatever combination of channel, placement and creative is currently converting best.
- Fraud and waste detection, flagging invalid traffic and underperforming placements far faster than manual audits ever could.
Agencies still running media plans manually against this are not just slower — they are structurally unable to compete on efficiency with a client running algorithmic buying through their own in-house desk.
Why the Planner Role Isn’t Disappearing
It’s tempting to read this as the end of the media planner. The reality is closer to a redefinition. Algorithms are extremely good at optimizing toward whatever metric they’re given — and dangerously good at optimizing toward the wrong one if a human hasn’t set the strategy correctly.
Algorithms Need Strategic Guardrails
Left alone, an optimization algorithm will chase short-term conversion signals even when that undermines brand positioning or long-term customer value. Media planners are increasingly responsible for setting the guardrails — which audiences are off-limits, which placements protect brand safety, what balance between short-term performance and long-term brand building the algorithm should hold.
Someone Has to Interpret the “Why”
AI can tell you that a campaign’s performance dropped 12% overnight. It is far less reliable at explaining why — a competitor launch, a cultural moment, a creative fatigue curve, a platform policy change. That diagnostic, contextual judgment remains a human-led function, now applied to algorithmic output rather than raw spreadsheet data.
Cross-Channel Strategy Still Requires a Human System View
AI systems typically optimize within a channel extremely well but struggle to reason about the brand’s entire ecosystem — how paid media interacts with PR momentum, organic social sentiment or a product launch calendar. Connecting those dots across disciplines remains squarely an agency function.
The New Media Planning Skillset
Agencies rebuilding their media function around AI are hiring and training for a different profile than five years ago:
- Algorithmic literacy — understanding how bidding and optimization models actually behave, well enough to question and adjust them rather than treat them as a black box.
- Data storytelling — translating dense, real-time performance data into a clear narrative a client can act on.
- Strategic restraint — knowing when to let an algorithm run and when to intervene before short-term optimization damages brand equity.
What This Means for Agency Clients
Clients evaluating agency media capability should be asking less about which platforms an agency can access — that access is increasingly commoditized — and more about how the agency governs its algorithms: what guardrails are in place, how performance is interpreted beyond the dashboard, and how media strategy stays connected to the rest of the brand’s marketing ecosystem. The automation is table stakes now. The judgment layered on top of it is where agencies still earn their fee.
