Performance Max and Advantage+ now pick your client's bids, audiences, and creative mix automatically. Here's the AI audit that proves what your paid media retainer still does that the algorithm can't.
by Ayush Gupta's AI
The problem
Paid media retainers used to be billed on campaign management: building ad sets, segmenting audiences, running bid strategy, and testing creative variants by hand. Google's Performance Max and Smart Bidding, and Meta's Advantage+ campaigns, now do most of that natively inside the platform, automatically, as the default recommended setup. The client doesn't need to ask the agency to 'test a new audience' anymore — the platform already tested dozens overnight and picked one. For a lot of paid media and performance marketing agencies, the manual, billable-hours version of the job that justified the retainer size is quietly disappearing into the ad platform's own algorithm, and the first sign is usually a client asking why monthly campaign management hours look the same as they did two years ago when the dashboard is doing most of the optimizing now.
The fix
Build an AI-assisted audit that separates what Performance Max, Smart Bidding, and Advantage+ now handle natively from the strategic, creative, and measurement judgment work the algorithm structurally can't do, then reprice and reposition the retainer around that judgment layer before the client runs the same comparison without you.
The Playbook
List every automated campaign feature actually running on the account right now
Pull the current campaign setup for every active account: which campaigns run on Performance Max or Advantage+, which use Smart Bidding or automated bid strategies, which use dynamic creative or automatically generated asset combinations. Most agencies underestimate how much of the account has quietly shifted to automated defaults over the past year, because each individual setting change felt small at the time.
Classify every billed campaign-management task against what the platform now does for free
Take the actual tasks in your SOW or time-tracking log and have Claude sort them into what the platform's automation now covers versus what still requires a human call.
Here is our current scope of billed campaign management tasks for a paid media retainer:
[PASTE SOW LINE ITEMS OR TIME-TRACKING TASK LIST]
Here is what's currently running on this account: [LIST AUTOMATED CAMPAIGN TYPES — e.g. Performance Max, Advantage+, Smart Bidding, dynamic creative].
For each task, classify it as:
1. Platform-automated — Performance Max/Advantage+/Smart Bidding now does this natively, with minimal human input required
2. Judgment work — requires context, strategy, or a call the algorithm can't make on its own (budget allocation across channels, audience exclusions tied to brand risk, creative strategy and testing hypotheses, interpreting results the platform's own attribution can't be trusted to interpret)
For every "platform-automated" item, note what's still worth doing manually even though the platform handles the baseline (e.g. feeding it better inputs, catching when it optimizes toward the wrong goal).Find the gap the automation can't see: budget strategy, exclusions, and incrementality
Performance Max and Advantage+ are built to optimize within the budget and goals they're given — they don't decide whether paid media should get more or less budget than organic or lifecycle this quarter, they don't know which audience segment carries brand risk even if it converts, and they can't tell you whether the conversions they're claiming credit for would have happened anyway. That's the real judgment layer, and it's specific to each account, not a generic pitch line.
For this account, list the channel and budget-allocation decisions, audience exclusions, and attribution/incrementality questions that Performance Max or Advantage+ cannot answer on their own, because they require context outside the platform:
Account context: [PASTE BUSINESS GOALS, BRAND SENSITIVITIES, CROSS-CHANNEL BUDGET SPLIT]
Be specific to this account, not generic. Flag anything where the platform's own reported results are likely overstating impact.Reposition the retainer narrative before the client runs the comparison
Go into the next performance review having already reframed the relationship from 'we manage your campaigns' to 'we make sure the automation is optimizing toward the right goal, catches what it can't see, and doesn't quietly waste budget on the wrong exclusions.' That pitch survives a client comparing your hours to what the dashboard already does for free, because it isn't competing with the dashboard.
Re-run the audit every quarter — Google and Meta ship new automation defaults constantly
Performance Max and Advantage+ both expand in scope every few months, often turned on by default for new campaigns. A quarterly check of what's newly automated keeps the retainer narrative current instead of getting caught flat when a client's procurement team notices the gap first.
What changes
A retainer that's defensible because it's priced around judgment the platform can't replicate, not campaign-management hours the platform already automates. Agencies get ahead of the client's own realization instead of discovering it at a budget review.
A paid media retainer used to be priced on labor: building out ad sets, segmenting audiences by hand, running bid strategy, testing creative variants one at a time. That was real, billable work, and it was the reason the retainer was sized the way it was.
Performance Max and Smart Bidding now run most of that automatically, as the default setup Google recommends. Meta's Advantage+ does the same thing on the social side — automated audience expansion, automated creative combinations, automated budget allocation across placements. The client doesn't need to ask for a new audience test. The platform already ran dozens overnight on its own.
Why this is different from losing a pitch
Nobody announces a decision to cut a paid media retainer because of automation. It shows up as a client asking, almost casually, why campaign management hours look the same as two years ago when "the dashboard seems to be doing most of it now." That question is reasonable, and an agency without a real answer loses the argument before it starts.
What the automation still can't do
Performance Max and Advantage+ optimize inside the budget and goals they're given. They don't decide whether paid media deserves more or less budget than organic or lifecycle this quarter. They don't know which audience segment carries brand risk even if it converts well. And they routinely overstate their own impact, claiming credit for conversions that would have happened anyway — something only a human running an incrementality check will catch. That's the judgment layer, and it's specific to the account, not a line that works in every pitch deck.
The fix is an honest audit, not a bigger pitch deck
Classify every billed task against what the platform now automates versus what still needs a human call. Be honest about the split — some of what agencies bill for genuinely is gone. Then rebuild the retainer narrative around the part that isn't: budget strategy across channels, exclusions tied to brand risk, and catching when the platform's own reported results are lying about impact.
Bottom line
The agencies losing paid media retainers right now aren't losing them to a competitor. They're losing them to the ad platform's own automation, one budget review at a time, because nobody put the judgment work into words before the client did the math themselves. Running this audit before the renewal conversation is cheaper than losing the account to a feature Google shipped for free.