What's left for a performance marketer when AI does the bidding?

(updated ) 6 min read ppcaistrategy
Platforms automate bidding, targeting and creatives. So where is the human value now? On strategy, creative direction and measurement in one pair of hands.

I get this question almost every week: if Google and Meta are soon handling the bidding, the targeting and the ads themselves, what’s left for a specialist? It’s a fair question, because automation is moving fast. Over the course of 2025, Meta made Advantage+ the default for new sales, app and lead campaigns step by step (announced in February 2025, then rolled out globally), and Google’s AI Max keeps extending the AI layer over your search campaigns. My answer is short: the levers are disappearing, the craft isn’t. It is moving to a very different place, though.

What the platforms are really taking over

This isn’t the distant future anymore. At Google, the vast majority of accounts now run on automated bidding. AI Max for Search, announced by Google in May 2025 and showcased shortly afterwards at Google Marketing Live (the announcement came on May 6, 2025 via the Google blog, and Google Marketing Live followed on May 21), puts an AI layer over your search campaigns that matches search terms, generates copy and picks the best landing page on its own. Google reports an average of around 14 percent more conversions at a similar CPA or ROAS in the first tests. Treat that as a direction, not a promise: these are numbers from the company selling the product.

At Meta, the shift is even more radical. At Cannes Lions in June 2025 it announced eleven new AI ad features, and it is openly working toward fully automated campaigns. The goal, as Zuckerberg outlined it: you enter a URL and a budget, connect your bank account, and the AI creates the creatives, picks the audience and adjusts the budget. All that’s left for you is reading the results.

Does that sound like the end of the profession? For people who see the job as adjusting bids and clicking together audiences: yes. For those who look further: no.

Where human value is moving

Automation is exactly as good as what you put into it and how you steer it. That’s where the work is now. Here are four places where a human makes the difference.

1. Strategy and objectives

An algorithm optimizes relentlessly toward the goal you give it, but it doesn’t come up with that goal. If you steer on as many conversions as possible, you get volume, even if half of them are worthless leads. The choice between growth and return, the question of which conversion is actually worth money, the trade-off between channels: that’s strategy, and it stays human work. The AI is a fast executor, not an entrepreneur.

2. Creative direction

Meta can now generate images, video and text on its own, and since 2026 it can even put AI music under your videos. But generation is not the same as direction. What’s your message, which angle resonates with your audience, what promise are you making? The platforms give you controls for this: think of Google’s AI Brief, where you provide messaging guidelines (“never mention prices”), or Meta’s brand kit, which locks in your logo, colors and fonts. Those are the reins a human uses to keep the output on brand. Without direction, you get polished ads that look like everyone else’s.

3. Measurement and data quality

For me, this is the heart of the story. The AI optimizes on the signals you pass along. Feed it messy data and the algorithm will optimize relentlessly toward mess. Duplicate conversions, a broken tag, leads that never become customers but still count just as heavily: all of that poisons your bids directly.

The real lever is better signals. Real conversion values instead of flat counts. Margin instead of revenue. Feeding your CRM data back through offline conversion import, so the algorithm learns from closed deals instead of from random form fills. The more the platforms automate, the more important it is that the measurement underneath is right.

4. Steering and control

An AI campaign is not a Christmas tree you put up and forget about. Someone has to judge whether the output is any good, whether budget is leaking to irrelevant traffic, and when to step in. The platforms are building in more and more controls for this, such as text guidelines and exclusions. But you do have to use those controls actively, otherwise the system optimizes toward the easiest result, not toward your result.

Specialization and measurement in one pair of hands

This is where it ties into how I work. The AI inside Google and inside Meta keeps getting better within its own channel. But no platform solves the connection between channels for you, and certainly not the measurement underneath. In fact, every platform wants you to believe that its own numbers are the truth. Someone has to look past those self-reported numbers and work out what really adds to the bottom line.

That’s why my approach is deep specialization per channel, combined with measurement in the same hands. The person running the campaigns and the person checking the numbers should be the same person. Otherwise you’re steering on separate platform numbers that contradict each other, and the AI happily keeps optimizing on wrong assumptions.

What this means for you as an advertiser

Concretely, here are a few points to get started with:

  • Judge your specialist on different things. Not on how often they adjust bids, but on the quality of their strategy, creative direction and measurement. That’s where the value is now.
  • Invest in your measurement first. Before you put more into the AI, make sure your conversion tracking is right, that you steer on value or margin and that your CRM data flows back. Clean signals are your biggest lever.
  • Give the automation a clear goal. Decide whether you’re steering for growth or for return, and which conversion is really worth money. The AI won’t choose that for you.
  • Keep control of your creatives. Use the platforms’ steering tools, such as brand kits and messaging guidelines, so your ads stay on brand.
  • Don’t blindly trust platform numbers. Keep measurement and execution in one pair of hands, so someone determines what really gets added.

My conviction: automation doesn’t make the profession smaller, it moves it up a level. Away from turning knobs by hand, toward strategy, direction and measurement. Advertisers who focus on that now make the AI work for them. Those who leave it to the default settings let the platform decide what’s good enough. And that’s rarely the same as what’s good for you.

Getting started

Want to know whether your account is ready for this shift and whether the measurement underneath is right? I’m happy to take a sharp look with you.

Free Performance Quick Scan

Sources

Free Quick Scan

Curious what this could bring you?

I review one ad account with read-only access and show you your 3 biggest opportunities. The report is yours to keep.

or book an intro call

Want to win back return?

Start with a Quick Scan and find your 3 biggest opportunities.