For years, being great at performance marketing meant being fast with a mouse, managing complex campaign structures in an account, running 500+ campaigns. You split-tested headlines by hand, nudged bids up or down by 10% of tCPA every two days up a few cents at a time. That version of the job is different now. AI has significant control of the levers now, for instance, smart bidding strategies, portfolio bid strategies in Google ads, whereas in Meta, you leave it to the algorithm to do the heavy lifting, while you pump out creatives. There are more and more of such use cases in targeting, bidding, and a growing share of the creative itself, and the skill that separates strong marketers from mediocre ones has moved somewhere else entirely.

Bidding stopped being a manual craft

Google’s own numbers make the shift concrete. Search campaigns using Smart Bidding Exploration see 27% more unique converting users on average, and that same capability is rolling out to Performance Max and Shopping campaigns. Campaign total budgets, which let advertisers set a budget across a flexible date range instead of a fixed daily cap, cut manual budget adjustments by 66% on average compared to daily budgets. Google is now pushing this further with demand-led pacing, which shifts spend toward high-demand days automatically while staying inside monthly caps.

Journey-aware bidding, still in beta, extends this logic to messier conversion paths: phone calls, form fills, newsletter signups, the whole non-biddable trail that used to be invisible to the bidding algorithm. Feed it the full journey and it optimizes toward outcomes that used to require a human analyst stitching together a spreadsheet of assisted conversions.

The practical effect is that manual CPC bidding is becoming a thing of the past. The PPC guys or gal’s job on the bidding side has shifted from adjusting numbers to feeding the algorithm better signals: cleaner conversion tracking, feeding more deeper CRM data like Qualified leads, the right optimization goal, enough budget to let the system learn.

Creative is becoming the real MOAT and AI-scaled creatives are…well, a part of it

On one hand, the adoption numbers are no longer early-stage. 83% of ad executives have deployed AI somewhere in their creative process, up from 60% just two years ago, and 86% of buyers are using or planning to use generative AI for video ad creative specifically. Dynamic Creative Optimization, in cases where Google ads and Meta ads make AI versions of your creatives and push into your campaigns whether you like it or not, swapping creative elements in real time based on audience signals, delivers a 32% higher click-through rate and a 56% lower cost per click compared to static creative. A joint study from researchers at Columbia, Harvard, TU Munich, and Carnegie Mellon even found AI-generated ads outperforming human-made ones on raw click-through rate: 0.76% versus 0.65%.

On the other hand, audiences haven’t caught up to advertiser enthusiasm, and the gap is wide enough that it should change how you brief creative. Only 45% of Gen Z and Millennial consumers feel positive about AI-generated ads, while 82% of ad executives assume their audience feels that way. Nearly 40% of Gen Z describes AI-made ads as negative, and 30% call AI-using brands outright inauthentic. Cost efficiency, not creative innovation, is now the top reason advertisers cite for using AI, a reversal from 2024 when it ranked fifth.

Put those two facts next to each other and the lesson for a manager is specific: AI creative tools can lift performance metrics of your team, while quietly damaging brand perception and reputation if the output looks and sounds generated. The teams winning with this aren’t the ones generating the most variants. They’re the ones using AI to produce more raw material faster, then applying the same editorial judgment to that material they’d apply to anything else before it goes live.

Is advertising ROI shifting?

Here’s where practitioners should slow down. A lot of the AI marketing ROI figures circulating right now measure different things and get flattened into one tidy narrative. McKinsey’s 2026 survey found 37% of respondents attribute a positive EBIT contribution to AI, but that’s a cross-industry enterprise figure, not a performance marketing return multiple, and it says nothing about the 63% who didn’t report that contribution, who may simply have no visibility into the number yet.

Before you build a budget plan or KPIs for your PPC team shaped around AI implementation, think twice. Track what a specific AI-assisted workflow, for example: automated bid strategy, AI-generated UGC creatives for Meta ads, or enabling AI Max feature in Search ads and hoping it would save your time and AI will take care of finding high intent traffic – actually costs against what it replaces or improves, including the review time it still requires. 

What this means for you when planning ahead

Pulling the threads together, a few shifts stand out for anyone running performance campaigns right now:

Manual bid management is slowly becoming a legacy skill in some cases. The value-add has moved to signal quality: conversion tracking hygiene, feeding the full customer journey into bidding systems, and choosing the right optimization target rather than tweaking numbers by hand.

Creative production speed is no longer the differentiator. With 73% of advertisers already using AI for display and 85% for social, raw output volume is table stakes. Brand authenticity and editorial oversight are the new competitive edges, especially with younger audiences who can spot generated creative and penalize brands for it.

Disclosure practices are inconsistent and getting riskier to ignore. 89% of advertisers using generative AI disclose it at least sometimes, but less than half always do. As regulation catches up, that gap is a liability waiting to surface.

Treat vendor ROI claims as a starting hypothesis, not a budget line. Build your own measurement before committing spend based on an industry average that was never calculated for your business.

AI picked the targets and set the bids. It’s drafting the ads too, often well enough to beat a human on click-through rate alone. What it still can’t do is know when fast and cheap starts to look fake to the people you’re trying to reach. Building a better positioning is the key. Building a good funnel is key. Doing measurable experimentation and A/B tests is key, AI can’t do that. The human and skilled approach to understand the gaps between the user journey and where drop-offs are happening, leading to low CTR, low CVR, very high CPA.