The new PPC skill set: From keyword manager to system optimizer

The old PPC skill set was built around control: define the keywords, choose the match types, set bids, write tightly aligned ad copy, and structure campaigns so the algorithm behaved the way you wanted.

The best ad managers of the past were great at Excel and pivot tables. Execution was the product and the differentiator for agencies and PPC experts. The more precisely you could control the variables, the better you were at the job, and that approach worked for the first decade of PPC.

Google Marketing Live (GML) 2026 made the next phase of PPC much harder to ignore. The biggest updates point to a shift from tactical control to system optimization, from keyword management to signal design, and from campaign setup to machine-aligned strategy.

The skills AI-driven Google Ads rewards

AI Max for Search is now out of beta. Smart Bidding Exploration is expanding into Shopping. Demand-led budget pacing will automate when your budget gets deployed. Business agent for leads can now qualify prospects directly inside a search conversation before anyone clicks your ad. Ads are showing up inside AI Mode conversations, matched not to keywords, but to conversational context Google’s AI interprets in real time.

The execution layer is being replaced outright. Selin Song, president of Google Customer Solutions, stated it directly during the keynote:

• “But things are changing. Execution is becoming a commodity and will no longer be a competitive advantage.”

Here’s what the new skill set looks like.

Your customers search everywhere. Make sure your brand shows up.

The SEO toolkit you know, plus the AI visibility data you need.

Start Free Trial

Get started with

Input design: The new keyword research

You need to know what inputs to give the system so it can find the right people on your behalf.

AI Max for Search, now out of beta and rolling out broadly, is a new Google Ads feature that uses a combination of broad match, keywordless targeting, text customization, and final URL expansion to find queries your keyword list never would’ve surfaced.

According to Google’s internal data, accounts using AI Max with text customization and final URL expansion see an average of 7% more conversions or conversion value at a similar CPA/ROAS.

That number is easy to wave away. What’s harder to ignore is the underlying mechanic: AI Max is finding converting queries your keyword list missed. The system has more access to user context than any keyword list you build, and it’ll keep getting better at using that access.

That means the skill is no longer “What keyword should I target?” It’s “What inputs do I need to give this system so it reaches the right people?”

That includes:

• Your conversion data: Smart Bidding can only optimize toward what you tell it matters. If your conversion actions are wrong, incomplete, or proxy metrics that don’t reflect business outcomes, the system is solving the wrong problem, and that’s the ad manager’s fault.
• Your product and feed data: For Shopping and ecommerce, Conversational Attributes — new Merchant Center feed attributes announced at GML for AI Mode surfaces — let you supply Q&A pairs, related products, and popularity signals. The AI uses that data to represent your products inside AI-generated responses. Thin feeds generate thin results. Rich feeds give the system something to work with, and as an ad manager, you need to optimize your feed with those questions at the center of the strategy.
• Your audience signals: New Customer Acquisition modes, also updated at GML, now include a “new prospects mode” that uses automated exclusions to reach brand-unaware users by filtering out people who’ve visited your site, searched your brand, or engaged with your content. This kind of upstream decision — who are we trying to reach? — shapes how the system operates. That’s not a campaign setting. It’s a strategic decision that now falls within the ad manager’s role.

If you’re still clinging to keyword lists as your main targeting strategy, you’re operating in a world that no longer exists. Today’s systems force you to think through business decisions, signal design, and the inputs steering automation.

Dig deeper: 10 keys to a successful PPC career in the AI age

Value signal architecture: The new bid management

The old version of bid management was about moving numbers. Then came automated bidding that factored in signals we couldn’t see. The job became deciding when a maximize strategy made sense versus when a target-based strategy was the better lever for the business.

That work isn’t gone, but the responsibility has expanded. Now it’s about how well you feed the system signals like first-party data, audience quality, and conversion value accuracy.

As Smart Bidding optimizes toward conversion values, you need to factor in a new layer of considerations.

Demand-led budget pacing, announced at GML and coming globally soon, will automate when your budget gets deployed throughout the day based on predicted demand signals. The system captures more on peak days, reduces spend on slower days, all within your monthly limits. You don’t control the pacing. You set the parameters the pacing operates within.

That means you also need to think through the economics of the offer. For example, if your store sells both electronics (10% margin) and home décor (55% margin), and you don’t model margin into your conversion values, Google may pace aggressively on days when electronics sell well even though those sales barely break even.

Product value adjustments, currently in global pilot, push this even further. You can now tell Google’s AI that a specific product, brand, or category should be weighted higher or lower in the auction. 

That helps nudge Smart Bidding toward actual business priorities instead of raw conversion value. You can optimize toward profit, seasonal sell-through, and best-sellers across Performance Max and Shopping campaigns without changing campaign structure.

The skill here is knowing what to signal, not how to set a bid. That requires clarity about:

• Margins: Which products can you afford to be aggressive on? Which low-margin items make aggressive bidding too expensive?
• Inventory position: What needs to move in the next 30-60 days?
• Lifetime value: Which products bring in repeat buyers? Which attract one-time purchasers?
• Cash flow timing: Where do you need revenue now versus where can you afford to be patient?

Journey-aware bidding, also newly out of beta, extends this to lead gen. You can now feed Google’s AI your full conversion journey, not just the final conversion event, and Smart Bidding will optimize across every stage of the funnel.

But to use it effectively, you need a fully instrumented conversion journey and a way to connect customer value back to the ad platform.

System prompting: The new copywriting

Here’s the skill with no real historical analog in PPC, and one of the most underestimated announcements from GML.

AI Brief, powered by Gemini, lets you guide AI Max for Search, Performance Max, and AI Max for Shopping using plain language. You write a brief describing your brand, your customer, your tone, and what to avoid. Google’s AI uses that brief in real time to shape how your campaigns find and represent you.

This isn’t copywriting or keyword strategy. It’s something closer to system prompting: the skill of giving AI enough context to act on your behalf without over-constraining it or leaving it to invent who you are.

Learning to prompt AI seems straightforward on the surface, but it isn’t. It requires attention, iteration, and a willingness to refine your thinking as you go.

Writing a brief that steers the system requires paid ads managers to understand things many advertisers haven’t had to articulate before:

• What makes this brand sound wrong?
• What searches are technically relevant but strategically damaging?
• What does the ideal customer look like in language specific enough to be useful?

Google’s example at GML was Cedar Pantry, a wellness grocery delivery brand. Its brief specified a tone that had to be “warm, calm, and confident, and never promotional,” while explicitly excluding price-driven language like “cheap,” “deal,” “fast,” and “bulk.”

One paragraph. Specific. Defensible. That brief shapes every impression the AI serves.

The practitioners who’ll be good at this aren’t necessarily the best keyword builders. They’re the ones who can distill brand strategy into operating instructions for a system that doesn’t already know the client.

And the strongest PPC experts will do that while maintaining confidence in the expertise they’ve spent years developing.

Dig deeper: The new PPC playbook: From media buyer to profit engineer

Budget architecture: The new budget management

Daily budget management used to be a significant part of the job. Watch pacing. Adjust if you’re under-delivering. Cap spend if you’re burning too fast. Build rules. Check in daily, all while managing the low-level stress of targeting that’s either limiting or overexposing your ad budget.

That’s compressing fast. Campaign total budgets, now generally available globally, let you set a fixed total spend with a defined start and end date. Google’s internal data says advertisers using it saw a 66% average reduction in manual budget adjustments compared to daily budgets.

The manual work has been automated. But a feature that looks great on paper raises a real question: How does a campaign using campaign total budgets perform compared to one using a daily budget?

That’s the part no GML announcement slide answers. Based on what I know about the ad auction, campaign total budgets likely work by forecasting demand across the entire flight and dynamically pacing spend based on predicted value, not daily ceilings.

It’s a prediction-led pacing model. From my experience, campaign total budget campaigns will almost always serve more aggressively on predicted high-value days, while daily budget campaigns will serve more consistently across all days.

That shifts the skill set toward interpreting auction behavior in a predictive system. It’s no longer “this is how the auction works.” It becomes “this is how the auction reacts” when pacing, budgets, and signals shift.

Demand-led budget pacing removes the daily pacing question entirely. The AI decides when to spend based on demand signals. You don’t control the daily rhythm, but you do set the ceiling and the objective.

What you still control is the architecture: how many campaigns share a budget, which budget parameters align with which objectives, and when to give the system room to operate versus when to constrain it.

Missed opportunity reporting, now generally available, provides visual insights into where bid and budget constraints are limiting growth opportunities. The data is there. The question is whether you can interpret it and make structural decisions from it.

Budget architecture is now the skill, not spreadsheet management and daily budget adjustments.

Measurement literacy: The new quality score management

Quality Score used to be the proxy for account health. CTR, ad relevance, and landing page experience were the three signals that told you whether your ads aligned with what users were searching for.

That proxy still matters today. But the upstream measurement question has become bigger and more complex.

Journey-aware Bidding requires conversion imports that reflect your actual funnel, not just the bottom of it.

Smart Bidding Exploration, which now shows 27% more unique converting users on average, only finds those users because it pulls signals from a broader range of performance data. The system’s ability to expand reach depends entirely on the quality of the signals you feed it.

Business Agent for Leads, also announced at GML, pushes this even further. An AI agent can now qualify leads directly inside a search conversation before anyone ever touches a landing page.

The leads those agents capture need to feed back into your bidding system for Smart Bidding to learn from them. That feedback loop doesn’t happen automatically. It requires integration, instrumentation, and someone who understands how conversion data shapes bidding behavior.

The skill is no longer optimizing toward Quality Score. It’s asking:

• What data does this system need upstream to make good decisions downstream?
• Do we have that data?
• How do we work with business partners to align that data with the ad account?

Dig deeper: In Google Ads automation, everything is a signal in 2026

---

A few things that still hold

The skill set is shifting. The fundamentals haven’t changed.

Conversion tracking is still nonnegotiable

Everything above — including AI Brief, Journey-aware Bidding, Smart Bidding Exploration, and Product Value Adjustments — operates on conversion data. If your tracking is broken or measuring the wrong thing, you’re giving the system a bad problem to optimize. Fix the measurement before touching the strategy.

Campaign structure still communicates intent

AI Max, Performance Max, and Smart Bidding Exploration have more room to operate in consolidated account structures with enough data to learn. Fragmented campaign architecture that made sense for manual bidding often works against AI learning now.

The brief you write is only as good as the thinking behind it

AI Brief doesn’t replace brand strategy. It amplifies it. If you don’t know what the client’s brand stands for or what searches would damage it, the brief will be vague, and the AI will behave vaguely.

Human oversight isn’t optional

The new skill set doesn’t remove you from the loop. It moves you to different points in the loop — upstream in the inputs, midcampaign in the signals, and downstream in the measurement. The job of the PPC practitioner is still to be the person who knows what the system should be doing and whether it’s doing it.

Skills that matter even more now

Asking better questions is now a core technical skill.

Predictive systems behave like mirrors. They reflect the clarity, structure, and intent of the questions you ask. If your questions are vague, the system’s behavior will be vague. If your questions are diagnostic and grounded in business reality, the system has something meaningful to optimize toward.

You need to know how to interrogate the system:

• What signal is it prioritizing? 
• What changed in the environment? 
• What does it believe is high‑value right now?

The quality of your questions shapes the quality of the system’s decisions.

Communicating system behavior to stakeholders is now part of the job

As execution becomes automated, the practitioner’s value shifts toward interpretation: explaining why the system behaved the way it did, what inputs shaped that behavior, and what adjustments come next. Stakeholders don’t see the signals, the pacing model, or the predictive logic. They see outcomes.

The role of the PPC expert is to translate volatility into meaning, model decisions into strategy, and system behavior into business language.

This isn’t a soft skill. It’s a survival skill in an environment where the work is increasingly invisible.

See the complete picture of your search visibility.

Track, optimize, and win in Google and AI search from one platform.

Start Free Trial

Get started with

The shift is already here

GML 2026 didn’t preview a future version of Google Ads. It confirmed the version we’re already operating in.

The practitioners who thrive now aren’t the ones who can recite how Google Ads used to work. They’re the ones who understand what the system needs to make good decisions and can provide those inputs clearly, consistently, and strategically to meet business goals.

The job has already shifted from keyword manager to system optimizer.

Dig deeper: What’s next for PPC: AI, visual creative and new ad surfaces