Key Takeaways

What AI Actually Does in Google Ads Now

Google Ads in 2026 is a fundamentally different product from Google Ads in 2020. The platform has progressively pushed account management toward AI-driven automation, and the manual levers that used to define skilled PPC management — exact match keywords, manual bid adjustments, granular audience targeting — have been deprecated, deemphasized, or wrapped in AI layers. The result: it’s now possible to launch a Google Ads campaign without ever opening the advanced settings, set the goal to “maximize conversions,” and let Google’s AI handle bidding, audience selection, ad rotation, and even ad copy generation.

The specific AI features that matter in 2026:

• Smart Bidding — AI-driven bid strategies (Maximize Conversions, Maximize Conversion Value, Target CPA, Target ROAS) that adjust bids in real time based on the user, query, device, location, time, and dozens of other signals.
• Performance Max — AI-driven campaign type that optimizes across Google’s full inventory (Search, Display, YouTube, Gmail, Maps, Shopping). The bidding, placement, audience, and creative are all managed by AI.
• Demand Gen — successor to Discovery campaigns, AI-driven for top-of-funnel inventory.
• Automatically created assets — AI generates headlines, descriptions, and sitelinks based on your landing pages.
• Broad match with AI bidding — when paired with Smart Bidding, broad match keywords are now Google’s recommended targeting model (a complete reversal from the 2018 advice).
• Asset optimization — AI determines which combinations of headlines and descriptions to serve to which users.
• Audience signals — AI uses your first-party data, customer lists, and behavioral signals to find lookalike audiences automatically.

The collective effect: Google has moved Google Ads management from a craft (skilled humans making granular decisions) toward an interface (humans setting goals and budgets, AI doing the execution). This is real progress in some ways. It’s also created new failure modes that didn’t exist before. The honest question isn’t “is AI good or bad” — it’s “what does AI do well, what does it do badly, and how should a Chicago small business actually run their account in 2026?”

Where Google’s AI Genuinely Wins

The most common counter-take to “you need a PPC manager” is dishonest — it pretends Google’s AI hasn’t gotten better. It has. The cases where AI genuinely outperforms manual management in 2026:

Bid optimization at scale. Smart Bidding makes hundreds of bid adjustments per hour, across thousands of signal combinations, in ways a human couldn’t replicate. For accounts with enough conversion data (50+ conversions in 30 days), Smart Bidding generally outperforms manual bidding on cost-per-conversion. This is real.

Audience expansion. AI is better than humans at finding lookalike audiences from your first-party data, identifying behavioral patterns in non-obvious customer segments, and adjusting audience targeting based on conversion patterns.

Real-time signal processing. Adjusting bids and creative selection based on time of day, day of week, weather, device, recent user behavior — AI handles this at a granularity that’s impossible manually.

Multi-asset rotation. Given 12 headlines and 4 descriptions, AI selects the combination most likely to convert for each individual impression. A human can’t do this in real time.

Cross-network optimization. Performance Max, when it works, can find conversions on inventory you’d never have targeted manually (YouTube, Display, Gmail) at acquisition costs better than direct buying on those surfaces.

Quality Score and Ad Rank optimization. AI is genuinely good at finding the small adjustments that lift Quality Score across thousands of ad-keyword pairs.

If you’re running a Google Ads account in 2026 and you’re not using Smart Bidding, you’re leaving real performance on the table. If you’re running a high-volume e-commerce account and you’re not testing Performance Max, you’re probably underperforming. The PPC manager who refuses these tools because “I prefer manual control” is fighting a war they’ve already lost — and costing the client real money.

Where Google’s AI Quietly Fails

The failure modes that matter aren’t dramatic. AI doesn’t crash your account or lock you out. It quietly does the wrong thing efficiently. The category that matters here is strategic failure — AI executing a goal that was wrong before the account ever launched. The tactical wasted-spend patterns (irrelevant search terms, geo leaks, duplicate audiences) sit in a separate section below; this section is about the higher-order judgment errors only a human catches.

The wrong goal, optimized perfectly. A Schaumburg HVAC contractor sets “maximize conversions” with the conversion defined as “form submission.” Google’s AI efficiently finds people who submit forms — including spammers, competitors gathering intel, students researching school projects, and price-shoppers with no intent to hire. Cost-per-conversion looks like it’s improving while actual booked jobs decline. The account dashboard tells one story; the contractor’s calendar tells a different one. The business is losing money faster, not slower.

Performance Max black box. When Performance Max works, it can be magical. When it fails, you have very limited tools to diagnose why. The placement reports are obscured, the audience targeting is automatic, and the search term reports are partial. We’ve audited Performance Max campaigns where a large share of the budget went to Display inventory the business would never have chosen — but the dashboard showed “conversions improving” so the business kept funding it. Without a human auditing the channel allocation, the trap is invisible.

Brand voice and compliance drift. Auto-generated ad copy pulls language from your landing pages and combines it in ways that occasionally produce off-brand or non-compliant ads. We’ve seen AI-generated ads for legal clients use prohibited language (FREE consultation when Illinois bar rules require disclaimers), AI-generated ads for healthcare clients claim outcomes that violate HIPAA marketing rules, and AI-generated ads for Naperville-area contractors include guarantees the business doesn’t actually offer. None of these get flagged by Google’s policy checks; they only get caught by a human reading the live ads.

Landing page mismatches AI won’t notice. Your ad promises “emergency garage door repair in Lockport.” Your landing page is a generic “we serve all of Chicagoland” page. Conversion rate is terrible because the landing page doesn’t match the ad intent. Same pattern in reverse: a Wicker Park bathroom remodel ad lands on a homepage that talks about all services equally — the user can’t tell if they’re in the right place. AI bidding will keep paying for clicks because the conversion is technically happening at a much higher cost than it should be. A human catches the mismatch in a 10-minute landing page audit.

Account structure decay. AI campaigns get added, paused, and reactivated. Asset groups multiply. Negative keyword lists go stale. Conversion actions accumulate. After 12–18 months of automated management, the account is a sediment of decisions nobody remembers making. We’ve audited Chicago SMB accounts with dozens of conversion actions where only a few were the ones the business actually cared about. The AI was confidently optimizing for “thank-you page view” events that had been broken since the last site redesign.

Seasonality misreads. AI bidding handles within-week seasonality well. It handles annual seasonality (back-to-school, holiday, tax season, Q4 home improvement) badly because the learning window is too short for annual patterns. A Naperville landscaper running AI-only ads will overspend in March (when the AI sees increasing search volume but doesn’t know it’s seasonal demand that will end) and underspend in September (when the AI sees declining conversions and pulls back, missing the late-season fall cleanup demand). A human PPC manager who’s worked one full year in Chicago knows the patterns; the AI has to re-learn them every cycle.

The pattern: AI fails at things that require business context, judgment, or signal beyond what’s in the account. AI is great at the optimization problem within the goals you set. It’s bad at deciding whether those goals are the right ones.

The Performance Max Trap

Performance Max deserves its own section because it’s the most aggressive expression of “let AI run everything” — and the most consequential decision a Chicago SMB makes about their account in 2026.

The pitch is compelling: one campaign, all inventory, AI does the work. For some account types (especially e-commerce with strong product feeds and clear conversion economics), Performance Max can genuinely outperform a portfolio of traditional Search, Display, and Shopping campaigns. We’ve seen it work. Google’s own Performance Max documentation is explicit that the campaign is goal-based: you set the goal and the AI decides how to reach it.

The trap is what you lose:

• Search term visibility. You can see some of the queries that triggered ads, but not all of them. The blackbox is partial.
• Placement control. You can’t see (or exclude) which specific YouTube channels, Display sites, or Gmail placements your ads ran on, except in aggregate.
• Audience attribution. The “audience signals” you provide are treated as suggestions, not constraints. AI will spend on audiences you didn’t specify.
• Brand safety. Display inventory and YouTube placements are vast; some of it is content you’d never associate with your brand voluntarily.
• Conversion attribution. When Performance Max takes credit for conversions, it’s often grabbing credit from existing branded search traffic or remarketing. The reported “incremental” conversions are sometimes not incremental at all.

The Performance Max trap is that the dashboard looks great while the underlying spend allocation is increasingly disconnected from what’s actually driving business outcomes. We’ve audited Chicago accounts where Performance Max was reporting a high ROAS in the dashboard while the business owner was reporting that booked leads had dropped over the same period. The campaign was attributing conversions that would have happened anyway through branded search; the actual incremental performance was negative.

The right approach for most Chicago SMBs in 2026: one or two Performance Max campaigns alongside traditional Search campaigns where you maintain control. Use Performance Max for the tail of inventory and conversion sources you couldn’t realistically buy directly. Use traditional Search for the high-intent commercial queries where you want to see exactly what you’re paying for. We covered the broader Google Ads budget strategy for Chicago SMBs in its own post.

Lead Quality vs Lead Volume: The Hidden Cost

The single biggest failure mode of AI-managed Google Ads — and the one most Chicago SMBs don’t realize is happening — is the lead quality decay that follows AI optimization toward higher lead volume.

The mechanism: AI optimizes for the conversion event you’ve defined. If your conversion event is “form submission” or “phone call lasting more than 30 seconds,” AI will efficiently find users who submit forms or make 30-second calls. The problem is that lead quality is bimodal — some of those conversions are real prospects who’ll become customers; some are spam, junk, or low-intent inquiries that consume your sales team’s time without producing revenue.

A human PPC manager auditing lead quality monthly will notice:

• Form submissions from competitors using generic email addresses
• Phone calls from telemarketers, vendors, or sales pitch agencies
• Inquiries from outside your service area
• Price-shopping leads who never close
• Job seekers who interpret your ad as a hiring page
• Spam submissions from automated bots that pass reCAPTCHA

AI won’t catch any of these. It will see the conversion event and continue optimizing toward the user patterns that produce them. Over 60–90 days, the account drifts toward lower-quality lead sources because they’re cheaper to acquire — even though they’re less valuable to the business.

The fix requires three things AI can’t do:

1. Define lead quality outside the ad platform. Connect CRM data, sales outcomes, or call tracking quality scoring back to Google Ads as a separate signal.
2. Audit leads manually. Once a week, a human reviews the leads that came in and tags them quality (real prospect, junk, price shopper, etc.).
3. Adjust strategy based on the audit. Pause ad groups producing junk. Add negative keywords. Tighten audience targeting. Improve form qualification questions.

A specific case: a Lincoln Park bathroom remodeler we audited had been running AI-only Google Ads for a year. The dashboard showed lead volume up year-over-year. The actual sales pipeline was flat. When we audited the leads, roughly half were price-shopping outside the firm’s target market, vendor outreach, or competitive intel. After we added offline conversion uploads (telling Google which leads actually closed), tightened the geographic targeting to Lincoln Park, Lakeview, Wicker Park, and the North Side neighborhoods that mapped to the firm’s actual jobs, and added negative keywords for adjacent low-margin services, total “conversions” dropped meaningfully — but qualified booked jobs went up. The AI was efficiently buying the wrong kind of conversion until a human told it what “good” actually meant.

The Wasted Spend AI Won’t Catch

Beyond lead quality, the other category of damage AI doesn’t catch is plain wasted spend — money going to clicks, impressions, or placements that have zero chance of converting. The most common patterns in Chicago SMB accounts we audit:

A real PPC manager runs these audits monthly. AI does not. On Chicago SMB accounts we’ve audited — across Naperville home services, Schaumburg professional services, Aurora retail, Oak Park healthcare practices, and downtown legal firms — the pattern is consistent: a meaningful slice of the ad budget is going to one of the wasted-spend categories above, and once it’s reallocated the spend produces real return for the first time. The categories show up in different proportions across industries, but they almost always show up. The negative keyword guidance from Google is foundational; following it weekly is non-optional for serious management.

Why Google’s AI Inherits Google’s Bias

Anyone who has run Google Ads accounts long enough has had the call with the Google rep. They’re cheerful, well-trained, and quietly convinced you should enable Performance Max, switch to broad match, add display expansion, and raise your daily budget by 30%. The recommendations almost always favor higher spend, more Google products, and looser targeting — almost never tighter geo-targeting, more negative keywords, or moving budget away from underperforming campaign types.

The reason isn’t conspiracy. It’s incentives. Google reps are commissioned salespeople with spend-tied targets. Their job is to grow your account, not to optimize it for your business outcomes. The way Google evaluates their performance is by how much your account spends after the call, not by how many qualified leads you booked, how good your unit economics looked, or whether your sales team had a good month. Most experienced PPC managers learn to take the rep’s advice as one input, run it through their own filter, and politely ignore the parts that don’t fit the business.

The same dynamic plays out with Google’s AI tools — quietly, at scale. Google’s official documentation is half technical reference and half marketing copy. It oversells Performance Max. It downplays the visibility loss that comes with handing control to automation. It promotes broad match as the default targeting model. The “best practice” guides recommend the campaign types and bidding strategies that move budget toward Google’s preferred inventory.

Google’s AI tools are trained on that documentation, plus support transcripts, plus marketing-team-authored explainers, plus the same “best practice” guides. Every recommendation the AI surfaces in your account comes from training data Google curated, shaped by the same incentives that shape what Google reps say on a call. When the AI suggests “expand to broad match for better reach,” or prompts you to “add a Performance Max campaign,” or recommends “increasing your daily budget to capture more impressions” — you’re not getting a neutral optimization. You’re getting Google’s house view, laundered through an interface that feels like a tool.

The implication for a Chicago small business running Google Ads in 2026: the AI is not a neutral advisor. It will recommend changes that benefit Google’s revenue more often than changes that benefit your revenue. A PPC manager who has run accounts across multiple verticals and clients develops the calibration to tell the difference. They know which “best practices” are real and which exist to move spend toward Google’s preferred inventory. They know when broad match works (high-intent, well-funnelled categories) and when it leaks budget (everything else). They know which Google rep suggestions to take and which to politely set aside.

AI doesn’t develop that skepticism, because skepticism wasn’t in the dataset. It can’t be — it would contradict the entity that trained it. The same way a Google rep can’t tell you to spend less on Google Ads, the AI can’t tell you that the campaign type Google promotes might be wrong for your business.

What a Real PPC Manager Actually Does

The honest job description of a Chicago PPC manager in 2026 — what they do that AI can’t:

Strategy and account architecture. Define what counts as a “good lead” for the business. Set realistic CPA targets based on customer lifetime value and margin. Choose which products or services to prioritize based on business capacity. Decide which campaigns run as Performance Max vs traditional Search, structure ad groups around themes AI can optimize within, and set up conversion tracking that actually maps to revenue — not just form submissions.

Lead quality monitoring and feedback loop. Define lead quality criteria, audit incoming leads regularly, and feed the quality signals back to the ad platform via offline conversion uploads or audience exclusions. This is the loop AI can’t close on its own.

Negative keyword and landing page discipline. Weekly review of search terms; ongoing negative keyword updates; landing page audits when an ad’s conversion rate drops. The same UX patterns that lift SEO conversion apply to PPC landing pages — covered in our UX for SEO post.

Creative oversight and compliance. Write or oversee ad copy that’s on-brand, differentiated, and compliant — especially in regulated verticals (legal in Illinois, healthcare under HIPAA, financial under FTC rules). Use AI variations for testing, not as the primary copy source.

Anomaly diagnosis. When performance shifts, diagnose whether it’s an algorithm change, seasonality, competitor activity, landing page issue, or creative fatigue. AI flags the change; only a human diagnoses the cause and chooses the right response.

Cross-channel context. Coordinate Google Ads with SEO, GBP, email, and other channels so they reinforce each other rather than cannibalize. The Chicago digital marketing playbook covers the broader channel orchestration.

Vendor and platform escalation. Manage relationships with Google account reps, dispute incorrect billing, handle policy violations, request account manager escalations. This alone often justifies the management fee for accounts with significant spend.

The unifying thread: a PPC manager is the strategic and judgment layer on top of AI execution. Not a replacement for AI, not a substitute — the brain that decides what the AI should be optimizing for and notices when something’s wrong.

When AI-Only Is the Right Answer

The honest counter-take: AI-only management is the right answer in some cases. The Chicago SMBs where we’d say “skip the PPC manager, run the AI yourself”:

Very small budgets. Below $1,000/month in ad spend, management fees would consume 30–60% of the budget. Better to put more money into ads than into management. Set up the account carefully once, use Smart Bidding with a reasonable goal, audit lead quality yourself monthly, and accept the suboptimal optimization for the cost savings.

Very simple businesses with unambiguous conversion goals. A one-product e-commerce store with consistent margins and a single conversion event (purchase) can run AI-only effectively. The AI’s optimization goal maps cleanly to the business outcome. No lead quality ambiguity. No service capacity issues. Performance Max with a strong product feed can work well here.

Owner-operators with PPC time and aptitude. A Chicago small business owner who’s genuinely willing to learn the platform, audit lead quality weekly, maintain negative keyword lists, and stay current on Google Ads changes — can manage their own account effectively. Not most owners, but some. The cost is the owner’s time, which is usually better spent on the business itself, but it’s a real option.

The alternative is a bad PPC manager. This is the most important one to be honest about. A cheap or careless PPC manager will lose more value than they create. If the choice is between a $300/month “set it and forget it” service and Google’s AI, take Google’s AI. The damage a bad manager does — wrong campaign structure, broken tracking, ignored account, poor strategy — is worse than no manager at all. The hire is only worthwhile if the manager is genuinely good.

The Chicago SMBs that should hire a PPC manager are the ones spending $1,500+/month on ads, with multi-step lead generation, multiple services or products, real margin per customer, and the ability to actually use better leads if they’re delivered. That’s most established businesses in Naperville, Schaumburg, Aurora, Lincoln Park, Wicker Park, Oak Park, Evanston, and the rest of Chicagoland — service businesses with operational capacity, retailers with margin, and professional firms with case selectivity. For everyone else — the very early-stage, the very small-budget, the one-product e-commerce — the AI alone is the right starting point, at least until the account grows into the management justification.

What a Chicago PPC Manager Costs in 2026

Honest cost ranges for PPC management in Chicago:

For most Chicago SMBs, agency or freelance retainer at the $1,500/month spend tier upward is the right model. In-house PPC hires are premature until ad spend exceeds $25K/month and there’s enough volume to justify a full-time role.

What to avoid in pricing models:

• Spend-based pricing only at very low spend tiers. A 20% fee on $500/month is $100 — not enough to fund real management. The manager is incentivized to push higher spend, not better results.
• Flat fees without scope. “We charge $1,000/month” without defining what’s included — sets up the manager to do the minimum.
• Performance-based fees alone. Sounds great but creates incentive misalignment (manager pushes high-converting but low-quality leads). Acceptable as a bonus on top of base fees; not as the entire fee.
• Setup fees with no clear deliverable. “We charge $2,500 to launch your campaign” — defensible if there’s a clear scope; usually a markup.

The right pricing model for most Chicago SMBs: a fixed monthly retainer with clear scope, monthly reporting, and a 60–90 day evaluation window.

DIY Framework If You’re Not Ready to Hire

If you’re spending less than $1,500/month on Google Ads and want to run the account yourself with AI assistance, here’s the honest framework:

Weekly (15 minutes):

• Review search terms report; add irrelevant queries to negative keyword list
• Spot-check lead quality from forms submitted that week
• Check for any campaigns that ran out of budget or hit spend caps
• Review any policy or disapproval notifications

Monthly (60–90 minutes):

• Pull the account-level report: spend, conversions, cost-per-conversion, conversion rate
• Audit lead quality from the past 30 days; tag leads as real, junk, or unclear
• Review Performance Max placement and audience data
• Adjust bid strategies if data volume supports it
• Update ad creative if performance is declining
• Test 1–2 new ad variants per campaign

Quarterly (3–4 hours):

• Full account audit: campaign structure, conversion tracking, tag verification
• Negative keyword list review
• Landing page audit on top 5 ads
• Budget reallocation based on ROAS by campaign
• Competitive review (run a competitor’s search terms; see what they’re bidding on)

Annually:

• Review of overall PPC strategy: budget level, channel mix, account structure
• Conversion tracking architecture review
• Consider whether the account has grown into management justification

The skills to learn first if you’re DIY: how to read a search terms report, how to set up conversion tracking that reflects real business outcomes (not just form submissions), how to audit lead quality outside Google’s platform, and how to interpret Quality Score and Ad Rank reports. Skip “how to manually adjust bids” — Smart Bidding handles that; your job is to set the right goal and feed it good signals.

How to Evaluate a PPC Manager

When hiring a PPC manager — agency, freelancer, or in-house — the diligence questions that separate good from bad:

Questions to ask any PPC manager before hiring:

Red flags in the relationship:

• Promises a specific ROAS or CPA before seeing the account
• Won’t share access to the actual ad account (only manages through a wrapper)
• Resists or delays giving you admin access to your own account
• Reports vanity metrics (impressions, click-through rate) without business outcome ties
• Refuses to discuss strategy, only tactics
• Pushes higher spend without business justification

Green flags:

• Asks about your business model, margins, and capacity before talking tactics
• Wants to see your sales pipeline and CRM data, not just account access
• Talks about what they wouldn’t recommend, not just what they would
• Has examples of clients they’ve recommended away from ads when ads weren’t the right channel
• Can articulate where AI helps and where it hurts in 2026

For a deeper take on agency evaluation across SEO and ads together, see our how to choose a Chicago SEO agency post — the principles transfer.

Where to Start

For a Chicago small business — whether you’re a Schaumburg HVAC contractor, a Lincoln Park dental practice, a Naperville law firm, or a downtown e-commerce retailer — deciding whether to hire a PPC manager or run Google’s AI alone in 2026:

1. Honestly assess your ad spend tier. Below $1,000/month, AI-only is usually right. Above $1,500/month with multi-step lead gen, a manager pays for itself.
2. Audit your current account before deciding. What’s the lead quality actually look like? What’s the search term report showing? Where’s the wasted spend? If the answers reveal real problems, hiring a manager will produce measurable improvement.
3. Don’t hire a bad manager. A cheap or careless PPC manager is worse than Google’s AI defaults. Use the evaluation framework above. Be willing to interview 3–5 managers before choosing.
4. Set a 60–90 day evaluation window. Give a new manager time to let AI bidding learn, negatives accumulate, creative refresh, and seasonality wash out. Evaluate against specific metrics agreed in advance.
5. Maintain owner-level visibility. Even with a great manager, the owner should see the account monthly, understand the strategy, and ask questions when something doesn’t make sense. The best managers welcome this; the bad ones resist.
6. Build in cross-channel coordination. PPC is one channel. The lift compounds when ads, SEO, GBP, and landing pages reinforce each other. We covered the Google Ads vs SEO tradeoff and the contractor-specific Google Ads playbook in their own posts.

The bigger frame: AI tools in Google Ads aren’t going away — they’re getting better. The PPC managers who’ll be valuable in 2027 are the ones who understand the AI tools deeply, use them well, and know exactly where human judgment still earns its fee. The PPC managers who refuse to use AI features will lose. The businesses that trust AI completely will leak money to optimization problems AI can’t see.

If you’d like a free audit of your current Google Ads account — including lead quality assessment, wasted spend identification, Performance Max audit, and a recommendation on whether you need management — request one at /contact. Our Chicago Google Ads management covers what an engagement looks like end to end. The right answer for your business depends on your spend tier, complexity, and what you’re actually trying to achieve — not on a generic “always hire a manager” or “always use AI” rule.