AI in PPC

What Does AI Actually Do in Google Ads?

How is AI being used in Google Ads campaigns?

At its simplest, AI in Google Ads does three things: it decides how much to bid for each auction in real time, it decides who sees your ads, and increasingly, it decides what your ads say and where they link to. These three areas correspond to the three main ways AI has changed campaign management over the past few years.

• Automated bidding (Smart Bidding) has been part of Google Ads for several years and is now the default approach for most campaign types. Rather than setting a fixed bid for each keyword, Smart Bidding uses machine learning to assess hundreds of signals in real time – device, location, time of day, search history, audience membership, and more – and adjusts your bid for each individual auction to maximise the likelihood of achieving your goal, whether that’s conversions, conversion value, or clicks. This works well when campaigns have sufficient conversion data to learn from, and requires careful management to get to that point.
• Automated audience and placement decisions have expanded significantly with Performance Max and AI Max campaigns, which we cover in detail below. These campaign types use AI to decide not just how much to bid, but which audiences to target, which channels to appear on, and – in some cases – which version of your website to send traffic to.
• Automated creative and content decisions are the newest and most significant development. AI Max’s auto-generated ad copy and URL expansion features mean Google can now serve ads with headlines it has written itself, linking to pages on your website it has chosen rather than the one you specified. This is either helpful or alarming depending on how well your website and campaign structure reflect your actual business – which is precisely why the quality of human input into these campaigns matters so much.

Performance Max: What It Is and When It Works

What is Performance Max and is it right for SMEs?

Performance Max (PMax) is Google’s all-in-one campaign type that runs ads across every Google channel simultaneously – Search, Shopping, Display, YouTube, Gmail, Discover, and Maps – using AI to allocate budget across all of them towards your conversion goals. You provide creative assets (headlines, descriptions, images, videos) and Google’s AI decides where, when, and to whom to show your ads.

We’ve run Performance Max campaigns for clients and seen strong results – particularly for ecommerce businesses with clear conversion signals, good visual assets, and well-structured product feeds. For this type of business, PMax’s ability to reach potential customers across multiple channels while optimising towards purchase is genuinely powerful and delivers results that Search-only campaigns can’t always match.

The limitations become apparent in specific contexts:

• Niche B2B service businesses often find Performance Max too broad. Without a physical product to anchor targeting, the AI can struggle to identify genuinely relevant audiences in niche markets. A specialist industrial supplier or a B2B SaaS business may find PMax serving ads to audiences that superficially resemble their customers but have fundamentally different intent.
• Limited asset businesses – those without quality video content, good product photography, or varied creative – will underperform in PMax, which relies heavily on having a range of assets to test and optimise across different placements.
• Low conversion volume businesses can find PMax’s learning phase painfully slow. The AI needs a meaningful volume of conversion data to optimise effectively, and for B2B businesses generating five to fifteen enquiries per month, this threshold can take a long time to reach.

The practical guidance: PMax is worth testing for ecommerce and businesses with clear, trackable conversion signals and quality assets. For niche B2B service businesses, well-structured Search campaigns with tight keyword control will typically outperform PMax, at least until sufficient conversion data exists to guide the AI effectively.

Not sure whether Performance Max is right for your campaigns? We’ll give you an honest assessment based on your specific situation.

AI Max: The Latest Evolution in Search Campaigns

What is AI Max and how is it different from Performance Max?

AI Max for Search campaigns is Google’s newest AI-driven campaign feature, distinct from Performance Max in an important way: it stays within the Search channel rather than expanding across all Google properties. AI Max enhances traditional Search campaigns with three main capabilities: broader keyword matching that expands beyond your exact keyword list to capture relevant queries, automatically generated ad copy that Google creates from your website content and existing assets, and URL expansion that can send traffic to the page Google’s algorithm deems most relevant rather than your specified landing page.

The potential upside is capturing relevant search queries you hadn’t specifically targeted, with ad copy and landing pages matched to those queries automatically. The potential downside – particularly for SMEs in niche markets – is that “relevant” as understood by Google’s AI may be considerably broader than “relevant” as understood by someone with deep knowledge of your sector. An AI that doesn’t understand the difference between commercial boiler servicing and domestic boiler servicing can send expensive clicks in entirely the wrong direction.

What is the September 2026 AI Max migration deadline?

This is important for any business currently running Dynamic Search Ads (DSA) campaigns. Google has announced it will stop allowing new Dynamic Search Ads campaigns from September 2026, and will automatically migrate existing DSA campaigns, along with automatically created assets and campaign-level broad match settings, into AI Max. If you are running DSA campaigns and haven’t actively reviewed them, your campaign structure is about to be changed by Google on your behalf.

The recommended action is to review your DSA campaigns now, understand how they perform, and either proactively transition them to AI Max under your own controlled terms or restructure them into standard Search campaigns before the migration happens. Businesses that ignore this deadline and allow the automatic migration may find their campaign structure and budget allocation changed in ways they haven’t anticipated.

If you’re running DSA campaigns and aren’t sure how the September 2026 migration affects you, we can review your account and tell you exactly what to do.

AI Agents: The Newest Development in PPC Management

What are AI agents in Google Ads?

AI agents are a step beyond automated bidding and campaign management. Where Smart Bidding and AI Max make automated decisions within your campaigns, AI agents analyse your account, surface insights, and make recommendations – functioning more like an intelligent advisor than an automated rule engine.

Google launched several AI agents through 2025 and into 2026, and they’re now available to most advertisers. Understanding what each one does is useful if you’re managing your own Google Ads or overseeing an agency relationship:

• Ads Advisor lives inside your Google Ads account and learns from your past interactions and campaign history to provide personalised recommendations over time. Unlike generic best practice suggestions, Ads Advisor is intended to tailor its guidance to your specific account data. It’s rolling out to all English-language Google Ads accounts and is worth exploring if you haven’t already.
• Analytics Advisor connects your Google Analytics data to your advertising performance, surfacing insights about which campaigns are driving genuine business results rather than just clicks and conversions. For SMEs that struggle to connect their ad spend to actual revenue impact, this cross-platform analysis has real value.
• Marketing Advisor is a Chrome browser extension that works across Google Ads, Google Analytics, and Google’s Help Centre, maintaining context across conversations so you can ask iterative questions about your campaigns without re-explaining your situation each time. According to WordStream’s independent testing of AI accuracy for PPC, Google’s Gemini-powered tools provided the most accurate PPC guidance among the major AI tools tested – a useful data point when deciding how much weight to give their recommendations.

Beyond Google’s own agents, a range of third-party tools offer more targeted AI analysis. Adzooma provides a consolidated dashboard for Google, Meta, and Microsoft Ads with specific optimisation recommendations and one-click implementation, and has a free tier that makes it accessible for smaller businesses. Optmyzr offers more sophisticated analysis for agencies and larger accounts. Dedicated audit agents, keyword agents, and reporting agents from various providers can handle specific tasks – identifying negative keyword opportunities, generating performance reports, analysing landing page conversion issues – without requiring full campaign management involvement.

The Honest Pros and Cons of AI Agents for SMEs

Are AI agents in Google Ads genuinely useful for small businesses?

This is the question worth asking honestly, because the answer isn’t simply yes or no.

Where AI agents genuinely help SMEs:

They’re excellent at surfacing data patterns that would take hours to find manually. An AI agent that reviews your search terms report, identifies irrelevant query patterns, and flags negative keyword opportunities is doing genuinely valuable work that saves time and improves campaign efficiency. Reporting and performance summarisation – turning complex campaign data into clear, readable insights – is another area where AI agents deliver real practical value for businesses that don’t have a dedicated analyst.

They can identify structural issues in campaigns – disapproved ads, budget limitations, quality score problems, conversion tracking gaps – quickly and without requiring expert knowledge to spot. For SMEs managing their own accounts, this diagnostic function is useful.

Where AI agents fall short for SMEs:

The fundamental limitation is context. An AI agent analysing your campaign data doesn’t know your business, your sector, your customers, or the specific competitive dynamics of your market. 43% of PPC managers in 2026 criticise the lack of granular control over campaign targeting and creatives in Google Ads, according to PPCsurvey.com – and AI agent recommendations often push further towards automation rather than addressing this concern.

For niche B2B businesses in particular, AI recommendations that look correct from a data perspective can be wrong from a sector knowledge perspective. An AI agent that recommends broadening keyword match types because it sees low impression share doesn’t know that your sector has highly specific terminology where a single word separates a qualified lead from an irrelevant click. It doesn’t know that your competitors are bidding on your brand terms. It doesn’t know that your average deal value makes a £200 cost per lead entirely acceptable in your market.

Google’s own recommendations within the platform are also worth treating with appropriate scepticism. Google’s recommendations are generated with Google’s revenue interests as a significant factor – they consistently push towards broader targeting, higher budgets, and more automation. Accepting all recommendations without reviewing them is one of the most reliable ways to increase your ad spend without proportionately increasing your results.

The right mental model for AI agents is the same one that applies to AI Max and Performance Max: they are powerful tools that work best when guided by genuine sector expertise, not substitutes for it.

The Human and AI Balance in PPC: What Actually Works

How should SMEs think about balancing AI automation with human expertise in their PPC campaigns?

The businesses generating the best returns from Google Ads in 2026 are not the ones that have handed everything to automation, nor the ones that are manually managing every bid. They’re the ones that have a clear understanding of what the AI does well and what it doesn’t, and have structured their campaigns and their management approach accordingly.

In practice, this means:

• Let the AI handle real-time bidding. Smart Bidding’s ability to process hundreds of signals per auction and adjust bids in milliseconds is genuinely superior to any manual process. Don’t fight this – give it the conversion data and targets it needs to work effectively and let it do its job.
• Keep human control over keyword strategy and negative keywords. The explosion in search query diversity driven by AI-powered matching means that the shift from keyword-based search to AI-driven discovery has automated the tactical layer of campaign management – but your negative keyword strategy and your keyword structure are still the primary levers through which sector expertise influences campaign performance. These require human knowledge of your market that no AI currently has.
• Review AI agent recommendations before implementing them. The one-click implementation that tools like Adzooma offer is convenient – but every recommendation should be assessed against your knowledge of your business and market before being applied. Use AI agents for diagnosis and data analysis; use human judgement for decisions.
• Build your asset library for AI campaigns. Performance Max and AI Max perform better with more high-quality assets to work with. Investing in good photography, short video content, and a range of headline and description variations gives the AI better material to test and optimise – and reduces the likelihood of it generating creative that doesn’t reflect your brand or positioning accurately.

Using ChatGPT, Claude, Gemini and Other AI Tools to Analyse Your PPC Campaigns

Can general-purpose AI tools like ChatGPT or Claude be useful for PPC management?

Many SMEs are already doing this informally – pasting a performance report into ChatGPT and asking what it means, or using Claude to generate ad copy variations. The question worth asking honestly is where these tools genuinely add value, where they fall short, and what risks come with using them for campaign decisions.

The significant advantage non-Google AI tools have over Google’s own agents is impartiality. Google’s recommendations – whether from Ads Advisor or the platform’s built-in suggestions – are generated by a company that earns more money when you spend more. That’s not a conspiracy theory, it’s a structural reality worth keeping in mind. ChatGPT, Claude, Gemini (when accessed outside of Google’s ad platform), and Perplexity have no financial stake in your campaign budget. When you ask them whether you should increase your bids or broaden your match types, they’re not working from an incentive to say yes.

Where non-Google AI tools are genuinely useful for PPC:

Ad copy generation and testing is probably the most immediately practical application. These tools are excellent at generating multiple headline and description variations from a brief, suggesting different angles for the same offer, or rewriting existing copy to better match a specific audience or intent. They can also help you identify whether your current ad copy is benefit-focused or feature-focused, and suggest improvements – useful for businesses that write their own ads without a copywriting background.

Interpreting reports and data you paste in is another strong use case. If you export your search terms report, your campaign performance data, or your auction insights report and share it with a capable AI tool, you can get a plain-language summary of what the data shows, which patterns stand out, and what questions are worth investigating further. For SMEs without a data analyst, this can be genuinely time-saving.

Brainstorming keyword lists, negative keyword patterns, and audience segments is also valuable. These tools can help you think through the range of terms someone might use when searching for your service, including irrelevant terms to add as negatives – particularly useful for niche B2B businesses where generic AI matching can misfire badly.

Explaining platform features and changes in plain language is something these tools handle well. If you’ve read that AI Max is migrating your DSA campaigns in September 2026 and want to understand what that actually means for your specific campaigns, a well-prompted AI tool will give you a clearer explanation than most official documentation.

Where non-Google AI tools fall short for PPC:

The fundamental limitation is that they have no live access to your Google Ads account. Everything they analyse has to be copied and pasted manually, which means their recommendations are based on a snapshot of data rather than a complete, real-time picture. They can’t see your auction data, your quality scores, your historical bidding patterns, or the full context of how your campaigns are performing across all dimensions simultaneously.

They also don’t know your business, your sector, or your competitive landscape unless you tell them – and even then, their knowledge is limited by their training data and by what you’ve chosen to share. An AI tool told that you run a B2B industrial cleaning supplies business will give you more relevant guidance than one working from a generic “PPC campaign” prompt, but it still won’t have the contextual market knowledge that an experienced specialist brings to an account review.

Hallucination risk is real, particularly for specific platform features and recent changes. AI tools can confidently describe Google Ads features that have been deprecated, give incorrect information about how specific bidding strategies work, or suggest settings that don’t exist in the current platform interface. WordStream’s testing of AI tools for PPC accuracy found meaningful variation in accuracy across tools, with Google’s own Gemini-powered tools performing best for platform-specific questions – which makes sense given their access to current platform documentation. For questions about how Google Ads actually works in 2026, cross-reference AI responses against official Google support pages before acting on them.

Data privacy is also worth considering. When you paste campaign data, keyword lists, or performance reports into a third-party AI tool, you’re sharing potentially commercially sensitive information with that tool’s provider. Most major AI tools have business-grade options with stronger data privacy commitments, but it’s worth understanding what you’re agreeing to before sharing detailed account data.

The practical verdict for SMEs:

Non-Google AI tools are most valuable as thinking partners and productivity aids – for generating copy variations, interpreting data you’ve exported, brainstorming keywords, and explaining concepts. They work best when treated as knowledgeable generalists who need context from you, rather than as PPC specialists with live account access. The combination of a capable AI tool for analysis and ideation, Google’s own agents for platform-specific diagnostic recommendations, and human sector expertise for strategic decisions is more powerful than any of the three alone.

What Should an SME Actually Do with All of This?

Where should a small business start with AI in their PPC campaigns?

If you’re feeling overwhelmed by the pace of change in Google Ads, the most useful thing to do is start with a clear-eyed audit of what’s already happening in your account. Many SMEs have AI features active in their campaigns that they haven’t consciously enabled – broad match keywords added by Google recommendations, automatically created assets turned on by default, Performance Max campaigns suggested by the onboarding flow. Understanding what’s currently running is the necessary starting point before making any changes.

From there, the practical priority order for most SMEs:

First, ensure conversion tracking is working correctly. Every AI feature in Google Ads depends on conversion data – without it, automation is working blind. Second, review whether your campaign structure is appropriate for your business type, using the guidance on Performance Max and AI Max above as a framework. Third, explore Google’s AI agents – particularly Ads Advisor and Analytics Advisor – as diagnostic tools, treating their recommendations as inputs to consider rather than instructions to follow. Fourth, if you’re running DSA campaigns, address the September 2026 migration before it addresses itself.

The overarching principle is the same one that applies to AI in marketing more broadly: AI amplifies what’s already there. A well-structured campaign with strong conversion tracking, clear goals, and good creative assets will benefit significantly from AI automation. A poorly structured campaign with weak data and generic creative will be amplified in the wrong direction.