How AI has shaped paid media, and what do you need to do to stay in control?
Artificial intelligence has stopped being a bolt-on to paid media and has become the operating system underneath it. For digital marketers, the question is no longer whether AI is reshaping PPC and programmatic. It is how fast you adapt, and where you choose to keep humans firmly in control.
Below is a grounded look at what has actually changed with AI in paid media, backed by research and real-world data, and what it means for advertisers trying to protect performance in an AI-first ad ecosystem.
From manual PPC to machine-led media buying.
Over the past five years, paid search has shifted from keyword-by-keyword optimisation to model-driven decision making. Smart Bidding, responsive search ads and, more recently, Google’s Performance Max campaigns, are all expressions of the same trend: hand the signals to the machine and let it decide where and when to bid.
Google’s own documentation is explicit. Performance Max uses AI across bidding, budget optimisation, audience selection, creative combinations and attribution, with the promise of maximising conversions against a set CPA or ROAS goal. What was once the job of a PPC specialist watching bid adjustments is now handled in real time by algorithms consuming thousands of signals per auction.This includes what Google refers to as automated bidding strategies in Google Ads, which form the backbone of Google Ads automated bidding and related systems.
This sits on top of a wider programmatic shift. Programmatic now accounts for over 85% of UK digital display ad spend, driven by automated, data-led buying. In practice, that means a growing share of paid media is now traded by machines, with humans guiding strategy, constraints and creative rather than pulling every lever manually.
The result is a structural change in what “doing PPC” actually means. Less time is spent on keyword hygiene and bid tweaks, more time on feeding the machine with the right goals, ad creative AI inputs and data to enable effective AI in PPC optimisation.
Productivity gains – and where they’re real.
There is no shortage of bullish commentary about AI’s impact on marketing productivity. McKinsey estimates that generative AI could add $2.6–$4.4 trillion to global economic output annually, with marketing and sales among the functions poised to capture a significant share of that value. Its 2023 State of AI survey found that around a third of organisations were already using generative AI regularly in at least one business function, with marketing one of the most common.
More recent data suggests adoption is deepening at leadership level. The Wall Street Journal reported a Wharton/GBK Collective survey of corporate leaders reported that 82% of executives use generative AI at least weekly and 46% use it daily; around three-quarters say they are already seeing positive ROI.
For paid media teams, that tends to show up in very practical ways:
- Faster campaign builds: AI-assisted tools accelerate keyword expansion, ad group structuring and ad copy generation, cutting hours from set-up cycles.
- Creative at scale: Responsive search ads, Performance Max mix and match AI-generated or AI-optimised assets to test more combinations than humans could feasibly manage, an example of how AI ads evolve in real time.
- Search term and audience insight: Large language models help cluster queries, summarise performance themes and identify new intents more quickly than manual spreadsheet work. This is a significant enhancement for AI audience targeting and AI technology in PPC more broadly.
- Forecasting and budgeting: Predictive models support scenario planning, especially for seasonal or promotional periods.
The important nuance is that these gains are not automatic. An MIT-linked study reported that around 95% of enterprise generative AI initiatives showed no measurable impact on P&L, largely because they were bolted onto existing workflows rather than thoughtfully integrated into them.
The same applies to PPC: dropping AI into a messy account seldom produces magic. Teams that see genuine uplift are those that redesign their processes around what the technology is good at (like bidding automation) and keep humans focused on judgment, not drudgery.
AI is changing the shape of search, not just the tools.
The impact of AI on paid media is not confined to how campaigns are managed, it is also changing the surface those campaigns run on.
On the search side, Google’s AI Overviews ad placement and similar generative search experiences are starting to compress the traditional results page. Early analyses suggest that AI Overviews can absorb informational queries that might previously have driven multiple clicks, with knock-on effects for both organic and paid visibility. That pressure is only likely to increase as generative answers roll out more widely in the UK and as Microsoft and others push their own AI search formats.
At the same time, audiences are spending more time in online video and walled platforms. Ofcom’s 2024 Media Nations report notes that overall TV and video viewing in the UK is now being driven by online platforms, including YouTube and broadcaster streaming services, while broadcast TV reach continues to decline.
That shift has direct consequences for paid media: formats like Performance Max that can reach across Search, YouTube, Discover and Gmail are designed to follow that fragmented attention, using AI to decide which surfaces will deliver the best outcome.
For advertisers, this means:
- Query volumes and click paths will evolve as more informational needs are met inside AI experiences.
- Attribution becomes harder, because AI systems are making cross-channel decisions inside black boxes.
- Creative must be more adaptable, capable of working across search, display, video and discovery placements without losing brand coherence.
Opacity, competition and control.
As AI-driven products grow in influence, scrutiny is increasing. According to Reuters, in June 2025, the Turkish Competition Authority opened a formal investigation into Google’s Performance Max to assess whether the way it bundles placements and data might restrict competition or disadvantage advertisers.
While this is not a UK case, it reflects a wider regulatory concern: when a single AI system controls allocation across multiple inventory sources, how transparent is that decision-making and is there room for meaningful choice?
From a practitioner’s point of view, the concern is simpler: black-box automation makes it harder to understand why certain queries, placements or audiences are being prioritised, and whether that aligns with brand safety, margin or strategic goals. AI may optimise for headline metrics like CPA, but without careful guardrails, it can:
- Over-index on branded and bottom-funnel traffic.
- Push impressions into low-context placements to meet volume targets.
- Mask creative fatigue or audience saturation behind blended performance numbers – a risk even when using automated bidding strategies Google Ads or AdWords automated bidding settings.
This is where the human role in PPC does not disappear, but instead becomes more strategic. AMs need to design experiments, set constraints, impose brand and data-use policies and know when to deliberately opt out of PPC automated bidding setups that do not meet their standards.
Governance, ethics and the UK context.
UK advertisers are not approaching AI in PPC in a vacuum. The ISBA and IPA have published 12 guiding principles for the use of generative AI in advertising, emphasising transparency, accountability, respect for intellectual property and the need for robust ethical frameworks around how AI is deployed. This is both a reputational and a regulatory issue: mishandling customer data or misrepresenting AI-generated content can erode trust quickly.
On the investment side, the IPA’s Q4 2024 Bellwether Report points to marketers increasing spend in direct and digital channels where performance is measurable, with commentators explicitly linking this to the uptake of programmatic and AI-powered personalisation. In other words, AI is being adopted not as a gimmick but as a response to budget pressure and the need to justify every pound.
Overlay this with Ofcom’s evolving role in regulating online services and safety, and it is clear that AI-driven advertising will sit under a brighter spotlight in the UK than in many markets. Agencies and in-house teams need to be ready to demonstrate not just performance, but responsible practice.
What does an expert, trusted practice look like in the AI era?
One of the quieter but more meaningful changes in the last year has been the way Google has adjusted the balance of power between its Shopping formats. For a long time, Performance Max enjoyed a built-in advantage: if a product existed in both PMax and Standard Shopping, the automated system would usually win the impression. That no longer holds. Google has rebalanced this so that the two compete on ad rank, which has brought Standard Shopping back into play in a way that is unusual for a company normally pushing advertisers further towards automation.
The practical outcome is that well-structured Standard Shopping campaigns are now competing more evenly with PMax, and in some cases outperforming it. The “levelling” effect exposes a simple reality: automation cannot rescue weak product architecture. Poor feeds, unclear margins or unfocused groupings will still lead to scattershot matching, only now you will see it reflected more openly because PMax no longer has priority by default. The advertisers seeing the cleanest returns are those treating manual Shopping as the foundation and using automation only where it genuinely adds reach or efficiency.
The same shift is evident on the Search side with the introduction of AI Max. It is best understood as an evolution of Dynamic Search Ads rather than a new campaign type. AI Max uses far richer signals than its predecessors, dynamically selecting landing pages and generating ad copy based on what it predicts will be most relevant to each user. In theory, this unlocks incremental queries and reduces the need for exhaustive keyword coverage. In practice, it behaves like a highly sophisticated expansion layer: powerful when the site and tracking are clean, wasteful when they are not.
What links these developments is that Google’s AI is becoming more capable, but not necessarily more aligned with advertiser priorities. PMax and AI Max can find demand you would never manually target, yet they can also cannibalise controlled Search activity, over-index on brand or push spend into placements that favour volume over quality. In the age of AI and smart bidding, the edge no longer comes from using automation, because everyone has access to the same tools. The advantage lies in how you constrain and shape the system.
The direction of travel is clear. The machine handles scale, but humans still define the strategy. Advertisers who build strong structural foundations and introduce automation deliberately, rather than by default, will be the ones who benefit.
Let’s talk paid media that works.
Whether you’re navigating Performance Max, testing automated bidding strategies, or rethinking how to approach AI in paid media, we can help you cut through the noise. Speak to our paid media team about building a strategy that keeps you in control, delivers results, and makes AI work on your terms.
