Paid Media Industry Updates

September Updates

September was the month AI features stopped being optional. Across Google, OpenAI, Meta and TikTok, platforms moved from offering automation to switching it on by default, often without advertisers needing to click anything.

OpenAI added conversion optimisation, platform targeting and automatic translation to ChatGPT Ads, just weeks after its European launch. Google automatically migrated Search campaigns into AI Max, removed language targeting from Search and pushed advertisers towards stronger first-party data. Meta pushed creators, messaging and AI agents further into its advertising ecosystem, while TikTok extended Smart+ into search.

At the same time, the EU AI Act made visible disclosure a legal requirement for realistic AI-generated ad creative shown in EU markets.

For advertisers, the opportunity is more automation, more reach and new ways to meet intent across search, social and conversational environments.

The risk is that changes are now happening inside accounts by default. Heading into Q4, the practical question is whether campaigns have the guardrails, data, and creative review in place to keep automation pointed in the right direction.

In Short: What Matters This Month

• ChatGPT Ads added impression-billed conversion optimisation, adjustable attribution, per-platform targeting, and opt-in text customisation with automatic translation.
• Google automatically upgraded Search campaigns using automatically created assets or campaign-level broad match to AI Max, and removed campaign-level language targeting on Search.
• Google also pushed first-party data and measurement, with new Data Strength tools and enhanced conversions coming to Google Analytics.
• Meta took its Creator Marketing Hub global, extended the free window for click-to-WhatsApp conversations to seven days, and launched Meta One paid business subscriptions.
• TikTok announced Smart+ Search Ads, Smart+ Creative Selection and wider access to its TikTok for Business MCP.
• EU AI Act transparency rules now require visible labels on photorealistic AI-generated people, products and scenes in ads delivered to EU audiences.

For most advertisers, September’s updates mean regular account audits are no longer optional. When platforms switch features on by default, human oversight, clean data and clear guardrails are what protect performance.

ChatGPT Ads Is Building Performance Infrastructure At Speed

Business Impact

One month after ChatGPT Ads launched across Europe, OpenAI shipped the kind of measurement, targeting and creative controls that took older ad platforms years to add.

For the UK and European advertisers, the platform is moving from an early test into something that can be planned, measured and optimised, more like other performance channels. The challenge is that it is changing quickly, so reporting definitions, campaign types, and best practices are still moving under advertisers’ feet.

OpenAI added conversion optimisation, platform targeting and AI-assisted copy

What happened

After ads began serving across 31 European markets on 24 August and self-serve Ads Manager opened in those markets on 31 August, OpenAI shipped a run of updates in mid-September.

These included a conversion-optimised campaign type that bills on impressions and optimises towards conversions from both ad clicks and ad views, alongside an Event Quality Score for conversion tracking. Attribution windows became adjustable, and advertisers can now target and report separately on Android app, Android web, desktop web, iOS app and iOS web.

OpenAI also introduced an opt-in setting that adapts existing headlines and descriptions to the context of each conversation and translates them into the user’s language. AI-suggested copy and images were added in Ads Manager, along with integrations for Shopify and HubSpot.

OpenAI also began testing new formats in the US, including Sponsored Agents, where an ad opens into a conversation with a business-sponsored agent. On 23 September, ChatGPT Ads expanded into Southeast Asia and Taiwan and is now available in more than 60 countries. Ads continue to show only to users on the Free and Go plans.

Why this matters

Impression-billed conversion optimisation is a different bet from paying per click. It asks advertisers to trust the platform to find the conversations where an ad influences a later conversion, which only works if tracking is reliable.

Platform targeting matters because performance can vary meaningfully by device and app. Advertisers can now act on those differences rather than accepting a blended result.

The automatic translation and text customisation features are particularly relevant for European advertisers. A single set of assets can now be adapted for users across markets and languages, but advertisers need to be comfortable with the platform adjusting their wording.

What advertisers should be paying attention to

Advertisers should check their Event Quality Score and fix tracking gaps before switching to conversion-optimised, impression-billed campaigns.

Results should be reviewed by platform, and view-through conversions should be reported separately from click-based conversions so month-on-month comparisons stay consistent.

Advertisers in regulated or brand-sensitive categories should test text customisation and translation carefully, reviewing outputs before scaling.

We continue to run live ChatGPT Ads campaigns for clients in the UK and Europe. Our experience so far is that clean product data, clear context hints, and strong landing pages still have more impact than any single new feature.

Google Is Making AI Max The Default For Search

Business Impact

September was the month AI Max stopped being optional for a large share of Search accounts.

Campaigns using older automation settings were moved into AI Max automatically, and Search campaigns stopped using the language setting altogether. Neither change required advertisers to click anything, which is exactly why they need checking.

For advertisers, the opportunity is broader reach into queries they may not have covered with keywords. The challenge is that guardrails such as negatives, brand controls and URL rules now do more of the work that keyword lists used to do.

Search campaigns were automatically upgraded to AI Max

What happened

Between 1 and 30 September, Google automatically upgraded Search campaigns using automatically created assets or the campaign-level broad match setting to AI Max. Campaign IDs and settings carried over, but the features switched on will depend on the starting point.

Campaigns that used automatically created assets moved across with search term matching and text customisation enabled. Campaigns using campaign-level broad match moved across with search term matching enabled, meaning they can now reach queries no keyword in the account covers.

Dynamic Search Ads were not part of September’s move. Following advertiser feedback, Google pushed the DSA migration to February 2027, with no new DSA ad groups from January 2027.

Google also removed campaign-level language targeting on the Search Network. Search ads now match based on the language of the ad itself and the user’s query and settings. In Performance Max, the language setting no longer applies to Search inventory but still filters YouTube, Display, Discover and Gmail.

On 23 September, Google announced a new AI Max report linking the search term, the assets shown and the landing page for each ad, alongside wider language support for AI Brief. The report has no confirmed release date yet.

Why this matters

For many advertisers, AI Max is now running whether or not it was planned. A broad match campaign may look unchanged on screen, but the queries it can enter have widened.

The language change matters for any account that splits campaigns by language, which is common for advertisers running across European markets. A campaign split no longer guarantees which language audience it reaches; the ad copy now does that job.

Together, these changes move Search further from keyword-led control and towards input-led control. Brand controls, URL rules, exclusions and copy guidelines are now the main levers.

What advertisers should be paying attention to

Advertisers should check every Search campaign for AI Max and decide feature by feature what should stay on. Text customisation can be turned off at campaign level, and search term matching can be managed at ad group level.

Search term reports for September should be reviewed for spend on irrelevant queries, with negatives, brand exclusions and URL exclusions updated where needed.

Multi-language and multi-market accounts should confirm that each campaign’s ad copy is written in the language of the audience it is meant to reach. Advertisers still relying on DSA should use the time before February 2027 to plan their own migration, rather than waiting for the automatic one.

Google Is Putting First-Party Data At The Centre Of Measurement

Business Impact

Google’s September measurement updates show that conversion data is becoming the main input behind automated bidding and AI-led Search.

The more automated campaigns become, the more performance depends on what the platform can actually see. Missing or low-quality conversion data means automation optimises towards an incomplete picture.

For advertisers in the UK and EU, where consent choices already reduce observable conversions, the quality of first-party data and tagging is now a direct performance lever.

Google introduced new Data Strength and measurement tools

What happened

On 10 September, Google presented new data and measurement tools, including a Data Strength Uplift metric that estimates the conversions recovered through an advertiser’s first-party data setup. Enhanced conversions are also coming to Google Analytics, with Data Manager integrated directly.

Google reported 11% more Search conversions with enhanced conversions compared with standard conversion imports, and 14% more conversions with Google tag gateway. These are Google’s own figures and should be treated as a direction rather than a benchmark.

Why this matters

A conversion the tag does not see is a conversion Smart Bidding cannot learn from. As AI Max widens matching, the quality of conversion data becomes the main thing stopping automation from optimising towards the wrong outcome.

What advertisers should be paying attention to

Advertisers should audit their tagging, consent mode setup and enhanced conversions before Q4 trading peaks.

Accounts using target CPA or target ROAS should also check targets against actual performance, as Google’s bidding now delivers closer to the configured target. Advertisers in government documents and related services should note that Google’s stricter authorisation policy, covered in our August update, took effect on 5 October.

Meta Is Turning Creators, Messaging And AI Agents Into Ad Infrastructure

Business Impact

Meta’s September updates were less about Ads Manager settings and more about the routes around the ad: creator content, messaging conversations and AI agents.

For advertisers, the opportunity is cheaper, faster ways to source creative and continue conversations after the click. The challenge is that more of the journey now happens inside Meta’s own environment, which makes response handling, product data and creator rights management part of paid social performance.

Creator Marketing Hub went global, and click-to-WhatsApp got a longer free window

What happened

At IAB Global Creator Week on 15 September, Meta announced a bundle of creator-marketing updates. The Creator Marketing Hub, which combines the Creator Marketplace and Partnership Ads Hub, is rolling out globally through the end of 2026 with better creator discovery, content permissions with expiry dates, and one-click ad creation.

Meta also launched Meta One, a paid business subscription with four tiers. It gates insights tools such as Competitive Insights and Custom Audience Insights, and more AI assistant usage.

On 28 September, Meta extended the free messaging window for conversations started from click-to-WhatsApp ads to up to seven days, up from 72 hours. From 1 October, WhatsApp service messages are billed beyond 1,000 free per business number each month.

Meta also confirmed that Threads ads no longer require an Instagram account, with a gradual rollout.

Why this matters

Creator content is becoming a standard input for paid social rather than a separate influencer workstream. Faster sourcing and cleaner rights handling make it easier to feed Advantage+ with the creative variety it needs.

The WhatsApp change shifts the economics of conversational journeys. A conversation started from an ad is now the cheapest place to talk to a customer, while organic service conversations have started to cost money. For lead generation, hospitality and high-consideration purchases, that strengthens the case for click-to-WhatsApp as an entry point.

Meta One is also a signal worth noting. Meta is starting to charge for the insight and AI advisory layer that sits on top of the auction, not only for media.

What advertisers should be paying attention to

Advertisers running click-to-WhatsApp campaigns should extend follow-up sequences to use the full seven-day window and check service message volumes against the new free allowance.

Brands using creator content should test the Creator Marketing Hub and review usage rights, expiry dates and disclosure on Partnership Ads.

Advertisers relying on Competitive Insights or Custom Audience Insights should check which Meta One tier now covers them before access changes. B2B brands without a strong Instagram presence may also want to test Threads once the change reaches their account.

Meta’s US launch of Muse, a personal AI agent that can shop on a user’s behalf, is not yet live in the UK or EU. It is still a reminder that product feeds increasingly act as shopper-facing copy, not just a technical catalogue.

TikTok Is Extending Smart+ Into Search And Agent-Led Workflows

Business Impact

TikTok is applying the same automation logic that runs Smart+ campaigns to search, creative selection and campaign management.

For advertisers, this makes TikTok a more complete performance channel, particularly for discovery-led categories where users search the platform before buying. The challenge is that keyword, creative and bidding decisions are increasingly made by the system, so the quality of the assets and conversion signals matters more than manual settings.

TikTok launched Smart+ Search Ads, Creative Selection and wider MCP access

What happened

At Advertising Week New York on 7 October, TikTok announced a set of AI-led advertising tools. Although just outside September, these updates build directly on the Smart+ changes covered in our August update.

A new Search Ads Campaign, powered by Smart+, automates keyword management, creative, bidding and budgets. Smart+ Creative Selection brings advertiser assets, Symphony AI-generated creative and creator content into a single workflow.

TikTok also made its TikTok for Business MCP available as a connector across several AI platforms, allowing agents and tools to plan, launch and manage TikTok campaigns. GMV Max for TikTok Shop now accounts for more commercial costs, including affiliate commissions, coupons and referral fees.

TikTok also cited 200 million people in Europe using the platform for discovery.

Why this matters

Search on TikTok is becoming a structured, automated buying route rather than an add-on. For brands whose audiences use TikTok to research products, places or services, this creates a new intent-led entry point alongside Google and conversational platforms.

The Creative Selection update also shows where TikTok sees its advantage. Advertiser assets, AI-generated variations and creator content are now treated as one creative pool, which makes creative supply a core performance input.

What advertisers should be paying attention to

Advertisers should review whether TikTok search behaviour is relevant to their category and plan a controlled Search Ads test once it is available in their account.

Creative pipelines should be set up to supply a steady mix of brand, creator and AI-assisted assets, with clear review steps for tone, accuracy and claims. Advertisers using TikTok Shop should check how the new GMV Max cost inputs change their return calculations.

AI Creative Now Comes With Disclosure Rules In The EU

Business Impact

As every major platform adds generative AI to its creative tools, the EU has made disclosure a legal requirement for some types of AI-generated ad content.

This matters for UK advertisers too. The rules apply based on where an ad is shown, so any UK brand running campaigns into EU markets is in scope.

The opportunity is that AI-assisted creative can still be used at scale. The challenge is knowing which assets need a label, and building that check into production rather than relying on platforms to catch it.

EU AI Act transparency rules are now in force for ads

What happened

The transparency obligations in Article 50 of the EU AI Act became binding on 2 August 2026, with Commission guidelines published in July. Penalties can reach €15 million or 3% of worldwide turnover.

For advertisers, the key rule covers photorealistic AI content that resembles real people, objects, places or events. This includes realistic AI avatars, AI-generated customers and photorealistic AI product shots. The artistic exception does not apply where commercial purposes predominate, so these ads need a visible label that a viewer can see without tools. Embedded metadata alone is not enough.

Clearly stylised illustrations, obviously impossible scenes and normal retouching such as colour correction fall outside the rule. AI-assisted text generally needs no label where a named person has genuinely reviewed it and holds editorial responsibility.

Platforms are building the plumbing. Google has introduced tools to help advertisers identify and label AI-generated creative, and Meta has added an AI-generated flag to its ad upload process.

Why this matters

Most platforms now generate or modify creative automatically, from Google’s text customisation and video resizing to Meta’s Advantage+ creative and TikTok’s Symphony. Advertisers remain responsible for what is shown, even where a platform produced the variation.

Pressure on platforms is also rising elsewhere in Europe. A Frankfurt court ruled in September that Meta can be held liable for third-party scam ads, a decision that is not yet final, and Meta now requires identity verification for all financial services advertisers targeting Poland. Both point towards stricter advertiser verification across the EU.

What advertisers should be paying attention to

Advertisers should sort their creative into AI-generated, AI-assisted, and human-made, and record which is which.

Any photorealistic AI people, products, or scenes in ads delivered to EU audiences should carry a visible label from the first frame, including in cropped story and vertical formats. Provenance metadata should not be stripped out in production workflows.

Platform-generated variations, including automated text and image enhancements, should be reviewed for EU campaigns. Business verification should also be kept current on every platform, particularly in financial services and other regulated categories.

What to Watch Next

Q4 is when these changes will be tested under real pressure, as auction prices rise and campaigns have less room for learning periods.

ChatGPT Ads remains the fastest-moving platform in paid media. Conversion-optimised campaigns, platform targeting and automatic translation make it more practical for European advertisers, while US tests such as Sponsored Agents show where the format is heading.

On Google, the priority is understanding how AI Max is behaving in accounts that were migrated in September. The promised AI Max report linking search terms, assets, and landing pages will be important when it arrives, and Dynamic Search Ads advertisers should plan ahead of the February 2027 migration.

On Meta, watch how click-to-WhatsApp economics and the global Creator Marketing Hub change the role of messaging and creator content in paid social plans.

On TikTok, watch for the rollout of Smart+ Search Ads in the UK and European accounts.

For EU campaigns, AI disclosure should now be treated as part of creative sign-off, not a platform setting.

What September Tells Us Overall

September showed platforms moving from offering AI features to switching them on by default.

OpenAI turned ChatGPT Ads into a measurable performance channel within weeks of its European launch. Google migrated Search campaigns into AI Max, removed language targeting, and pushed harder on first-party data. Meta pushed creators, messaging and AI agents deeper into its ad ecosystem. TikTok extended Smart+ into search. And the EU made disclosure of realistic AI creative a legal requirement.

The opportunity is more reach, more automation and more ways to meet intent across search, social and conversational environments.

The risk is that changes now happen inside accounts without anyone clicking anything. Without regular audits, budgets can drift towards irrelevant queries, unreviewed creative and unclear reporting.

That makes paid media management less about operating settings and more about governing systems. Data quality, guardrails, creative review, compliance and clear measurement are what keep automation working in an advertiser’s favour.

Need Help?

If your Search campaigns were migrated to AI Max in September, or you are planning activity on ChatGPT Ads, Meta, TikTok, or across EU markets, now is the time to review how your accounts are set up before Q4 peaks.

The question is no longer only whether your campaigns are live and spending efficiently.

It is whether your tracking, guardrails, feeds, creative, and compliance processes are strong enough to keep automation working towards your goals.