What Changed in Digital Marketing This Week: ChatGPT Ads, Google AI Labels, and the DSA Deadline

Key Takeaways

ChatGPT Ads launched audience list targeting on 7 July, letting advertisers upload hashed emails and phone numbers as audience filters, alongside a new overview dashboard and suggested ad drafts. OpenAI’s ad platform is starting to look credible.
Google rolled out mandatory AI ad disclosures on 9 July across Search, YouTube, and Discover. A “How this ad was made” label now appears on any ad created with Google’s own AI tools, automatically. Third-party AI use is still optional to disclose.
Dynamic Search Ads will be auto-migrated to AI Max in February 2027. Campaigns using Automatically Created Assets face an earlier September 2026 deadline. The window to upgrade on your own terms is open now.
Google updated its Local Services Ads requirements on 6 July, primarily an administrative tidy-up, but the link between compliance and badge eligibility has been made more explicit.
New Google Ads terms of service took effect on 1 July, with expanded language around how advertiser inputs, including website URLs and conversational AI interactions, feed into Google’s systems.
The core theme across all of it: AI is taking over the execution layer, and advertisers who have not cleaned up their data and creative inputs are going to feel the shift most.

Three major developments landed this week across Google and OpenAI, each pointing in the same direction. AI is now running more of the campaign, and the human advertiser’s job is shifting to managing the inputs and reviewing the outputs. Here is what actually changed, and what to do about it if you run paid advertising for a New Zealand service business.

ChatGPT Ads Is Starting to Look Like a Real Platform

OpenAI’s advertising product has moved through several phases since it launched: curiosity piece, US-only beta, and now, in the space of two days this week, something that resembles a functional ad platform.

On 7 July, Search Engine Land reported that OpenAI had quietly rolled out custom audience list targeting inside ChatGPT Ads. The feature lives under a new “Tools” section labelled Audiences. Advertisers can upload raw or hashed email addresses and phone numbers to use as audience filters. Lists of 25,000 contacts or more can be used to include or exclude specific segments from campaigns, with bid multipliers available at the ad group level. It is the same core mechanic that Google and Meta have offered for years, and its arrival signals that OpenAI is building a proper targeting stack rather than a novelty product.

Two days later, Search Engine Land covered a batch of additional ChatGPT Ads updates. A new overview dashboard gives advertisers account health monitoring, recommended performance actions, and trend charts for key metrics. Suggested ad drafts can pull an image, headline, and description directly from your existing website metadata, though OpenAI has been specific: this does not generate new copy or imagery using AI, it surfaces what is already there. A redesigned static ad card format with larger visual elements rolled out across web and mobile. The platform also expanded into Japan and South Korea.

ChatGPT Ads remains a US-first product and confirmed access for New Zealand advertisers is not yet open. But there are two reasons it matters here. First, the platform is expanding geographically at speed, moving from the US into major Asia-Pacific markets within months of launch. Second, OpenAI is building toward a real advertising alternative to Google and Meta, and the trajectory is faster than the market expected. A daily active user base of 200 million is not a small audience.

The practical question for NZ businesses is not if ChatGPT Ads arrives here but when. If you have not thought about how your business would advertise inside an AI assistant, this is a reasonable week to start.

What to do: For now, track it rather than act on it. Draw up a simple brief covering your core service, your customer list format, and the website copy you would use for suggested ad drafts. When access opens, you will move faster than competitors who are starting from scratch on that day. Clean, specific website copy will matter more on this platform than it does on Google Search, where you write every headline yourself.

Google Mandates AI Labels on Every Ad Its Tools Help Create

On 9 July, Google announced the global rollout of AI ad disclosures across Search, YouTube, and Discover. When Google’s own generative AI tools are used to create or alter an ad, the system now embeds SynthID markers into the content and surfaces a “How this ad was made” label in My Ad Center. Users find it through the three-dot or info icon on any ad. The label shows whether the creative was AI-generated or AI-modified, and by which tools.

For advertisers using Google’s Asset Studio, Performance Max creative generation, or the conversational ad creation workflow, these labels appear automatically. You do not need to trigger them. Google is doing it for you.

For advertisers using third-party AI tools, including independent image generators, AI copywriting products, or anything built on a model that is not Google’s, disclosure is currently optional depending on local regulation. That last qualifier matters. New Zealand does not have specific AI content disclosure requirements today, but markets like the EU are moving in that direction, and running undisclosed AI-generated content in international campaigns is an increasing compliance risk.

This is not a creative constraint. Google is not restricting what AI can do to your ads. It is saying that when AI is involved, users will be told. That has two practical effects. First, it normalises the expectation of transparency, and that expectation will spread across all platforms. Second, it creates a visible connection between “AI helped make this” and your business name. An AI label on a well-crafted, accurate ad is not a problem. An AI label on a hallucinated product claim or a misleadingly generated image is.

For NZ businesses running Google Ads: if you have been using Google’s AI creative tools and the outputs are not reviewed before going live, build that review step in now. The label makes AI authorship visible to users. You want what follows to be worth owning.

What to do: Pull up your current Google Ads creative workflow and identify any campaigns where Asset Studio or automated creative generation is active. Sample a set of live ads and ask honestly whether you would be comfortable with “Made with AI” appearing next to your brand name on each of them. Fix the ones where the answer is no.

Your Dynamic Search Ads Have a Deadline

If you are running Dynamic Search Ads, the migration calendar has been confirmed. According to updated guidance from Google, DSA campaigns will be automatically upgraded to AI Max or to standard Search campaigns with broad match and Smart Bidding starting in February 2027. That is seven months away.

Campaigns using Automatically Created Assets and campaign-level broad match settings face an earlier cut-off: auto-upgrades begin in September 2026. That is nine weeks away.

Google’s history on this migration is worth noting. The company announced the DSA phase-out in early 2026, paused new DSA campaign creation, then reversed that decision in June and allowed new campaigns again. The confirmed position now is that existing DSA campaigns will keep running, but February 2027 is when Google takes over the migration if you have not done it yourself. Doing it yourself means you control what ports across, you preserve historical data correctly, and you are not surprised by what the AI has decided to do with your campaigns.

The new reporting documentation Google released alongside this update is genuinely useful. Advertisers can now segment performance data by ad format, compare AI Max and DSA results in separate views, and analyse search term and landing page performance for each campaign type independently. If you have avoided reviewing DSA performance because the data was hard to isolate, these new report views solve that.

Google’s optimisation guidance has also shifted in a meaningful way. The recommendation is to focus on conversion goals rather than keyword relevance, review search terms and item groups every one to two weeks, and use negative keywords sparingly rather than over-filtering traffic that the AI can find value in. That last point deserves attention. DSA was built on the principle of giving Google your site and managing the negative list to block irrelevant traffic. AI Max operates the same way but across a much larger surface area. Advertisers who built their DSA strategy around aggressive negative keyword lists may find that those lists undercut AI Max’s ability to work.

We covered AI Max when it moved out of beta at Google Marketing Live in May, including the performance data Google cited for full-feature campaigns. The DSA deadline makes that conversation more urgent than it was then.

For a plumbing company in Auckland or a law firm in Wellington running DSA campaigns against their service pages: the February 2027 migration will happen whether you are ready or not. The difference is whether you port your settings and historical data in an orderly way, or whether Google does it for you and you spend a quarter unwinding the results.

What to do: Run a DSA performance report this week. Identify your best-performing landing pages and search terms. Use the new segmented report views to understand what DSA is actually doing before you migrate. Either upgrade now using Google’s porting tools, or schedule the migration for October to give yourself a clean run before the September Automatically Created Assets deadline.

Two Smaller Changes Worth Knowing About

On 6 July, Google updated its Local Services Ads requirements. The changes are primarily administrative: renaming “Local Services platform policies” to “Local Services Ads requirements,” cleaning up outdated provisions, and aligning the documentation with the current badge system. The underlying verification framework did not change. But the restructuring makes the connection between compliance standards and badge eligibility more explicit, and for any tradesperson, clinic, or legal firm using Local Services Ads in markets where it is available, badge removal triggered by a compliance issue during a busy period is a painful outcome. Review the updated requirements if you are running LSA anywhere.

On 1 July, updated Google Ads terms of service took effect. The headline addition is expanded language explaining how advertiser-provided inputs, including assets, URLs, and information entered into Google’s conversational AI tools, may be used across Google Ads features to improve campaign performance. New provisions also cover what Google can access and crawl when you authorise it for automated campaign setup.

None of this is surprising given where Google is heading, but it has implications for advertisers handling sensitive client information through Google’s ad tools. If your business is in healthcare, financial services, or legal, and you have started using Google’s AI-assisted campaign setup features, the new terms warrant a read before you authorise automated crawls of anything you would not want processed broadly.

What This Week Tells You About Where Things Are Heading

Strip the product names and the pattern is consistent. AI is moving from “optional feature in your account” to “default execution layer across every major ad platform.” Google is auto-migrating your campaign types on a fixed calendar. OpenAI is building audience targeting from scratch and moving faster than the industry expected. Google is labelling AI involvement in creative whether you asked it to or not. None of these things requires you to do anything, and that is exactly the risk.

Passive advertisers, people who set campaigns and check them monthly, are going to find that the campaigns they think they are running are not the ones actually live. AI Max does not behave like a well-managed DSA campaign. An AI-generated creative set reviewed once at launch is not a managed rotation. A suggested ChatGPT ad draft pulled from a thin or outdated website is not a strong ad.

The businesses that will do well in this environment are the ones treating the AI as a capable but unsupervised employee. Brief it properly. Give it real conversion data and real product descriptions. Check what it produces before a label or a bill arrives. The AI marketing agents guide we published earlier this year covers the operational side of building that workflow if you are doing it for the first time.

For a service business in New Zealand, the practical to-do list from this week is short.

Audit your DSA campaigns before September if you are using Automatically Created Assets. Do not wait for February.
Review what Google’s AI creative tools have produced on your behalf and confirm you are comfortable with a disclosure label attached to each one.
Build a simple brief for ChatGPT Ads now, covering your audience list, service description, and website metadata, so you are ready when access opens in this market.

The AI ad era is not something coming in a future update. It has been running for two years. What changed this week is that the platforms stopped pretending the seams were invisible.

Frequently Asked Questions

What are ChatGPT Ads audience lists and can NZ businesses use them? OpenAI added audience targeting to ChatGPT Ads on 7 July 2026, letting advertisers upload hashed emails and phone numbers to include or exclude audiences from campaigns, with bid multipliers at the ad group level. The ChatGPT Ads platform is currently US-focused with recent expansion into Japan and South Korea. Confirmed access for New Zealand advertisers is not yet available, but the geographic rollout is accelerating.

What does Google’s AI ad disclosure label mean for my campaigns? If you create or modify ads using Google’s own AI tools, including Asset Studio and the conversational ad creation workflow, Google now automatically adds a “How this ad was made” label visible to users through the info icon on the ad. You do not need to add the label yourself. You do need to review what those tools have produced, because the label publicly connects AI authorship to your brand name.

When will my Dynamic Search Ads stop working? Google will automatically migrate DSA campaigns to AI Max or to standard Search campaigns starting February 2027. Campaigns using Automatically Created Assets or campaign-level broad match settings face an earlier auto-upgrade from September 2026. You can migrate yourself at any time using Google’s porting tools and keep more control over the transition.

Did the July 1 Google Ads terms update change how my data is used? The updated terms expanded the language explaining how advertiser inputs, including website URLs, uploaded assets, and information entered into conversational AI tools, may be used to improve campaign performance across Google Ads. If your account handles sensitive client data in healthcare, legal, or financial services, review the updated terms before authorising Google’s automated crawl features for campaign setup.

Should NZ service businesses be preparing for ChatGPT advertising now? Yes, as a planning exercise rather than an immediate action. Document your core service offer, get your customer list into a format that an ad platform can ingest, and audit your website copy so it would make a credible suggested ad draft. When access opens here, businesses that have done that prep will have a meaningful head start on the ones who begin thinking about it on launch day.

If any of the changes above affect how your campaigns are set up and you want a second opinion on what to do, we audit Google and Meta accounts for NZ businesses and can usually identify the highest-impact fixes in a single session. Book a free strategy call and bring your account.