Weekly Signal · September 1–7, 2026
This week, a more capable AI model arrived while Google changed how existing Search campaigns can reach people. The useful response is to test new capabilities with good business context, understand what campaign settings actually changed, and measure lead quality before scaling.
1. GPT-6 Astra raises the bar for AI work
What happened
OpenAI released GPT-6 Astra on September 3, highlighting stronger research, computer use, and complex, multistep work. Its launch material also describes improvements in following templates and using the relevant business context in an output. (source)
Why it matters to marketers
A model that can work through several steps may be useful for a campaign brief, content plan, or analysis rather than only a single piece of copy. But useful output still depends on the background it receives: the audience, offer, approved claims, examples, goals, tools, and limits. Without that information, even a stronger model can produce generic marketing work.
GrowthGary’s take
Treat context as a business asset. Build a small, maintained source pack for one recurring task, such as a blog brief or weekly report. Include approved facts, a few strong examples, clear rules, and a definition of success. This is the thinking behind a guided system such as GrowthGary AI Marketing ChatOS: the work starts with the business, not a blank prompt.
What to do next
Choose one repeated marketing task and list everything a skilled employee would need before beginning it. Give the AI the same context, test the output against a real example, and remove information that distracts it. Keep a person reviewing consequential changes before connecting the workflow to publishing, customer records, or spending.
2. Google starts moving legacy Search features into AI Max
What happened
On September 1, Google began progressively upgrading eligible Search campaigns that used campaign-level broad match or standalone automatically created assets to equivalent AI Max settings. Existing brand lists are preserved. Dynamic Search Ads are on a later migration timetable, now set for February 2027. Google’s documented defaults leave final URL expansion off in both of the September migration paths. (source)
Why it matters to marketers
The immediate issue is understanding the settings and measuring the result, not assuming that every AI Max feature has been switched on. A change in matching or asset handling can affect the searches and messages that produce leads. Conversion volume alone will not show whether those leads are a good fit.
GrowthGary’s take
Export a baseline for search terms, cost, conversion rate, cost per qualified lead, and landing-page performance. Compare the same measures after the change. Review brand controls and the settings in each migrated campaign rather than relying on the product label.
What to do next
Audit active Search campaigns for the affected legacy features and record current settings. Check migration dates and any changed controls in the account. Keep human approval around budget, offers, and brand language. The broader principle applies to every AI-driven search workflow: make the inputs and the resulting actions visible.
3. Traditional keyword ads begin appearing in AI Mode tests
What happened
Google confirmed a small experiment in which text ads from regular Search campaigns using exact and phrase match keywords can appear in AI Mode when a query has explicit, direct intent. The confirmation came on September 4. (source)
Why it matters to marketers
This could give advertisers some reach inside a conversational search experience without rebuilding every campaign around a more automated product. It is still a limited test. The appearance of an ad in AI Mode does not, by itself, establish that the placement brings better leads or sales.
GrowthGary’s take
Keep tightly themed ad groups and clear intent-based landing pages. A person asking a detailed question in AI search may expect a precise answer when they click. That makes page relevance and a clear offer as important as the new placement.
What to do next
Review search terms and conversion quality regularly, separate brand from nonbrand performance, and document sudden changes in impressions or clicks. If the test reaches your campaigns and reporting permits a useful comparison, measure it against a clean baseline before expanding spend.
The simple takeaway
AI marketing is becoming more capable and more connected to live business decisions. Give AI better context, inspect automated platform changes, and measure qualified outcomes instead of celebrating reach alone. Clean inputs, clear limits, and reliable baselines help a small team benefit from these shifts.