From September, no new DSAs – and existing DSA, ACA and campaign‑level broad match setups will be auto‑upgraded to AI Max.
Google is upgrading Dynamic Search Ads to AI Max – and from September, your legacy DSA, automatically created assets, and campaign‑level broad match will start to move over. This isn’t a plot twist so much as the next step in a direction Google’s been heading for years: less energy on building endless keyword lists, more on giving the system better signals to work with.
If you’ve used DSA as a “catch‑all” for gaps, or as a light discovery layer, this change matters. AI Max will do that job differently – and you have a window now to shape how it’s set up, tested and explained before the automatic upgrades kick in.
What’s actually changing?
Whether you’re using Dynamic Search Ads, automatically created assets or the campaign‑level broad match setting today, those setups are now on a path to AI Max from September via Google’s automatic upgrades (or earlier if you choose to do so).
Google’s line is simple: AI Max is an optimisation layer for Search campaigns that uses Google AI to expand reach, tailor creative and pick better landing pages. It blends what DSA used to do (dynamic matching to your site) with broader AI‑driven matching and asset optimisation, all inside your existing Search campaigns rather than as a separate campaign type.
Key points from Google’s announcement and docs:
- DSA campaigns and ad groups will be upgraded to AI Max starting in September, with tools available now to migrate and test early.
- AI Max is not a new campaign type; it’s a suite of features you can switch on within Search (search term matching, text optimisation, final URL expansion, plus controls like brand/location settings and text guidelines).
- Google cites internal data showing incremental conversions and broader query coverage when AI Max is enabled with Smart Bidding, but those numbers should be treated as directional, not gospel.
Around that, the discourse is loud: “end of DSA”, “end of keywords”, “more AI creep”… Underneath the headlines, the practical reality for most advertisers is quieter: your discovery layer is getting smarter and more tightly integrated – and it will reward clean inputs and clear guardrails more than clever match‑type tricks.
What does this actually mean?
This continues a trend, it doesn’t start one
Smart Bidding, Broad Match, Responsive Search, Performance Max: Google Ads has been AI‑shaped for a long time. AI Max is another iteration in that story, not a sudden lurch into an AI‑only future.
The useful way to think about this isn’t “we’re losing control” – it’s “the levers that matter have shifted”. You already see better outcomes when your structure, site and assets are in good shape; AI Max simply leans harder on those ingredients, and less on whether you’ve built every long‑tail keyword yourself.
The old “DSA catch‑all” habit has a use‑by date
Using DSA as a separate box to hoover up whatever the rest of your account missed has always been a bit of a hack. It works, but it hides structural issues (gaps in coverage, messy taxonomy, weak landing pages) behind a single campaign that quietly cleans up after everyone else.
AI Max pushes you towards a more deliberate setup:
- Discovery becomes something you design into your core Search campaigns, not something you bolt on.
- Queries, ads and URLs are all being chosen by the same optimisation layer, so it’s worth thinking about how your campaigns, ad groups and site sections signal intent and value to that layer.
The upside? You get a more responsive, AI‑assisted discovery engine that can follow changing behaviour faster than a manual keyword build ever will.
“Understanding enough” is the goal, not total line‑by‑line visibility
AI Max won’t give you the comforting illusion of control that came from building every keyword by hand. But Google is adding more reporting – including ways to see where AI Max is matching, what assets it’s leaning on, and which URLs it’s sending people to.
The useful question becomes: do you have enough understanding to run this like adults?
- Can you isolate AI Max’s contribution in your reporting (queries, conversions, incremental themes vs existing coverage)?
- Can you tell a simple story internally about what it’s doing, where it’s working, and where you’ve chosen to put guardrails on it?
If you can answer those, you’re in a good place – even if you can’t see every decision the model makes.
What do you do now?
This week: take stock
- Quantify your DSA footprint
- Which campaigns/markets use DSA or similar features, and what % of spend and conversions do they drive?
- Decide your AI Max stance
- Are you going to test and tune before September, or accept the default upgrade and monitor?
- Fix obvious inputs
- Prioritise key categories where DSA leans on messy site structure or weak landing pages.
This quarter: test with intent
- Run like‑for‑like experiments
- Use Google’s tools to test AI Max against current DSA, keeping everything else comparable.
- Add a simple AI Max view to reporting
- Track AI Max‑influenced spend, conversions, CPA/ROAS and top query themes.
- Collect a few real examples
- One great win, one miss you fixed with guardrails, one “no real change” case.
Before September: write your one‑pager
Have a short internal note that answers:
- Where we will / won’t use AI Max by default.
- The minimum standard on site, assets and structure before we switch it on.
- How we’ll judge success and when we’ll intervene.
