Quick answer: how is AI changing digital marketing?
AI is changing digital marketing mainly inside the ad platforms you already pay for. Meta now runs Advantage+ by default for sales, app and leads campaigns. Google adds AI Max to Search and places ads inside AI Overviews. You get more automated reach, fewer manual controls and some placements you cannot report on separately.
Why most AI marketing example lists don't help a performance budget
Most example lists show enterprise stunts: a custom AI campaign from a global drinks brand, or a recommendation engine built by a data science team. You can't switch those on in your ad account on Monday morning.
The changes that hit a brand spending ₹10 lakh to ₹1 crore a month are quieter. Platforms switched defaults on, removed manual controls and added placements inside AI answers. Several of the popular lists still describe Meta and Google as they worked in 2023.
For each of the 12 examples below we answer four things:
- What changed, and when.
- What you lost: a control, a setting or a report.
- Where it applies: some changes are market-specific.
- What to do: test it, accept it or work around it.
For the wider picture of AI across SEO, content and analytics, see our complete guide to AI in digital marketing for 2026. This article stays with ad accounts.
Examples 1-4: what Meta's AI now decides for you
Meta has moved from "AI is an option" to "AI is the default". Here are the four changes that matter most for a sales-focused account.
Example 1: Advantage+ became the default for sales campaigns
In February 2025 Meta announced a streamlined campaign setup, first as a test. Advantage+ is now on by default when you create a campaign with the sales, app or leads objective. Advantage+ sales campaigns replaced Advantage+ shopping campaigns.
What you lost: the quiet ability to restrict. Narrowing the audience, excluding placements or running several ad sets without a campaign budget turns Advantage+ off. Several industry write-ups say Meta's help centre gives a 9% average improvement in cost per conversion. We could not confirm this directly from Meta. Either way it is a platform figure, so treat it as a claim to test, not a promise. Our Advantage+ setup guide covers the settings in detail.
Example 2: The separate Advantage+ shopping campaign is being retired
If a guide still calls it "Advantage+ Shopping" with one ad set, it is out of date. Version 24.0 of Meta's Marketing API, released 8 October 2025, blocks creation of new legacy shopping campaigns. Version 25.0, released 18 February 2026, extended the block on creating or updating them to all API versions from 19 May 2026, and Meta says remaining legacy campaigns are paused at version 26.0.
What you gained: an Advantage+ sales campaign can now hold several ad sets. Industry reports say each ad set takes up to 50 ads, against one ad set with 150 ads in the old version. We could not confirm these limits from Meta, so check Ads Manager. If you use third-party tools built on the API, ask your vendor which version they run.
Example 3: The existing-customer budget cap is gone
The old setup let you cap how much budget went to existing customers. The new setup has no such cap. The risk is real at scale: the algorithm can find cheap conversions among people who would have bought anyway.
What to do: track new vs returning revenue from your store data, not just Ads Manager. The worked example later in this article shows why this matters.
Example 4: Meta uses AI chat conversations as an ad signal
From 16 December 2025, Meta uses people's interactions with its AI tools to personalise content and ads. It applies only to people who use Meta AI, and they cannot opt out. Sensitive topics such as religion, sexual orientation, politics, health and ethnicity are not used for ads.
It does not apply in the UK, the EU or South Korea. For a UK brand, nothing changes. Elsewhere, Meta has another signal to match your ads with buyers, which makes your creative the main targeting lever you still control.
Examples 5-9: how Google Search changed around AI
Google's changes are about where your ads show and who writes them. Most of them sit inside campaigns you already run.
Example 5: AI Max expands your keywords without you adding them
AI Max for Search is not a new campaign type. It is an optimisation layer you switch on inside an existing Search campaign, and it is documented in Google Ads Help and the Google Ads API. Some 2025 articles still call it a new beta, so check the live Help page before you rely on any status label.
Search term matching lets Google reach queries you never added as keywords. With Manual CPC, AI Max does not get the signals it needs for its advanced optimisations. Google recommends a conversion or conversion-value bidding strategy.
Example 6: AI rewrites your copy and picks your landing page
AI Max also includes text customisation, formerly called Automatically Created Assets, and final URL expansion. Google can now write headline and description variations and send people to a different page on your site.
For D2C stores, final URL expansion can push traffic to a weak collection page or an old blog post. Review landing pages in your reports during the first two weeks. Google says advertisers who turn on AI Max typically see 14% more conversions or conversion value at a similar CPA or ROAS (cost per acquisition, return on ad spend). That comes from Google's internal 2025 data on non-retail advertisers, so a Shopify store should test, not assume.
Example 7: Your ads now appear inside AI Overviews
Ads from existing Search, Shopping and Performance Max campaigns are automatically eligible to show inside AI Overviews when they win the auction and are relevant to both the query and the answer. Ads can show above or below AI Overviews in all 200+ markets where Overviews exist.
Inside the answer itself, ads run in English in a listed set of countries, including India, the US, Canada, Australia, Singapore, Malaysia, Indonesia and the Philippines. The UK is not on the list and no Gulf market is. Sensitive verticals such as alcohol, gambling, finance and healthcare are excluded.
Example 8: Reporting you can no longer see
There is no separate reporting for ads inside AI Overviews. They are counted as Top Ads, you cannot target the placement alone and you cannot opt out.
Track Top Ads click-through rate, the share of question-style search terms and conversion rate by search term over time. A sudden shift in these, with no change on your side, is often the only sign that AI placements are moving your numbers.
Example 9: Conversational ad formats in AI Mode
Google announced Conversational Discovery ads and Highlighted Answers on 20 May 2026 at Google Marketing Live as tests inside AI Mode. Coverage at the time described them as US tests with no public rollout date, and we found no international timeline from Google. Non-US brands should prepare by keeping product feeds complete and landing pages answering real buyer questions, since those are the inputs any AI answer format draws on.
Examples 10-12: how your team's daily work changes
Once the platforms run targeting and bidding, your team's job moves to inputs and checking. These three changes follow from that.
Example 10: Bidding on value, not just conversions
AI Max and Meta's automated campaigns optimise towards what you tell them a conversion is worth. If you send gross order value, the AI chases revenue, including orders that are later returned or refused on delivery.
Brands with COD (cash on delivery) in India and the Gulf feel this most, because RTO (return to origin, when a COD parcel is refused) can wipe out the margin on an order. Work out your break-even ROAS after product cost, shipping, tax and returns, and make sure purchase events reach the platform cleanly through the Meta Pixel and Conversions API (CAPI).
Example 11: Creative becomes the targeting
With audiences broad by default, the ad itself decides who stops scrolling. That means you need more distinct concepts, not just colour swaps. AI creative tools help you produce volume. Treat vendor claims about time saved as marketing until you measure them on your own team.
More creative also means faster wear-out. Our guide to spotting creative fatigue before ROAS drops covers the signals.
Example 12: AI that reads the account before anyone acts
The newest shift is AI that analyses performance instead of only producing ads. ScaleWin's assistant, for example, pulls the last 7-10 days and breaks results down by age, gender, placement, device, region, hour, ad set and ad. It checks learning-phase status and compares landing-page views with link clicks before suggesting a change.
This matters because when the platform decides more, a wrong human edit costs more. Diagnosis first, then action.
What most guides get wrong about AI in digital marketing
The biggest error is presenting old brand stunts as current examples. Many lists still lead with campaigns from 2022-2024 that tell a performance team nothing about its own account.
Four other claims need correcting:
- "You need AI Max to show in AI Overviews." False. Existing Search, Shopping and Performance Max campaigns are automatically eligible. Broad match and AI Max can widen query coverage, but neither is required.
- "Advantage+ Shopping has one ad set." Out of date. It is now Advantage+ sales, with several ad sets allowed.
- "Platform figures prove it works." Meta's 9% and Google's 14% are the platforms' own numbers. Google's comes from non-retail advertisers.
- "AI decides, so you can stop checking." The opposite. Fewer controls and hidden placements mean you need your own profit tracking more than before.
How to test AI features without breaking a working account
Test one change at a time, against a recorded baseline, and judge it on profit. Google offers a one-click experiment for AI Max that runs a 50/50 split in eligible accounts. On Meta, run one Advantage+ sales campaign alongside your current setup.
Our order of checks before switching anything on
- Tracking: if platform-reported purchases differ from store orders by more than 20%, fix tracking first.
- Volume: our rule of thumb is at least 30 purchases a week in the campaign you plan to test. Below that, results are mostly noise.
- Baseline: record 14 days of CPA, break-even ROAS, new vs returning revenue and RTO rate.
- Split: use Google's 50/50 experiment, or give the Meta test 20-30% of the budget.
- Duration: run at least 3 weeks before judging. Don't edit mid-test.
- Decision: keep the change if profit holds or improves and new-customer revenue does not fall. Roll back if profit drops by more than 10%.
Worked example: Advantage+ sales against a manual setup
Step 1: Work out break-even ROAS
Example: a brand spends ₹30 lakh a month on Meta (₹1,00,000 a day). Average order value is ₹2,000 including 18% GST. After GST, net revenue is about ₹1,695. Product cost is ₹500, shipping ₹200 and the returns and RTO allowance ₹195, leaving ₹800 contribution per order.
Break-even ROAS
Order value ÷ contribution before ads = ₹2,000 ÷ ₹800 = 2.5Step 2: Compare profit per ₹1 lakh of spend
After 21 days, Advantage+ (₹30,000 a day) reports a ROAS of 3.2 and the manual campaigns (₹70,000 a day) report 2.9. Per ₹1 lakh spent, Advantage+ brings ₹3,20,000 in revenue (160 orders) and manual brings ₹2,90,000 (145 orders).
Profit per ₹1 lakh spend
Advantage+: 160 × ₹800 − ₹1,00,000 = ₹28,000. Manual: 145 × ₹800 − ₹1,00,000 = ₹16,000Step 3: Check where the revenue came from
Store data shows 45% of Advantage+ revenue came from existing customers, against 25% for manual. New-customer revenue per ₹1 lakh is ₹1,76,000 for Advantage+ and ₹2,17,500 for manual. Without the old existing-customer cap, Advantage+ found cheaper repeat buyers.
Decision: Advantage+ makes more profit, but its new-customer revenue is lower than manual, so it does not yet pass the rule above. Keep it at its test budget alongside the manual prospecting campaigns, and raise the Advantage+ budget in 20% steps only once new-customer revenue holds.
What differs by market: India, the Gulf, the UK, the US and Asia-Pacific
The same campaign settings behave differently by country. Check this table before copying a playbook from a brand in another market.
| Market | Ads inside AI Overviews | AI Mode ad tests | Meta AI chat signal |
|---|---|---|---|
| India | Yes (English) | No | Yes |
| Gulf (UAE, Saudi Arabia) | Not listed | No | Yes |
| UK and EU | No | No | No |
| US | Yes | US tests (from May 2026) | Yes |
| Canada, Australia | Yes | No | Yes |
| Singapore, Malaysia, Indonesia, Philippines | Yes (English) | No | Yes |
Status as of the latest Google and Meta updates in the brief. Ads above and below AI Overviews run in all 200+ markets where Overviews exist.
Tax changes your profit maths too. Meta and Google record whatever purchase value your store sends, so reported revenue may include GST in India or VAT in the UK, EU and Gulf, while US store prices usually exclude sales tax. Strip tax out before you compare ROAS across markets. COD and RTO matter mainly in India, the Gulf and parts of South-East Asia, so build your RTO rate into the value you feed the bidding.
Run AI tests outside Diwali, Ramadan or BFCM (Black Friday and Cyber Monday). Auction costs swing during these periods, which makes a 3-week test hard to read.
Read next in this series
- 7 AI Digital Marketing Mistakes That Quietly Waste Ad Spend in 2026
- AI Digital Marketing in India: Adoption, Real Rupee Costs and the Tools Worth Paying For
- The Future of AI Digital Marketing, 2026-2030: What Has Shipped, What Is Still a Bet, and What You Still Control
- AI Digital Marketing for Lean Teams Spending ₹10-20 Lakh a Month
- Pros and Cons of AI Digital Marketing in 2026: Which Defaults to Keep, Test or Override
Key takeaways
- Advantage+ is on by default for Meta sales, app and leads campaigns. Narrowing the audience or excluding placements switches it off.
- The old Advantage+ shopping campaign is being retired, and its existing-customer budget cap is gone. Watch your new vs returning split yourself.
- Google AI Max is an optional layer inside Search campaigns. It needs conversion-based bidding, and you do not need it to appear in AI Overviews.
- Ads inside AI Overviews have no separate reporting and no opt-out. They run in India, the US and parts of Asia-Pacific, but not the UK, and no Gulf market is listed.
- Test each AI feature against a recorded baseline and judge it on profit after tax, shipping and returns, not on platform ROAS.
Ai digital marketing: frequently asked questions
Is Advantage+ shopping still available on Meta?
Can I turn off Advantage+ in Meta Ads Manager?
Do I need AI Max to show ads in Google AI Overviews?
Can I opt out of my ads appearing in AI Overviews?
Are ads in AI Overviews available in the UK or UAE?
Does AI Max work with Manual CPC?
Is AI Max the same as Performance Max?
Will AI replace my media buyer or agency?
Sources
- Google Ads Help: About AI Max for Search campaigns
- Google Ads Help: AI Max bidding and experiments
- Google Ads Help: Ads in AI Overviews
- Google Ads API: Getting started with AI Max for Search campaigns
- Reuters via TradingView: Meta to use AI chats to personalise content and ads
- PPC Land: Meta deprecates legacy campaign APIs for Advantage+ structure
Let ScaleWin read your ad data every day
ScaleWin checks your Meta and Google Ads daily, judges every ad on profit and tells you exactly what to change.
Start free, no card



