Google Search Ads and AI Max SKILL.md

This is the exact file agents read, including the credit header and licence notice. Save it as google-search-ads/SKILL.md.

---
name: google-search-ads
description: "Build and improve Google Search campaigns in the automated era: conversion goals and values for Smart Bidding, simple structure by economics and intent, AI Max for Search and broad match governed by experiments, brand controls, URL rules and negatives, responsive search ad assets, landing page match, weekly search term reviews and Microsoft Ads equivalents. Use whenever someone asks about Google Ads search campaigns, PPC, AI Max, broad match, Dynamic Search Ads migration, responsive search ads, negative keywords, rising CPCs or brand versus non-brand search."
license: CC-BY-4.0
metadata:
  author: Afeez Adebayo
  author_url: https://afeezadebayo.com
  author_linkedin: https://www.linkedin.com/in/afeez-adebayo/
  author_x: https://x.com/za_icon
  source: https://afeezadebayo.com/skills-md/marketing/performance-marketing/google-search-ads/
  category: marketing
  subcategory: performance-marketing
  version: "1.0.0"
  created: "2026-10-10"
  updated: "2026-10-10"
---

# Google Search Ads and AI Max

Google Search ads used to be a keyword game: match types, bids per keyword, and ad copy written line by line. In 2026 the work has moved upstream. Broad match and AI Max for Search (launched May 2025) find queries beyond your keyword list, rewrite text from your assets and choose landing pages; Smart Bidding sets bids per auction. Google has said AI Max is becoming the default for new Search campaigns and that remaining Dynamic Search Ads move to AI Max, with that migration delayed to February 2027. [Reported from Google announcements, 2026] The advertiser's job is now to govern the inputs: conversion goals and values, structure, assets, landing pages, brand controls and negatives, and to watch search terms closely.

This skill builds, audits and improves Google Search campaigns (and Microsoft Ads equivalents) in that automated world.

## Operating principles
1. **Feed bidding the right goal.** Smart Bidding optimises for whatever conversions and values you give it.
2. **Structure by economics and intent,** not by every keyword theme.
3. **Govern automation with controls:** negatives, brand exclusions or inclusions, URL rules and asset reviews.
4. **Read search terms every week** when using broad match or AI Max.
5. **Separate brand from non-brand** to see true acquisition cost.

## Account structure
| Campaign | Purpose | Bidding | Notes |
|---|---|---|---|
| Brand | Protect branded queries | Target impression share or tCPA/tROAS with caps | Watch incrementality; test where competitors don't bid |
| Non-brand core | High-intent product or service queries | tCPA or tROAS on qualified values | Main acquisition engine |
| Non-brand discovery (AI Max or broad) | Find new queries | tCPA/tROAS with search term monitoring | Promote winners to core |
| Competitor | Competitor names, where allowed | Strict tCPA | Follow trademark rules |
| Geography or language splits | Different economics or budgets | As above | Only where economics differ |

## When to use, and when not to
Use it for new Search accounts, audits, AI Max adoption decisions, rising CPCs, falling conversion rates, and migrating from DSA.

Don't use it for Shopping and Performance Max (`performance-max-shopping`) or for overall budget allocation (`media-planning-budgeting`).

## Inputs to ask for, in one message
1. Account access or exports: campaigns, ad groups, keywords, search terms, assets, conversions, 90 days.
2. Conversion actions, values and offline imports.
3. Margins or target CPA and ROAS by product or service.
4. Brand guidelines and terms that must not be used in ads.
5. Landing pages and their conversion rates.

## Workflow

### Step 1. Check goals
Primary conversions, values, offline imports, new-customer goals. Fix tracking first (`conversion-tracking-capi`).

### Step 2. Simplify structure
Consolidate thin ad groups and campaigns so bidding has enough data; keep splits only where economics, intent or budgets differ.

### Step 3. Choose matching and AI Max settings
Test AI Max in a controlled way: an experiment against the current setup, with brand controls, URL inclusions or exclusions, and text customisation reviewed. Keep exact and phrase for proven high-value queries where control matters.

### Step 4. Build negatives and controls
Account-level negative lists (jobs, free, irrelevant meanings), brand exclusions on non-brand campaigns, URL exclusions for pages that shouldn't receive ads.

### Step 5. Write assets
Responsive search ads with distinct headlines covering benefit, proof, offer, objection handling and the query's language; descriptions with clear next steps; sitelinks, callouts, structured snippets, images, prices and promotions where relevant.

### Step 6. Align landing pages
Message match, fast mobile pages, clear call to action; review the landing page reports when AI Max chooses final URLs.

### Step 7. Set bidding and budgets
tCPA or tROAS based on margin; avoid budget limits on campaigns that should scale; change targets gradually.

### Step 8. Monitor
Weekly search terms, asset performance, landing page reports, impression share lost to budget or rank, and new-customer share.

## Search term review routine
| Finding | Action |
|---|---|
| Irrelevant query with spend | Add negative (right level and match type) |
| Relevant, converting query in discovery | Add as exact or phrase in core campaign |
| Brand query in non-brand campaign | Add brand exclusion |
| Competitor query where you don't want it | Add negative or move to competitor campaign |
| Query sent to a poor landing page | Adjust URL rules or create a better page |

## Testing AI Max safely
1. Use a Google Ads experiment: control is the current campaign; treatment has AI Max on.
2. Turn on brand controls (exclusions on non-brand) and URL rules before starting.
3. Review text customisation outputs in the first week; restrict if they break brand or legal rules.
4. Run for at least four weeks or until the experiment reaches significance on your primary goal.
5. Judge on qualified value or profit, not clicks or raw conversions.
6. Check the search terms and landing page reports for new queries and pages AI Max introduced.

## Asset checklist for responsive search ads
- 10 to 15 distinct headlines: the main benefit, proof (numbers, reviews), offer, objection handling, query language, brand.
- 4 descriptions with a clear next step.
- Sitelinks to the most useful pages, callouts, structured snippets, images and price assets where relevant.
- Avoid near-duplicate headlines; they limit combinations.
- Pin only where legally or brand-required.

## Weekly monitoring checklist
| Check | Look for |
|---|---|
| Search terms | Irrelevant spend, new converting queries, brand leakage |
| Landing page report | Pages chosen by AI Max with poor conversion |
| Asset report | Low performers to replace |
| Impression share lost (budget/rank) | Where scaling or quality work is needed |
| Conversion lag | Avoid judging targets before conversions mature |
| New vs returning customers | Whether search is acquiring or harvesting |

## Brand search: protect or pay?
Brand campaigns often show very high ROAS because people searching your name were going to find you anyway. Test incrementality by pausing brand ads in regions or periods where competitors aren't bidding and measuring total brand clicks and conversions (paid plus organic). Keep brand coverage where competitors bid aggressively or where sitelinks and offers clearly help.

## Edge cases and judgement calls
| Situation | What to do |
|---|---|
| Low conversion volume | Consolidate campaigns; consider maximise conversions before targets; optimise on a valid earlier event if needed |
| Strict brand or legal language | Restrict text customisation; pin approved headlines where necessary |
| Lead generation with poor lead quality | Import qualified outcomes and bid on them |
| Rising CPCs | Check auction insights, quality and landing page experience before raising bids |
| Microsoft Ads | Import Google campaigns, then adjust bids and audiences; check search partners |

## Handoffs to related skills
- Tracking and values: `conversion-tracking-capi`
- Shopping and PMax: `performance-max-shopping`
- Landing pages: `landing-page-cro`
- Ad copy testing: `ad-creative-testing`
- Brand search incrementality: `incrementality-testing`

## Worked example
A B2B software company ran 140 Search campaigns with tiny budgets, exact match only, optimising for demo form fills. CPCs were rising and sales rejected most demos. The rebuild consolidated into brand, core non-brand, competitor and a discovery campaign testing AI Max as an experiment, with offline imports of qualified demos as the bidding goal. Weekly search term reviews moved converting queries into core and added negatives for job seekers and free tools. Qualified demos rose and cost per qualified demo fell over two months, while the AI Max experiment was kept after it beat the control on qualified demo value.

## Output format
```
# Google Search plan: <account>, <date>
## Goals and values
## Structure (campaigns, purpose, bidding)
## AI Max and matching approach (with experiment design)
## Negatives and brand controls
## Assets and landing pages
## Monitoring routine
```

## Common mistakes
- Turning on AI Max or broad match without reviewing search terms.
- Hundreds of thin campaigns starving bidding of data.
- Brand and non-brand mixed.
- Bidding on form fills that sales rejects.
- Assets that repeat the same message in every headline.

## Questions to ask before you change the account
- Which conversions and values reflect real business value?
- Where do margins differ enough to justify separate campaigns?
- What brand or legal language rules apply to ad text?

## Self-check before you answer
- Are goals and values aligned with business value?
- Is the structure simple enough for bidding to learn?
- Are AI Max and broad match governed by controls and reviews?
- Is brand separated and tested for incrementality?

## Official documentation
- Google Ads Help (search "AI Max for Search campaigns"): https://support.google.com/google-ads/
- Smart Bidding: https://support.google.com/google-ads/answer/7065882
- Responsive search ads: https://support.google.com/google-ads/answer/7684791
- Negative keywords: https://support.google.com/google-ads/answer/2453972

---
Created by Afeez Adebayo, Marketing, SEO & Growth Expert
Website: https://afeezadebayo.com
LinkedIn: https://www.linkedin.com/in/afeez-adebayo/
X: https://x.com/za_icon
Source: https://afeezadebayo.com/skills-md/marketing/performance-marketing/google-search-ads/

Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). You may copy, share and adapt this skill, including for commercial use, as long as you credit Afeez Adebayo and link to the source URL above. Keep this notice in any copy or adapted version.