Google Ads in 2026 requires reliable foundations, not a blind switch to every AI format
What matters is what is genuinely available, controllable and measurable in a German account
Google Ads is not changing in 2026 because of one new campaign type. Google Search itself is becoming more conversational through AI Overviews and AI Mode, existing ads may appear in certain AI-powered search environments, AI Max expands Search campaigns, and new advertising formats are being tested. None of this creates an obligation for an SME to activate every innovation immediately.
The sensible response is a controlled readiness process. It separates official product status from announcements, verifies conversion goals and data quality, assigns every campaign a clear role, reviews landing pages and claims, and defines which reports can support a decision before an experiment begins. That is the purpose of the Google Ads 2026 readiness register introduced in this guide.
The register promises neither access to a test format nor additional conversions. Its value is more practical: it prevents a team from mistaking a blog announcement for an account feature, launching automated expansion without suitable pages, or constructing a supposedly precise success story from AI placements that cannot be segmented. A change is approved for launch only after its status, purpose, inputs, controls, measurement and accountable owner have been documented.
Distinguish between available, automatically included, limited rollout, test and announcement-only states.
Assess goals, campaign signals, the website, feeds, creative assets and first-party data for suitability.
Set brand, URL, search-query, budget and approval boundaries before activation.
Observe technical operation, ad delivery and business outcomes as separate layers.
Separate paid AI Search from organic visibility and the forthcoming detailed AI Max guide
This page explains the 2026 landscape and preparation, not every individual account setting
This guide addresses the search intent behind “Google Ads 2026”: what has changed for advertisers, which product changes and claims apply in Germany, and which preparatory work is useful now? It is not a complete operating manual for AI Max. Search-term matching, text customisation, final URL expansion and their detailed controls will receive a dedicated technical review in the next article, so the two pages do not compete for the same primary query.
Unpaid visibility also remains a separate subject. The guide AI Search for SMEs: SEO, AI Overviews, AI Mode and visibility explains how sources, indexing, publisher controls, AI Overviews and AI Mode relate from an SEO perspective. A paid campaign does not purchase an organic citation; conversely, an organic source is not a campaign setting.
The scope here covers status verification, account readiness, campaign-type roles, landing-page and creative foundations, conversion data, experiments, reporting and stop rules. It excludes detailed Shopping feed optimisation, complete account migration, legal advice, individual privacy approval and performance forecasts. This boundary keeps the page maintainable when Google changes rollouts or interfaces.
2026 status, decision logic, the readiness register, release gates and defensible measurement limits.
Organic AI visibility, advanced AI Max configuration, feed projects, legal matters and account-specific optimisation.
First separate the search feature, ad inventory and rollout status
A visible AI Mode does not automatically mean every new ad format can be booked
Google introduced AI Mode in Germany, Austria and Switzerland, supporting complex, multi-part questions and follow-up queries. That changes the research journeys available to users. It does not prove, however, that conversational ad formats are open to every German Google Ads account or can be selected separately.
Likewise, AI Overviews were introduced in Germany in German and English for signed-in users aged 18 and over and appear when Google considers them helpful for the query. Advertising plans must nevertheless distinguish among ads above or below an Overview, ads inside the Overview, and new tests within AI Mode. These positions have different availability and do not share a single button called “buy AI Search”.
The register therefore gives every innovation a field titled “official status on review date”. Acceptable evidence is current documentation, a visible account setting or written confirmation of participation—not a screenshot from another market. Every update records the date, source and affected region so outdated assumptions do not continue unnoticed.
| Change | Status in Germany | Campaign relationship | Control | Measurement limit | Next step |
|---|---|---|---|---|---|
| AI Overviews | Search feature available | Existing ads may be eligible | No dedicated campaign type | Placement not always separately reported | Check current documentation |
| AI Mode | Search feature available | New formats in testing | No universal booking commitment | No general-purpose report | Document account eligibility |
| AI Max for Search | Optimisation layer available | Existing Search campaigns | Review features separately | Use AI Max reporting | Run an experiment, not a full rollout |
| Conversational formats | Limited testing | Dependent on rollout | Not yet a standard workflow | Do not generalise test data | Exclude from forecasts |
| Legacy transitions | Timetable published | DSA, ACA and campaign-level broad match | Follow account notices | Record before and after states | Maintain inventory and deadlines |
Build the Google Ads 2026 readiness register as a shared basis for decisions
Manage an innovation as a verifiable work item rather than a vague idea about the future
Each row in the register describes exactly one change or hypothesis: ads around an AI Overview, for example, an AI Max experiment within an existing Search campaign, or the import of qualified leads. The row contains the source, market status, affected campaign, business goal, required inputs, controls, measurement, owner, approval status and next review date. Product monitoring and operational implementation therefore remain distinct.
An entry begins with “Monitor” when the feature is unavailable to the business or not yet sufficiently defined. “Prepare” means data, pages and responsibilities can be improved without activating the format. “Test” requires an approved hypothesis, stable measurement and limited scope. “Scale” becomes an option only after a defensible evaluation; “Stop” remains a legitimate decision.
The register is not a one-off audit. Product announcements, help pages, default values and account interfaces can change. Every row therefore has an evidence document and an expiry date. A recommendation that appears current but has no review date must not be transferred directly into a campaign. For automatic migrations in particular, the team records the previous setting and the state that is genuinely visible in the account afterwards.
| Topic | Status and evidence | Business task | Required inputs | Control and measurement | Owner and decision |
|---|---|---|---|---|---|
| AI Max in a brand campaign | Available in account; review date | Test incremental relevant demand | Goal, assets, approved URLs | Experiment, search terms, landing pages | Paid Search; prepare |
| Ads around AI Overviews | Environment may be included automatically | Support complex research | Relevant ad and page | Not separately targetable; overall performance | Paid Search; monitor |
| Qualified lead | CRM connection reviewed | Bring bidding closer to lead quality | Consent, identifier, status logic | Import diagnostics and reconciliation | CRM and Marketing; test |
| New conversational ad | No account access confirmed | No operational task | Document readiness only | No budget or forecast | Product monitoring; wait |
| Legacy setting | Migration notice available | Transition without functional break | Baseline and configuration | Before-and-after reports | Account owner; schedule |
Assess ads around AI Overviews realistically within existing campaign logic
Germany supports certain delivery environments, but not separate AI placement management
Current Google Ads documentation on ads and AI Overviews distinguishes ads above or below an Overview from ads placed directly inside it. Existing campaigns can be considered for the adjacent areas in supported markets. Ads inside the Overview, by contrast, are available only in an explicitly listed set of countries and languages that did not include Germany on the review date.
In practice, this means a German SME does not set up a separate “AI Overview campaign”. Google identifies existing Search, Shopping and Performance Max inventory as potential sources. Delivery follows established auction and relevance systems; for ads inside an Overview, the query and the content of the Overview provide additional context. Eligibility, auction entry, an actual impression and a click remain separate events.
Google currently offers neither direct targeting exclusively for this environment nor a separate opt-out for ads within Overviews; according to its documentation, segmented reporting is also unavailable. A performance change must therefore not be attributed to the AI placement without evidence. The register records what can actually be observed and preserves the boundary of the analysis.
A campaign and ad may be eligible for consideration in a particular environment.
Existing ranking and relevance signals continue to determine delivery.
Above, below and inside an AI Overview do not have the same availability.
Without segmentation, a causal statement about the placement is not defensible.
Treat Conversational Discovery, Highlighted Answers and Business Agent as tests
A product announcement is a monitoring signal, not a commitment for a German media plan
In its announcement of a new generation of ads for the AI era of Search, Google describes Conversational Discovery Ads, Highlighted Answers, AI-powered Shopping Ads and Business Agent for Leads, among other products. The formats are intended to answer questions in context or explain products and providers across longer research journeys; paid content continues to be labelled as sponsored.
The decisive words, however, are “test”, “pilot” and “in the coming months”. A format can be publicly announced and technically real while remaining unavailable to a particular account, language, sector or region. The readiness register therefore does not move a format into launch planning until access, permitted use, billing, inputs, controls and reporting have been confirmed inside the relevant account.
There is still plenty that can be prepared. Service pages need to answer concrete questions, product data must remain current, claims require sources and approval, CRM processes must identify qualified leads, and the company needs clear boundaries for automated copy or dialogue. This work also improves existing campaigns; it should not be justified exclusively by a future ad format.
Question-led creative explanation; confirm access and markets before planning.
A sponsored recommendation within eligible answers; do not present it as an organic mention.
Product data and creative become more important; feed quality remains a prerequisite.
A conversational lead journey; clarify sources, handover, privacy and quality in advance.
Plan AI Max and legacy transitions against the updated timetable
DSA, automatically created assets and campaign-level broad match do not share one deadline
Google's updated announcement on the transition from Dynamic Search Ads to AI Max moves the start of the automatic DSA transition to February 2027. The automatic transition of automatically created assets and the campaign-level broad match setting is still due to begin in September 2026. Older summaries that mention only a September deadline are therefore incomplete.
Before a migration, create an account inventory covering the affected campaign, legacy feature, current status, URL rules, brand boundaries, search-query baseline, conversion goals, budget status and accountable person. A screenshot by itself is not enough. Exported settings and reporting data document the baseline and help assess temporal associations. Causally separating migration effects from seasonality and shifts in demand still requires a suitable comparison or experiment design.
An automatic transition must not be interpreted as a recommendation to open every AI Max function as widely as possible immediately. Google may mirror settings or move them into new areas; the business must then verify what was actually activated. The register tracks deadlines, defaults and suitability separately, preventing both panic migration and unnoticed account changes.
Do not migrate on the basis of an old date. First review the current Google documentation, the notice in your own account and the specific legacy feature affected.
Understand AI Max as an optimisation layer for existing Search campaigns
Search-term matching, asset optimisation and landing-page selection solve different problems
Google describes how AI Max works for Search campaigns as a continuous optimisation layer rather than a new campaign type. Search-term matching can supplement existing keywords with broad-match, asset-based and landing-page-based methods. Asset optimisation encompasses text customisation and final URL expansion.
These functions must not be reduced to a single “AI Max on or off” decision. Search-term matching changes the demand that may be reached, text customisation changes the messages that may appear, and final URL expansion changes potential destinations. Each layer requires its own inputs, exclusions, quality criteria and reports. Where a URL is selected dynamically, pinned RSA assets may be disregarded according to the documentation because they may not fit the new destination.
Google also notes that AI Max may not be effective for campaigns that are significantly limited by budget. An experiment therefore begins not with additional automation options but with a review of budget status, conversion foundations, campaign objective and suitable landing-page coverage. The subsequent detailed article series can examine each control in depth; here, the essential decision is whether an experiment should take place at all.
Which additional queries may be inferred from keywords, assets and pages.
Which copy may be customised or generated, and which claims must remain protected.
Which URLs may be served and which destinations must be excluded.
Define brand, location, copy and URL boundaries before expanding reach
Document control as a concrete rule with an example and an accountable owner
The official guide to setting up AI Max for Search campaigns shows settings at campaign and ad-group level, as well as separate areas for search-term matching and asset optimisation. The interface may change during rollout, so the register stores the intended business rule in addition to the location of a control.
AI Max replaces neither the role of other campaign types nor a clean account structure. The existing guide to Performance Max for SMEs explains the cross-channel model. A Search campaign remains focused on search intent and Search inventory. The comparison should be based on task, data and control requirements—not on a promise that one solution is inherently more modern.
Store at least one permitted and one prohibited example for every function. For URLs, this includes not only obvious error pages but also careers, login, legal notice pages, unavailable offers, pages in the wrong language and irrelevant blog articles. For brands and copy, review spelling, claims, prices, mandatory information and sensitive categories. An abstract rule without a test case is not ready for approval.
- Document approved and excluded URL families, each with one live example page.
- Obtain business approval for brand inclusions, brand exclusions and permitted references to competitors.
- Distinguish the service location from location interest and the user's physical location.
- Record copy boundaries for pricing, guarantees, superlatives, mandatory disclosures and brand voice.
- Test tracking templates, redirects and dynamic landing pages together.
Organise Search, Shopping, Performance Max and new formats by the job they perform
More AI in Search is no reason to move proven campaigns into a single system
Search campaigns address explicit search demand and permit detailed analysis of search terms, ads and landing pages. Shopping uses structured product data. Performance Max optimises across channels towards defined conversion goals. New AI-powered advertising environments may draw on existing inventory, but that does not give all campaigns the same role.
Allocation starts with the business problem. Should the campaign answer clearly expressed demand, offer a specific product from a feed, reach additional demand across several Google properties, or qualify a long lead journey? Only then are campaign type, bidding, assets and reach options selected. Technical availability is not a sufficient business case.
The readiness register stores one primary task and one explicit non-goal for each campaign. A brand campaign, for example, may cover brand queries but should not become responsible for general category expansion. A Performance Max campaign may scale across channels, but it does not automatically replace search-term analysis. These boundaries make later experiments easier to interpret and prevent double counting in analysis.
Connect explicit intent, search terms, ad messaging and destination pages in a controlled way.
Serve products from current feeds with price, availability and attributes.
Optimise across several Google properties towards a clearly defined conversion goal.
Plan only when access, an appropriate task and a dedicated approval process are confirmed.
Clean up conversion goals before automation receives additional reach
Bidding must not learn from actions that are easy to measure but have little business value
Google distinguishes between primary and secondary conversion actions. Primary actions can be used for bid optimisation and the “Conversions” column when their associated goal is used by the campaign. Secondary actions normally support observation in “All conversions”; custom goals have special rules that must be reviewed before use.
The guide to GA4 and Consent Mode conversion tracking for Google Ads examines technical implementation and the separation of marketing and consent signals in greater depth. For 2026 readiness, the team also verifies which action represents actual business value: an initial enquiry, qualified lead, appointment, sale, margin or another confirmed state.
Micro-actions such as a page view, scroll depth or phone-link click can provide diagnostic value, but they must not receive the same bidding weight as a confirmed lead without review. Duplicate imports, changing count methods and campaign-specific goals also need to be cleaned up. Additional reach becomes a sensible test only when goal definition, source, value, deduplication and ownership are clear.
Use only actions that represent a clear business outcome or a reliable preceding stage.
Analyse micro-actions and diagnostic signals separately instead of artificially upgrading them.
Regularly recheck CRM status, values, deduplication and campaign assignment.
Treat Enhanced Conversions and CRM feedback as a dedicated data project
More data sources help only when purpose, consent, matching and diagnostics are sound
Google has updated Enhanced Conversions settings for 2026: Google Ads can accept user-provided data simultaneously through website tags, Data Manager and API connections, while the previously separate web and lead settings are being unified. For certain uploads, the move to the Data Manager API has become relevant. Existing implementations should therefore be assessed against current diagnostics rather than an old setup screenshot.
For B2B accounts, Google describes using Google Ads Data Manager with Enhanced Conversions for leads. Qualified or converted lead statuses can be returned from eligible CRM sources. This requires suitable identifiers, field mapping, timestamps, status logic and compliant data handling; a connected CRM does not prove that the import is correct.
Hashing does not automatically create GDPR compliance. The company must have the specific data flow, applicable Google policies, consent requirements and its own legal basis reviewed. The register documents technical and organisational evidence only; it is not legal advice. No upload should begin as an experiment without a confirmed purpose or where sensitive data creates unresolved risk.
- Define the data source, accountable role, purpose and affected conversion action.
- Test identifiers, normalisation, timestamps, a unique transaction or order ID, and deduplication.
- Approve consent signals, Customer Data Terms and sensitive categories separately.
- Align qualified-lead and converted-lead definitions with Sales and CRM owners.
- Monitor diagnostics, match quality, import errors and a business-level sample together.
Test one change with a hypothesis, control state and termination rule
The experiment answers a limited question, not the future of every Google Ads campaign
Google provides a dedicated workflow for AI Max experiments. Depending on the starting position, a control/trial within a campaign may be available. The documentation also lists exclusions and incompatibilities involving certain network, budget, bidding or existing experiment configurations. Verify eligibility in the account before designing the test.
A useful hypothesis describes the change, audience, primary metric, quality safeguard and expected decision direction without assuming a guaranteed effect. For example: “If search-term matching is activated in this adequately funded non-brand campaign, it will generate additional qualified leads without allowing the share of commercially irrelevant queries or landing pages to exceed the agreed threshold.”
During the test, avoid competing foundational changes to goals, budget logic, landing pages and offer definition unless they are part of the design. Record seasonality, promotions, tracking incidents and Sales changes in an event log. Duration should reflect the data available and experiment diagnostics; a universal number would be misleading without account context.
| Hypothesis | Control and trial | Primary signal | Quality boundary | Confounding factors | Decision |
|---|---|---|---|---|---|
| AI Max matching expands valuable demand | Existing state versus trial | Qualified conversions or value | Irrelevant queries and URLs | Seasonality, budget, tracking | Adopt, adjust or stop |
| Text customisation increases relevance | Approved assets versus customisation | Goal-related performance | Claim and brand violations | New offers, creative changes | Retain or remove assets |
| CRM feedback improves the signal | Before and after a stable import | Confirmed lead status | Match and import errors | Sales process, time lag | Approve or repair signal |
| URL expansion finds better pages | Fixed URL versus approved URL family | Conversion and page relevance | Incorrect or unapproved destinations | Relaunch, redirects | Constrain expansion or stop |
Evaluate the query, asset, actual landing page and conversion separately
AI Max transparency is useful, but it is not a separate AI Mode performance report
Google's documentation on reporting in AI Max for Search campaigns covers search-term, keyword, asset and landing-page reports, among others. New labels such as the AI Max match type or a source column help explain why a match occurred. Reports can also reveal the actual headlines and destination pages that were served.
This transparency does not automatically answer whether an impression occurred in AI Mode or around an AI Overview. AI Max reporting and placement segmentation are different questions. For every metric, the readiness register therefore stores its exact definition, available dimension, data gap and permitted interpretation. What cannot be segmented must not be reconstructed from an aggregate trend.
Evaluation follows an evidence chain. First, verify that the function was active and technically stable; then observe queries, assets and destinations; next, validate conversion assignment and lead quality; and only then assess the business effect within the permitted experiment design. A higher conversion count is not a complete decision without goal quality, cost, value and Sales confirmation.
Which setting was genuinely active, when it was active and to what extent.
Which search terms, ads, assets and landing pages were observed.
Which action was counted, attributed, deduplicated and imported.
Which leads, sales, values or margins were confirmed outside the advertising platform.
Assign clear ownership for product monitoring, campaigns, content, data and Sales
AI features increase the number of handovers and therefore the need for explicit responsibility
Paid Search cannot own a new feature alone. Marketing maintains the register and hypothesis, website or content owners review landing pages and claims, Analytics verifies measurement and data flows, CRM or Sales confirms lead status, and an authorised role decides on brand, legal and sensitive-content questions. In a small team, one person may hold several roles, but the tasks should still remain separate.
Every change receives a release sheet containing the account, campaign, date, starting state, approved functions, exclusions, a landing-page sample, conversion goals, dashboard or reports, owner, stop rule and retest date. Verbal approval should not become permanent policy. The record makes clear why automation was opened, constrained or declined.
A monthly product review follows official documentation and account notices. An operational campaign review observes budget, goals, queries, assets, URLs and diagnostics. A business-quality review evaluates leads or sales. These cadences should not become daily micromanagement: learning systems and experiments require stable conditions, while genuine errors can and should be stopped immediately.
- Product Watch: document new guidance, rollouts, deadlines and available account features.
- Paid Search: own the hypothesis, settings, budget status and reporting.
- Content and Web: maintain approved claims, page quality and URL families.
- Analytics and CRM: review goals, data flows, deduplication, diagnostics and lead status.
- Brand and subject-matter approval: confirm tone, claims, mandatory information and sensitive boundaries.
- Decision-maker: start, expand, constrain or stop an experiment and document the reason.
Stop automation when status, data or the available control surface is unreliable
Not every uncertainty requires inactivity, but every uncertainty needs a defined response
A lack of access to a new test format is not a problem that should be solved with budget. The correct status is “Monitor”. Poor conversion quality, incorrect landing pages, unapproved claims or sensitive data are operational blockers. They must be addressed before an experiment because otherwise the system learns towards the wrong outcome or scales an avoidable risk.
Insufficient stable data can also make a broad rollout unsuitable. That does not automatically mean returning to entirely manual optimisation. Options include reducing scope, choosing a different primary goal, observing for longer or repairing measurement first. The decision should follow the business task and the evidence that can realistically be produced, not a blanket pro- or anti-AI position.
Define stop rules before activation. They should cover not only performance thresholds but also technical errors, incorrect URLs, brand violations, unexplained data changes and missing CRM feedback. After a stop, keep the cause open in the register until remediation and retesting are documented; switching off a control is not, by itself, a completed diagnosis.
| Signal | Risk | Immediate response | Required evidence | Release condition |
|---|---|---|---|---|
| Format absent from account | Planning based on an assumption | Monitor; allocate no budget | Official availability | Access and rules confirmed |
| Incorrect landing pages | Irrelevant or risky delivery | Constrain URL expansion | Live sample and exclusions | Approved URL families |
| Unapproved copy | Brand or claim violation | Remove asset or stop function | Creative and source review | Business approval |
| Weak conversion signals | Optimisation towards micro-actions | Review goals and bidding | CRM and tracking reconciliation | Reliable primary goal |
| Data or consent gap | Unlawful or faulty upload | Stop the data flow | Technical and legal approval | Diagnostics and purpose confirmed |
Frequently asked questions about Google Ads 2026 and AI Search
Concise answers with clear boundaries for German SMEs
The following answers reflect the documented status on 21 August 2026. Google may change availability, defaults, terminology and reporting. Before implementation, current documentation must therefore always be considered alongside the state visible in the advertiser's own account.
Are ads in AI Mode available to every German advertiser?
No. AI Mode is available in Germany as a search feature, but the new conversational advertising formats are being tested or rolled out on a limited basis. A public example is not universal confirmation of access. Check current Google documentation and the options visible in your account.
Do I need a dedicated campaign for AI Overviews?
No. Google identifies existing Search, Shopping and Performance Max campaigns as potential inventory. There is no standalone campaign type exclusively for AI Overviews. Delivery and position depend on availability, the auction and relevance.
Can I target or disable ads in AI Overviews separately?
According to current Google documentation, neither exclusive targeting of this placement nor a separate opt-out is available. This is an important control boundary and must be recorded in the register before anyone tries to infer placement performance from aggregate reports.
Does AI Max replace my keywords?
No. AI Max can supplement keywords through broad-match, asset-based and landing-page-based methods. Keywords, search-term reports, brand rules, negative keywords, pages and conversion goals retain distinct roles. The right combination must be tested at campaign level.
Is AI Max a new campaign type?
No. Google describes AI Max as an optimisation and feature layer for existing Search campaigns. Components such as search-term matching, text customisation and final URL expansion perform different jobs and should not be treated as one indivisible switch.
When will Dynamic Search Ads be transitioned automatically?
Under Google's timetable updated in June 2026, the automatic DSA transition begins in February 2027. September 2026 remains relevant for automatically created assets and the campaign-level broad match setting. Check the specific notice in your own account.
Do Enhanced Conversions make the setup automatically GDPR-compliant?
No. The technology can support measurement, but it does not replace purpose assessment, a legal basis, consent, data minimisation, policy checks or appropriate handling of sensitive data. The specific implementation requires suitable technical and legal approval.
How can I tell whether a change is working?
Start with technical activation and stable diagnostics, then review search terms, assets and actual destination pages, followed by correctly assigned conversions and confirmed lead or sales quality. A predefined experiment provides the strongest basis for a cautious causal decision; an aggregate trend does not establish why a result changed.
Turn an AI announcement into a controllable Google Ads operating process
Google Ads in 2026 is expanding Search, automation and the range of possible advertising formats. For German SMEs, however, the most important capability is not being first to activate every feature. It is the ability to connect rollout status, campaign role, the website, assets, conversion goals, first-party data, controls and reports in a decision that can be reviewed.
The Google Ads 2026 readiness register brings these elements into one decision framework. It distinguishes Monitor, Prepare, Test, Scale and Stop; documents evidence and ownership; and prevents unavailable formats from entering forecasts or non-segmentable delivery from being presented as reliably measured success attributable to an AI environment.
The defensible sequence is straightforward: verify official status, define the business task, repair goals and data, approve pages and creative assets, establish controls, run a limited experiment, interpret reports correctly and only then decide. This keeps the account open to new opportunities without exchanging control, privacy or measurement logic for an innovation promise.
Would you like a prioritised review of your account, measurement and landing pages for the current and upcoming Google Ads changes in 2026? Develop your Google Ads programme systematically with Salestudia.