
AI Search Algorithm and Platform Change Tracker
A sourced tracker of material AI search platform changes affecting publisher eligibility, citations, measurement and user discovery.
This tracker records sourced changes that materially affect publisher eligibility, citations, visibility measurement and user discovery in AI-assisted search. It excludes rumor, unsourced ranking claims and routine model speculation.
Last reviewed: 13 July 2026. Product access varies by market, account and rollout. Check the linked primary source before making a platform decision.
How Changes Are Selected
A change is included when an official provider source documents at least one of the following:
- a new search or answer surface;
- a publisher crawler or control;
- a change to source links or citation reporting;
- a new measurement report;
- a material market or user expansion;
- a documented retrieval or query behavior;
- a publisher action required to preserve eligibility.
Each entry separates the documented change from TotalAuthority's practical interpretation. Interpretation is not a claim about an undisclosed ranking algorithm.
Current Operating Summary
As of July 2026:
- Google says normal Search eligibility and SEO fundamentals apply to AI Overviews and AI Mode; no special AI schema or machine-readable AI file is required.
- Google has begun rolling out dedicated generative AI performance reports in Search Console to a subset of sites.
- Bing Webmaster Tools offers AI Performance in public preview across supported Microsoft AI experiences.
- OpenAI documents OAI-SearchBot for search visibility and GPTBot separately for potential training controls.
- ChatGPT search, Claude web search and Google AI Mode provide linked web research experiences.
- Cross-platform visibility still requires surface-specific prompt monitoring because official metrics differ.
Timeline
31 October 2024 — OpenAI launches ChatGPT search
Documented change: OpenAI introduced ChatGPT search with timely answers and links to web sources. OpenAI said ChatGPT could choose to search based on the question or the user could select web search.
Publisher impact: Brand and content discovery could occur inside a conversational answer with a source panel rather than only through a conventional results page.
Action: Establish ChatGPT-specific prompt and referral monitoring. Do not assume conventional rank position predicts citation.
16 December 2024 and 5 February 2025 — ChatGPT search access expands
Documented change: OpenAI's launch page records expansion to logged-in users in December 2024 and then to everyone in supported regions without signup in February 2025.
Publisher impact: A broader audience could use the search experience, increasing the need to separate ChatGPT referrals from total AI visibility.
Action: Preserve utm_source=chatgpt.com and referring-host data where available.
5 March 2025 — Google introduces AI Mode in Labs
Documented change: Google announced an experimental AI Mode using Gemini capabilities for complex questions, comparisons, reasoning, follow-ups and web links.
Publisher impact: The unit of discovery expanded from one query to a conversational research journey.
Action: Add multi-part and follow-up tasks to Google-specific prompt research.
20 March and 27 May 2025 — Claude web search launches and expands globally
Documented change: Anthropic announced web search for current information with direct citations, later updating availability to all Claude plans globally.
Publisher impact: Claude became a distinct citation surface that should be tested independently rather than inferred from ChatGPT or Google performance.
Action: Add Claude to the benchmark only where it is relevant to the audience and record whether web search is active.
20 May 2025 — Google rolls AI Mode out in the United States
Documented change: Google announced U.S. rollout without a Labs signup and described query fan-out, in which the system breaks questions into subtopics and issues multiple searches.
Publisher impact: Supporting pages may be discovered through subquestions that do not exactly match the original wording.
Action: Build topical coverage around real tasks and evidence, not hundreds of keyword-swapped pages.
21 August 2025 — Google expands AI Mode and agentic features
Documented change: Google announced availability in more than 180 countries and territories in English, alongside agentic and personalized capabilities in selected experiences.
Publisher impact: Market and account context became even more important to testing; one U.S. session could not represent global output.
Action: Segment prompts by country, language, account and product mode.
30 September 2025 — Google expands visual exploration in AI Mode
Documented change: Google announced conversational visual results with links, supported by multimodal understanding and its Shopping Graph.
Publisher impact: High-quality images, product data and visual context became more relevant to applicable discovery tasks.
Action: Maintain image quality, descriptive context, merchant data and rights alongside text evidence.
18 November 2025 — Gemini 3 enters Google Search AI Mode
Documented change: Google announced Gemini 3 in Search, starting with AI Mode, with new reasoning and generative interface capabilities.
Publisher impact: Model and interface changes can shift answer and link patterns without a publisher-side change.
Action: Annotate the event in visibility reports and avoid attributing every trend to site work.
10 December 2025 — Google publishes consolidated AI feature guidance
Documented change: Google's Search Central guidance states that existing SEO fundamentals remain relevant, supporting links require normal Search eligibility and snippet eligibility, and no special AI markup or file is needed.
Publisher impact: Technical SEO, useful textual content, internal links, accurate structured data and current business/product information remain the documented foundation.
Action: Reject unsupported “AI schema” or mandatory AI-file claims. Use the technical GEO audit.
10 February 2026 — Bing launches AI Performance public preview
Documented change: Bing Webmaster Tools introduced total citations, average cited pages, sampled grounding queries, page-level citation activity and visibility trends across supported Microsoft AI experiences.
Publisher impact: Publishers gained official citation reporting beyond referral traffic.
Action: Verify the property, establish a baseline and treat citation counts as visibility—not rank, authority or traffic. Microsoft explicitly warns against those interpretations.
3 June 2026 — Google launches generative AI performance reports
Documented change: Google announced dedicated Search Console reports for generative AI visibility in Search and Discover, initially for a subset of websites. Documented dimensions include impressions, pages, countries, devices in Search and dates.
Publisher impact: Google AI visibility became more directly observable without relying only on overall Web performance.
Action: Use the dedicated view where available, preserve rollout context and combine it with page analytics and cross-platform prompt tracking.
June 2026 — OpenAI updates publisher guidance
Documented change: OpenAI's publisher FAQ states that public sites can appear in ChatGPT search, recommends allowing OAI-SearchBot for summaries and snippets, documents utm_source=chatgpt.com, and distinguishes GPTBot for potential training controls.
Publisher impact: Search inclusion and potential training choices can be governed separately.
Action: Review robots.txt, WAF behavior and analytics. Document the legal and business decision for each crawler instead of using one blanket rule.
What Is Not in This Tracker
- Unverified ranking-factor lists.
- Anecdotes from one prompt run.
- Model benchmark releases with no publisher impact.
- Minor interface experiments without a primary source.
- Vendor claims that a tactic guarantees citations.
- Rumored traffic or click-through statistics.
Change Impact Framework
When a new platform event occurs, assess five areas.
Eligibility
Did crawler, index, snippet or product-data requirements change?
Measurement
Did the provider expose a new report, dimension, referrer or limitation?
User behavior
Did the experience add follow-ups, multimodal input, personalization or agentic actions?
Evidence
Did the provider document new source types, data partners or retrieval behavior?
Operations
Does the team need to change prompts, annotations, dashboards, crawler policy or content maintenance?
How to Maintain an Internal Tracker
Use columns for:
- effective date;
- announcement date;
- platform and surface;
- official source;
- documented change;
- affected markets;
- publisher impact;
- required action;
- owner;
- implementation status;
- evidence after change;
- review date.
Review monthly, but issue an alert only when the event changes a material decision. Daily noise can cause teams to rewrite strategy around temporary interface tests.
How to Analyze Performance Around an Update
- Mark the event date in analytics and visibility tools.
- confirm the affected product and markets.
- preserve an unchanged prompt benchmark.
- compare pre- and post-period distributions.
- inspect technical releases and campaigns that occurred concurrently.
- segment pages, prompts and platforms.
- report association rather than causality unless the design supports it.
Use the measurement frequency guide and statistical confidence framework to prevent false alarms.
Frequently Asked Questions
Is there one AI search algorithm update list?
No. Each provider operates different products, retrieval systems, models and rollouts. Track exact surfaces.
Should content be rewritten after every model launch?
No. First identify whether the change affected publisher eligibility, evidence selection, measurement or user behavior. Keep useful accurate content stable unless a real gap exists.
Are all rollout dates global?
No. Many changes begin in selected markets, languages, accounts or subscription tiers. Check the primary source and record the test environment.
How often is this page reviewed?
Monthly and after material official announcements. The current review date is 13 July 2026.
Sources
- OpenAI: Introducing ChatGPT search
- OpenAI: Publishers and Developers FAQ
- Google: Introducing AI Mode
- Google: AI Mode at I/O 2025
- Google: AI Mode global and agentic expansion
- Google: Visual exploration in AI Mode
- Google: AI features and your website
- Google: Generative AI performance reports
- Microsoft: AI Performance in Bing Webmaster Tools
- Anthropic: Claude web search




