New features, improvements, and policy updates in chronological order.
We refreshed the ChatGPT measurement baseline to match recent changes in how ChatGPT builds answers. ChatGPT results from October 9 onward should not be compared directly with earlier ones. Gemini and Perplexity measurement is unchanged.
The Privacy Policy now lists what we receive at Google sign-in, the payment records we store for paid purchases, the service emails we send, and how long payment records are kept after account deletion. Effective October 1, 2026.
We are adding clauses to the Terms of Service on how we announce changes to paid-plan allowances such as monthly measurement credits, when they apply to existing subscribers, and how far in advance we announce terms updates. Effective October 31, 2026; your current plan price and credits do not change.
The overall share (donut) in Competitor Comparison now uses the same method as the four individual metrics: ChatGPT, Gemini and Perplexity are weighted equally. This removes the effect where the AI with the most answers dominated the result and aligns the cards and the donut on one basis. Earlier overall-share values should not be compared directly.
Monthly measurement credits for subscriptions started on or after September 5 are now Starter 12, Growth 80 and Pro 200. Existing subscribers keep their current credits. A biweekly auto-check option and credit add-on packs were added.
As ChatGPT, Gemini and Perplexity changed the way they build answers, we refreshed all three measurement axes to the current answer environment at once. Results from September 5 onward should not be compared directly with earlier ones. Eleven new categories and a "none of these fit" report were added.
We revised list-item counting, robots.txt allow-rule parsing, page language detection, page-type-aware schema recognition, and product offer relationships. Treat your first diagnostic under the new criteria as a fresh baseline rather than comparing it with past results
Google Analytics integration is now available from the Starter plan. Once connected, the dashboard trend chart and the monthly report show your score and AI referral traffic side by side.
Body length is counted in characters for Korean and words for English, but both used the same numbers, grading English pages too strictly. English thresholds now match the word-count scale
The overall trend chart now shows grade ranges as background bands, so you can see which grade tier your score sits in and how far the next boundary is. Summary cards are consolidated, and dashboard sections are reordered to surface frequently checked information first.
Paid plan subscriptions have started. Pay with Korean cards, with pricing published on the site. The free diagnostics and trial measurement remain unchanged.
AI composes answers by splitting one question into several sub-questions. Content Coverage judges whether your site has a page to answer each expected sub-question, and turns gaps into a content backlog with suggested directions.
Three additions to Visibility Opportunities results: per-question ghost citation flags, a category ownership verdict showing whether your brand effectively owns its space in AI answers, and a comparison card that shows what changed when you re-measure.
Our first category benchmark measured 12 Korean B2B SaaS domains in real ChatGPT, Gemini, and Perplexity answers — median score 51/100 and a large platform gap in mention rates (33% vs 83%), published as anonymized aggregates.
The dashboard now tells you whether your current grade is stable, fluctuating, or trending — based on the flow of your recent measurements.
We now distinguish sources the AI verifiably referenced from URLs it merely wrote into the answer — so the screen also tells you how much to trust the diagnosis.
The gauge now reflects how far you have come within your current grade, so improvement shows on screen before the grade itself changes.
Fixed duplicate domains in the page structure diagnosis "Recent results" list, and a bug where going back from a report re-ran the diagnosis.
The competitor comparison page now shows your recent comparison runs. Click any entry to fully restore that past result, including share metrics, the heatmap, and citation sources. Available on Starter plans and above
A new "Report" item has been added to the Overview menu. Review your registered brand's monthly grade changes, visibility share, per-LLM grades, competitor comparison, and next-month recommendations in a one-page report and save it as a PDF. Available on Starter plans and above
The Visibility Opportunities start screen now shows a list of your recent analyses, so you can reopen past results and track how your exposure gaps improve over time. Available on Starter plans and above
Brand mention detection in Korean AI answers no longer mistakes names embedded inside unrelated compound words for mentions, and topic relevance now correctly handles category names joined with the Korean middle dot (·)
Measurements that receive no response from any AI engine are no longer treated as results. Credits are refunded automatically, and no misleading zero score is saved to your results, history, or dashboard trends
Diagnosis used to fail for sites with HTTPS (SSL) configuration issues; it now automatically continues via the www address when available, and clearly explains the cause when the site is unreachable
We published and refreshed cornerstone guides and glossary entries on what makes a brand get recommended in AI answers, how Korean GEO tools compare, and how AI citation decay and content freshness work
You can now switch the prompt coverage trend on your dashboard between weekly and monthly views, and see a summary of the prompts where your brand recently went unmentioned
Sharing a diagnostic report link on social media or messengers showed an empty or broken preview image; the score card now renders correctly
Site Diagnostic now follows pages that auto-redirect to their real content, and when a page has too little content to score, it guides you to Page Structure Diagnostics instead of showing an error
We added Yeti (Naver crawler) checks for the AI Briefing era, reclassified llms.txt as a reference item based on evidence, and added a reference card that maps your signals to Google and Naver AI surfaces
Visibility Opportunities now validates AI responses more strictly so invalid answers never reach your results, and automatically refunds credits whenever a measurement run completely fails
Mention detection is now more precise, fixing cases that were miscounted; competitor comparison billing edge cases were corrected; and Korean text no longer appears on English screens
Page diagnostics, brand analysis, competitor comparison, and the dashboard all respond faster; the dashboard trend chart now uses accurate ratio baselines; and visual consistency was cleaned up across screens
A single account-wide master switch now controls all alerts, while each domain picks its own email and Slack channels — making alert behavior predictable at a glance
Improved Korean typography site-wide for a sharper, more consistent screen, and separated blog & glossary content into white cards so long reads are easier
Added a citation sources card to your Visibility Opportunities results. It classifies the sources cited in AI answers by type, showing where brand recommendation is weak
Precise analysis of only the sentence around the brand mention and a clear tri-state view of positive·neutral·negative
RanketAI launched a monthly AI answer citation measurement report hub — "AI Visibility Benchmarks". The first issue discloses how RanketAI progressed from grade D to C in 25 days using own media only.
A new menu that helps you find where your brand exposure is weak in AI answers, and get improvement guidance — all in one screen
Unified schema.org metadata across the site so AI answers from ChatGPT, Perplexity, and Gemini consistently recognize RanketAI as a single brand entity
Connect Google Analytics to see click-through visits arriving from AI answers like ChatGPT, Perplexity, and Gemini, right on your dashboard
A market comparison guide and industry meta-term glossary are now available
Measurement stability and result consistency have been reinforced together
A record save error that occurred under specific conditions has been fixed
Two new long-reads: a 13-question visibility playbook and a Google AI Mode shift analysis
Broader Korean answer recognition and measurement that reflects market context
Scheduled diagnostics, change detection, and alerts now work as one flow
How major Schema.org types affect AI answer citation, with practical usage notes
Official policies and robots.txt setup for the major AI crawlers, in one place
The "what to do next" guide after a diagnosis is now a simple four-step workflow
Brand entity now links to academic and standards-body authorities to improve AI citation signals
Two pillar guides comparing GEO and AEO analysis tools, with selection criteria
Evaluate a domain’s AI citation readiness with a single overall grade
The most important indicator now leads the dashboard, with step-by-step progress during diagnostics
Diagnosed pages now accumulate per domain so you can track changes over time
Score changes now come with the context needed to interpret them
AI Brand Exposure Diagnostics now measures across ChatGPT, Gemini, and Perplexity
Unified visual language across the dashboard, blog, and marketing pages
Full answer text, action guides, and cross-answer analysis now live in the result page
We unified the service name as RanketAI and moved to ranketai.com
A free tool that scores how prepared a URL is to be cited in AI answers