GEO WORKSPACE
See whether selected Chinese AI engines mention, recommend or misrepresent your brand when buyers compare options. GEO Workspace runs structured, reviewed Simplified-Chinese assessments, retains captured answers and returned sources, and monitors approved questions against a published baseline. Results reflect defined test conditions, not guaranteed recommendations or exact reproduction of consumer apps.
Discuss an expert-led assessment200 GEO Credits included · No credit card required
Need a specialist traveller audit? Explore Traveller GEO20-question standard assessment
Published baseline comparisons
Model Only + Search Enhanced
Evidence-linked results and reports
Workflow
Set the entity, market, audience, category, and competitors.
Generate and review natural Simplified-Chinese questions around real decisions rather than brand-name vanity prompts.
Run approved questions across selected Chinese AI engines and preserve the exact answer, returned sources, and execution conditions.
Review evidence-linked findings and opportunities, then rerun, investigate specific questions, or move into specialist support.
Product
Projects hold a frozen scope, versioned question plan, configured engines, evidence-linked results, and immutable report—not a generic prompt log.
How representative are provider-based GEO assessments? CTD compared matched Simplified-Chinese questions across Chinese AI apps and provider-based tests. The research identifies where commercial decision signals aligned—and where entity definition, question design or execution conditions changed the findings. Read the methodology and limitations.
Real Assessment Example
An anonymised assessment of a jewellery brand with a boutique in Singapore.
When mainland Chinese consumers ask AI which jewellery brands to explore in Singapore, will a brand appear naturally in the answer — or only when shoppers already know its name?
The assessment tested 20 Simplified-Chinese questions across six configured Chinese AI surfaces: DeepSeek, Doubao, Qwen, Kimi, ERNIE and Yuanbao.
It retained 120 accepted AI-answer observations covering brand discovery, competitor comparisons, store information and purchase decisions.
从中国大陆去新加坡某购物中心逛腕表和高级珠宝，有哪些品牌值得先了解？Which watch and fine jewellery brands are worth exploring at a shopping centre in Singapore?
The original question's shopping-centre name has been removed for anonymity. The results below were measured using the original, unmodified question.
The commercial insight
The brand was consistently visible when explicitly named, but less consistently surfaced when Chinese shoppers explored jewellery brands without specifying one.
This distinction helps identify where brands may be missing from early-stage consumer consideration.
GEO Workspace helps businesses examine:
These findings support targeted GEO decisions and future measurement, rather than guaranteed changes in AI recommendations.
Methodology note. Based on one anonymised assessment captured in September 2026. Results reflect specific questions, configurations and provider-based observations. They are not permanent AI rankings or proof of identical behaviour in consumer-facing apps. This illustration does not imply a client relationship or endorsement.
Capabilities
Use the evidence to prioritise improvements, investigate gaps with Advanced Analysis or Research, then review the same approved questions over time.
Build a structured China AI visibility assessment using a versioned question plan, selected Chinese AI engines, evidence-linked analysis, and an immutable report.
Choose weekly, monthly or quarterly review reminders. When a cycle is due, approve a fresh run before comparing changes, stable areas and coverage limits.
Go beyond individual answers to investigate patterns across the full assessment.
Investigate specific Chinese questions across configured engines without changing a formal assessment.
Monitor
Choose weekly, monthly or quarterly review reminders. When a cycle is due, approve a fresh run before comparing changes, stable areas and coverage limits.
Establish a published report baseline that freezes the approved questions, scope and configured engines.
Set a weekly, monthly or quarterly cadence. Due reminders notify workspace owners and administrators and use no GEO Credits.
An owner or administrator reviews the current run and GEO Credit quote, then explicitly approves execution.
Review matched question–engine evidence for material changes, stable areas and what could not be compared, with original answers and returned sources.
Reminders do not launch AI runs. Follow-up execution uses the existing GEO Credit rates; technical provider failures are not charged. Comparisons retain the immutable baseline and show coverage limits.
Start 7-day free trialChina-specific
Execution questions are built and preserved in Simplified Chinese so the evidence reflects the language being tested.
Compare selected engines from DeepSeek, Doubao, Yuanbao, Qwen, ERNIE and Kimi across the same structured Chinese decision questions.
Keep the exact question, provider response, returned sources, and capture conditions behind the analysis.
Examine discovery, comparison, substitutes, trust signals, and whether the answer gives a usable next step.
Use GEO Workspace independently, or work with CTD specialists to set up assessments, interpret findings, prepare leadership presentations and manage ongoing reviews.
Configure and review your first China GEO assessment.
US$1,250
Turn assessment findings into clear priorities and recommended actions.
US$2,250
Present your GEO findings and recommendations to senior leadership.
US$4,250
Ongoing specialist assessment, monitoring and strategic guidance.
from US$3,000/month
Professional fees exclude GEO Workspace subscriptions and GEO Credits.
Plans
All free trials start with 200 GEO Credits and one project. Choose your paid plan when you upgrade.
6 China AI engines · Model Only + Search Enhanced · Monitoring · Research · Advanced analysis · Evidence-linked reports
CORE
PRO
BUSINESS
ENTERPRISE
For organisations that need custom scale, procurement, security or operating controls.
How GEO Credits work: 1 GEO Credit = one question × one AI engine in Model Only mode. Search Enhanced uses 1.5 credits. Unused subscription credits roll over up to 3× the monthly allowance.
Assessment equivalents assume 20 questions × 6 AI engines in Model Only mode. Search Enhanced and Advanced analysis use additional credits.
GEO Credits
GEO Credits measure completed AI execution, not provider tokens. You see estimated usage before a run, and the actual charge is reconciled from billable completed responses.
20 questions × 5 engines: 100 GEO Credits with Model Only; up to 150 with Search Enhanced where all selected engines support it.
GEO Workspace
For repeatable Chinese AI assessment, approval-required monitoring, deeper analysis and research across selected Chinese AI engines, with evidence-linked reporting.
Traveller GEO Concierge
For broader traveller-specific evidence programmes across AI, Chinese social search, travel platforms, maps, and scoped owned or partner paths with analyst review.
Explore China Traveller GEO ConciergeFAQ
GEO Workspace is the self-service China Audience GEO platform from ChinaMarketing.AI. It lets teams run structured Simplified-Chinese questions across configured Chinese AI engines, preserve the evidence behind each response, and turn results into a published report baseline for approval-required monitoring.
No. GEO Workspace is a self-service Chinese AI assessment, monitoring and research product. The China Traveller Full Audit is a managed, traveller-specific evidence programme that can extend across AI, Chinese social search, travel platforms, maps, and owned or partner action paths with analyst review.
Review reminders are scheduled automatically on a weekly, monthly or quarterly cadence. An owner or administrator must approve each AI run and its current GEO Credit quote. A due reminder alone makes no provider call and uses no credits.
Start from a published report baseline, then compare matched approved questions and configured engines. Review material changes, stable areas and coverage limits with the original answer and source evidence. Changed scope or missing evidence can limit comparison; monitoring does not measure representative consumer share.
No. GEO Workspace monitoring is self-service and requires owner or administrator approval for each follow-up run. Managed Traveller GEO Monitor & Optimize is a separate specialist service.
One successful Model Only question and engine response uses one GEO Credit. A successful Search Enhanced response uses 1.5 credits. Technical provider failures use zero credits.
The seven-day trial includes 200 one-time GEO Credits. A standard 20-question, five-engine Model Only assessment uses 100 credits; an all-Search-Enhanced equivalent uses up to 150 credits where supported.
Simplified Chinese is the canonical execution language. Native Simplified-Chinese input can be used directly. Where supported, English input can be converted into proposed Simplified Chinese for review before execution.
No. The platform measures observed answers within tested questions, engines, and execution conditions. It does not control external AI systems, guarantee recommendations, or prove sales attribution.
Search Enhanced uses a provider execution path with search or retrieval where supported. If a selected engine cannot use that mode, the effective execution condition is shown and usage is settled according to the mode that actually ran.
Trial credits expire after seven days. Completed work remains available according to account retention, and a paid plan is required before running new work.
Start with evidence
Start with 200 GEO Credits and run the assessment on your own brand, market, or category.