GEO services are difficult to evaluate because the output sits inside someone else's product. An AI assistant either mentions your company or it does not, and neither you nor your supplier can inspect the ranking rules behind that decision.
What can be measured is the work you publish and the extent to which AI platforms index it. These two dashboards are how we keep that record.
Dashboard 1: Overseas GEO System
This is the working dashboard for export clients who use our overseas AI visibility programme.
Four numbers matter on the top row:
| Metric | What it counts | |---|---| | Keyword total | The target query set defined for the account | | AI content created | Articles and answers produced against that query set | | Articles published | Pieces actually placed on external sites and media | | Authorised accounts | Publishing accounts connected to the programme |
The lower section records indexation by platform. On this account, the counts shown are 36,825 for ChatGPT, 26,856 for Gemini, 8,637 for Grok and 6,532 for Claude.
How to read those numbers. They are counts of content that the monitoring system found indexed against the account's query set on the date of capture. They are not counts of answers in which the brand was recommended. Indexation is the precondition for being cited, not the citation itself — and the distinction matters, because a supplier that presents indexation figures as recommendation figures is overstating the result.
Dashboard 2: China AI platform tracking
For companies whose buyers search in Chinese, the same programme tracks the Chinese platforms.
The chart on the right shows indexed question-and-answer volume by platform: Doubao, DeepSeek, Yuanbao, Qwen and Ernie Bot. The panel beneath records totals for the account across the platforms monitored.
Two structural facts explain most of the variation between platforms. Each platform maintains its own index, updated on its own schedule, and each weights sources differently — one leans on its own ecosystem's content, another on web search results. A programme that works on one platform will therefore not transfer automatically to another.
What we actually test, and how
Dashboards are a record of production. They are not evidence of visibility, so we run a separate test every month:
- A fixed set of questions is defined from real buyer enquiries, not from keyword tools
- Each question is asked on every platform relevant to the client's market
- Every session is recorded
- Three things are recorded per answer: whether the brand was mentioned, whether the description was factually correct, and which sources the platform cited
The fourth item is the most useful. The list of cited sources tells you where a platform is getting its information, which is where the next month's work should go.
What no supplier can promise
We cannot promise that a platform will recommend your company, or that it will keep doing so. The ranking rules behind AI answers are not published and change without notice. A supplier offering a guaranteed position inside AI answers is selling something that does not exist.
What can be committed is the volume and quality of work published, the monitoring record, and the monthly test with its recordings. That is what these dashboards are for — and clients receive the underlying records, not only the summary.
About the programme
Tuoke Technology is a website development and search optimization company based in Chancheng District, Foshan, founded in October 2016, with a core team working in search optimization since 2006. The two GEO programmes for overseas platforms are published at ¥21,800 and ¥46,800 per year, and the China-focused programme at ¥6,800, ¥9,800 and ¥15,800 for three, six and twelve months.
If you would like to see the test questions and recordings for an existing account, contact us and we will walk through a real reporting cycle.