As new content gets published, competitors gain authority, or AI systems update the information they rely on, the sources and brand names included in AI responses change. So, while a single AI visibility check tells you whether your business appears in the response at a given point in time, it cannot tell whether that visibility is improving, declining, or fluctuating across topics.
This is where historical AI search data becomes useful. Instead of treating AI visibility as a one-time snapshot, it allows marketers, SEO teams, and PR teams to investigate how brand mentions change over time.
In this article, we’ll look at practical ways to use historical endpoints in the DataForSEO LLM Mentions API to track a brand’s AI visibility and create real GEO product features and reporting workflows.
Contents
What are LLM Mentions?
What you can measure with historical data in the LLM Mentions API
1: Track AI visibility trends over time
2: Measure AI visibility change between two dates
3: Monitor newly gained and lost AI mentions
What to do next: investigate the causes and optimize for more AI visibility
Final takeaway
What are LLM Mentions?
An LLM mention is a general term for any occurrence of a brand, domain, product, or link to your website in AI search environments. This can mean, for example, your brand name showing up in ChatGPT’s answer or your blog post being referenced in a Google AI Overview. The latter is a specific type of an LLM mention called a citation. It means the AI model includes a reference to your content, such as a blog article or landing page. However, a brand can be mentioned without being cited. For instance, an AI answer might name your company as a relevant solution but link to a third-party article instead of your website. This distinction is important because mentions show AI visibility, while citations show source authority and can also drive referral traffic.
Together, mentions and citations create a structured data layer for AI visibility tracking and measuring changes over time, and our LLM Mentions API makes this data accessible at scale. Instead of manually checking results for individual prompts or building custom scrapers and maintaining large historical datasets, teams can access structured JSON data on AI search performance through an API. It’s easy to use and ready to integrate into a wide range of workflows: from monthly reporting and brand monitoring in AI search to building GEO tools, dashboards, and related product features.
Depending on the endpoint, the LLM Mentions API can show up-to-date information, historical trends, changes between two dates, newly gained and lost mentions, and the brands or categories most frequently mentioned for a topic. With historical data, teams can monitor changes over time and connect them to real business activities, such as product launches, PR campaigns, or content updates.
What you can measure with historical data in the LLM Mentions API
A number of endpoints in the LLM Mentions API now offer historical AI search data that’s available back to August 1, 2025. These endpoints are designed to answer three simple but important questions.
First, are we becoming more or less prominent in AI answers over time? The Historical LLM Mentions endpoint helps to answer this with detailed month-by-month metrics for the domains or keywords you specify. For each month, it provides the total number of mentions or citations (depending on your settings) and their overall AI search volume estimation.
Second, the Timeseries Delta endpoint helps to measure exactly how much your AI visibility changed between two dates. This is useful when you want to compare a before-and-after, such as the month before and the month after a new feature release campaign.
The third question – how many mentions were gained and how many disappeared between two dates – is best answered with the data from the New and Lost endpoint. It provides detailed metrics on newly acquired and recently lost mentions or citations in AI responses.
These endpoints can support a practical AI search-tracking workflow, where your team can monitor trends, compare changes, and measure what exactly changed. And it can be built in three simple steps that we’ll cover in more detail in the following sections.
1: Track AI visibility trends over time
The first step is to look back and assess whether your AI visibility is increasing, declining, or staying flat. This is important because a single current snapshot does not explain the direction, context, or cause of changes in visibility. So, before a company like Notion can plan its next move, they need to understand in which direction they are moving in AI search and why.
To get this critical context, their team can use the Live LLM Mentions Historical endpoint, which returns historical monthly counts of AI mentions for a domain or keyword.
Endpoint:
POST https://api.dataforseo.com/v3/ai_optimization/llm_mentions/historical/live
Example request:
[
{
"language_name": "English",
"location_code": 2840,
"platform": "google",
"target": [
{
"domain": "notion.com",
"search_filter": "include",
"search_scope": ["sources"],
"include_subdomains": true
}
]
}
]
In this request, the API checks historical metrics for notion.com citations in Google AI Overviews for the English language and the United States location. The response returns monthly items, and each item contains a year, month, and metrics, as in this example:
{
"version": "0.1.20260610",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.5396 sec.",
"cost": 0.101,
"tasks_count": 1,
"tasks_error": 0,
"tasks": [
{
"id": "07071240-8284-0662-0000-191db36bc4c0",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.4741 sec.",
"cost": 0.101,
"result_count": 1,
"path": [
"v3",
"ai_optimization",
"llm_mentions",
"historical",
"live"
],
"data": {
"api": "ai_optimization",
"function": "historical",
"language_name": "English",
"location_code": 2840,
"platform": "google",
"target": [
{
"domain": "notion.com",
"search_filter": "include",
"search_scope": [
"sources"
],
"include_subdomains": true
}
]
},
"result": [
{
"items_count": 12,
"items": [
{
"year": 2026,
"month": 6,
"metrics": {
"mentions": 1777,
"ai_search_volume": 1575610
}
},
{
"year": 2026,
"month": 5,
"metrics": {
"mentions": 3277,
"ai_search_volume": 1915510
}
},
{
"year": 2026,
"month": 4,
"metrics": {
"mentions": 2149,
"ai_search_volume": 431120
}
},
{
"year": 2026,
"month": 3,
"metrics": {
"mentions": 2684,
"ai_search_volume": 2603630
}
},
{
"year": 2026,
"month": 2,
"metrics": {
"mentions": 1242,
"ai_search_volume": 231570
}
},
{
"year": 2026,
"month": 1,
"metrics": {
"mentions": 1201,
"ai_search_volume": 776820
}
},
{
"year": 2025,
"month": 12,
"metrics": {
"mentions": 1678,
"ai_search_volume": 1198700
}
},
{
"year": 2025,
"month": 11,
"metrics": {
"mentions": 2930,
"ai_search_volume": 1863780
}
},
{
"year": 2025,
"month": 10,
"metrics": {
"mentions": 2836,
"ai_search_volume": 1675290
}
},
{
"year": 2025,
"month": 9,
"metrics": {
"mentions": 1834,
"ai_search_volume": 865770
}
},
{
"year": 2025,
"month": 8,
"metrics": {
"mentions": 2752,
"ai_search_volume": 1304480
}
}
]
}
]
}
]
}
The mentions value indicates how many times the target domain was mentioned on the selected platform in the set location and language that month. The ai_search_volume is an estimate of the total search demand associated with those AI mentions.
For reporting, you can turn this API response into a simple table:
| Month | Mentions | AI search volume |
| January | 1,201 | 776,820 |
| February | 1,242 | 231,570 |
| March | 2,684 | 2,603,630 |
| April | 2,149 | 431,120 |
| May | 3,277 | 1,915,510 |
| June | 1,777 | 1,575,610 |
Besides total monthly mentions and AI search volume, your “AI Visibility Trend” report could also include comparisons between Google AI Overviews and ChatGPT, or across primary markets or languages.
Helpful tips: choosing the right target for each business goal
Different goals require different target settings in your LLM Mentions API request. When tracking brand visibility over time, start with your main domain and brand name.
For example:
"target": [
{
"domain": "notion.com",
"search_filter": "include",
"include_subdomains": true
},
{
"keyword": "notion",
"search_scope": ["answer"],
"match_type": "word_match"
}
]
Looking at both can provide a more complete view of visibility, as AI answers may mention the brand without citing the website, or cite the website without using the brand name directly. The target array in LLM Mentions API requests can include up to 10 entities, so you can easily monitor several product names, relevant category terms, or important branded phrases in one request.
To track product visibility, use your product name(s).
"target": [
{
"keyword": "notion ai",
"search_scope": ["answer"],
"match_type": "word_match"
},
{
"keyword": "notion api",
"search_scope": ["answer"],
"match_type": "word_match"
}
]
To track topic visibility, use a non-branded keyword and search in questions or answers. This will help you identify whether a topic is appearing in AI search environments.
"target": [
{
"keyword": "best project and task management tools",
"search_scope": ["question"],
"match_type": "word_match"
}
]
If you’d like to exclude your own domain and study the market, use search_filter: "exclude" for your domain and combine it with a keyword target. With this setup, you can analyze AI responses around a specific topic where your domain currently does not appear in AI conversations.
"target": [
{
"domain": "notion.com",
"search_filter": "exclude"
},
{
"keyword": "best project and task management tools",
"search_scope": ["answer"],
"match_type": "word_match"
}
]
2: Measure AI visibility change between two dates
Historical trend data can show whether mentions are generally increasing or decreasing over time. But teams often need a more direct comparison, such as this month versus last month, before and after a product launch, or before and after a major content update. This is where the Timeseries Delta endpoint becomes useful.
A typical request compares performance between the selected periods for the same target, platform, location, and language settings. In the example below, we compare the AI visibility of the notion.com domain on April vs May and May vs June.
Endpoint:
POST https://api.dataforseo.com/v3/ai_optimization/llm_mentions/timeseries_delta/live
Example request:
[
{
"language_code": "en",
"location_code": 2840,
"platform": "google",
"date_from": "2026-05-01",
"date_to": "2026-06-01",
"group_range": "month",
"target": [
{
"domain": "notion.com",
"search_filter": "include",
"search_scope": ["sources"],
"include_subdomains": true
}
]
}
]
The response will help you understand how mentions and AI search volume changed between the comparison dates.
{
"version": "0.1.20260610",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.4523 sec.",
"cost": 0.102,
"tasks_count": 1,
"tasks_error": 0,
"tasks": [
{
"id": "07071301-8284-0663-0000-14c4b8d4c160",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.3902 sec.",
"cost": 0.102,
"result_count": 1,
"path": [
"v3",
"ai_optimization",
"llm_mentions",
"timeseries_delta",
"live"
],
"data": {
"api": "ai_optimization",
"function": "timeseries_delta",
"language_code": "en",
"location_code": 2840,
"platform": "google",
"date_from": "2026-05-01",
"date_to": "2026-06-01",
"group_range": "month",
"target": [
{
"domain": "notion.com",
"search_filter": "include",
"search_scope": [
"sources"
],
"include_subdomains": true
}
]
},
"result": [
{
"items_count": 2,
"items": [
{
"date": "2026-05-01",
"delta_mentions": 1128,
"delta_ai_search_volume": 1484390
},
{
"date": "2026-06-01",
"delta_mentions": -1517,
"delta_ai_search_volume": -352140
}
]
}
]
}
]
}
If the delta is positive, the brand’s visibility between the two dates increased; if the delta is negative, the brand lost visibility; and if the value is small or close to zero, there was little measurable change. This data is especially useful for before-and-after analysis around important business activities. For example, a team may use this endpoint to compare visibility:
- from one month to the next;
- before and after a product launch;
- before and after a PR campaign;
- before and after entering a new topic or category.
When combining the data from the Timeseries Delta and the Historical endpoints, you can create a simple but more informative reporting table like this:
| Month | Mentions | Change | AI search volume | AI SV Change |
| May | 3,277 | +1,128 | 1,915,510 | +1,484,390 |
| June | 1,777 | -1,517 | 1,575,610 | -352,140 |
The results show that between June 1 and May 1, the brands total citations dropped by 1,517, and the AI search volume of the queries for which they had citations in June decreased by 352,140. These two metrics are best interpreted together to understand both the scale and demand context of the change, that is, approximately how many searches are performed with prompts that trigger a brand mention. If mentions increased but the search volume decreased, the brand may be appearing in more AI answers, but for lower-demand queries. If mentions decreased but AI search volume has grown, the brand may be appearing in fewer answers, but ones connected to higher-demand searches.
Practical tips for this use case
Choose the date_from and date_to settings in your API request based on the type of comparison you want to make.
For regular reporting, compare the same day across your reporting periods, for instance, the first day of one month with the first day of the next:
"date_from": "2026-05-01", "date_to": "2026-06-01"
If you want to perform a before-and-after analysis, choose dates that reflect your target activity. For example, if you launched a new product page on May 15, you would want to compare visibility before that event and two or four weeks later.
When reviewing the response, do not look at mentions alone. A positive mention delta means the target appeared more often, but ai_search_volume adds demand context: a brand may gain mentions in lower-volume AI answers, or lose mentions while still appearing in higher-volume answers.
Finally, use this endpoint as the measurement step, not the full explanation. If the Timeseries Delta endpoint shows a meaningful increase or decrease, follow up with the New and Lost endpoint to measure how many AI mentions were gained or lost.
3: Monitor newly gained and lost AI mentions
If your business had 50 mentions in January and 51 mentions in February, at first glance, the net numbers look stable as if nothing changed much. But your AI visibility profile could have gained 20 new mentions and lost 19 old ones, even though the total count barely moved.
To investigate this, your team can turn to the New and Lost endpoint. It provides new and lost LLM mention counts, as well as AI search volume, for the domain or keyword specified in the target field. With the request below, we can check how many citations of Notion brand were newly gained and lost between May 1 and June 1.
Endpoint:
POST https://api.dataforseo.com/v3/ai_optimization/llm_mentions/timeseries_new_lost/live
Example request:
[
{
"language_name": "English",
"location_code": 2840,
"platform": "google",
"date_from": "2026-05-01",
"date_to": "2026-06-01",
"group_range": "month",
"target": [
{
"domain": "notion.com",
"search_filter": "include",
"search_scope": ["sources"],
"include_subdomains": true
}
]
}
]
The new values in the API response indicate AI answers that contain the target at the date_to timestamp but did not contain it at the date_from timestamp. The lost values indicate AI answers that contained the target at the date_from timestamp but no longer contain it at the date_to timestamp.
{
"version": "0.1.20260610",
"status_code": 20000,
"status_message": "Ok.",
"time": "1.1182 sec.",
"cost": 0.102,
"tasks_count": 1,
"tasks_error": 0,
"tasks": [
{
"id": "07071301-8284-0664-0000-424495833e23",
"status_code": 20000,
"status_message": "Ok.",
"time": "1.0592 sec.",
"cost": 0.102,
"result_count": 1,
"path": [
"v3",
"ai_optimization",
"llm_mentions",
"timeseries_new_lost",
"live"
],
"data": {
"api": "ai_optimization",
"function": "timeseries_new_lost",
"language_name": "English",
"location_code": 2840,
"platform": "google",
"date_from": "2026-05-01",
"date_to": "2026-06-01",
"group_range": "month",
"target": [
{
"domain": "notion.com",
"search_filter": "include",
"search_scope": [
"sources"
],
"include_subdomains": true
}
]
},
"result": [
{
"items_count": 2,
"items": [
{
"date": "2026-05-01",
"new_mentions": 2933,
"lost_mentions": 1805,
"new_ai_search_volume": 1870320,
"lost_ai_search_volume": 384970
},
{
"date": "2026-06-01",
"new_mentions": 1717,
"lost_mentions": 3234,
"new_ai_search_volume": 1539140,
"lost_ai_search_volume": 1892460
}
]
}
]
}
]
}
If we add this perspective to the data layers from the previous two steps and create a comprehensive table for reporting, it could look something like this.
| Month | Mentions | Change | AI SV | AI SV Change | New | New AI SV | Lost | Lost AI SV |
| May | 3,277 | +1,128 | 1,915,510 | +1,484,390 | 2,933 | 1,870,320 | 1,805 | 384,970 |
| June | 1,777 | -1,517 | 1,575,610 | -352,140 | 1,717 | 1,539,140 | 3,234 | 1,892,460 |
Here we see that in June, the Notions’ citations between May and June decreased by 1,517, but the brand also gained 1,717 new mentions and lost 3,234, which means the overall change was not just the result of a few small fluctuations. The brand earned some new visibility while losing substantial existing presence, and the newly gained mentions were also connected to a lower estimated search demand (an AI SV of 1,539,140) than the mentions that disappeared (1,892,460). So, this combination of data helps teams to not only measure change over time but also understand if growth came from newly won AI answers, whether losses affected high-value queries, and whether the overall visibility trend is supported by meaningful search demand.
What to do next: investigate the causes and optimize for more AI visibility
Once you have a historical AI search visibility report, the next step is to turn that data into action. The numbers can show whether your AI search presence is growing or declining, but the real value lies in understanding why those changes happened and what to optimize next.
If mentions are increasing and your brand is gaining prominence in AI answers, the goal is not only to celebrate the growth, but to understand what’s driving it. Look at the topics and query patterns through which your brand started appearing more often. These may point to new opportunities worth supporting with more content and external references.
Here’s an example request for the Search Mentions endpoint configured to capture new, growing mentions. It filters for recently recorded AI responses with variations of your brand name (partial_match) and high AI search volume of related queries.
[
{
"language_name": "English",
"location_code": 2840,
"platform": "google",
"target": [
{
"keyword": "notion",
"search_scope": [
"answer"
],
"match_type": "partial_match"
}
],
"filters": [
[
"ai_search_volume",
">=",
"500"
],
"and",
[
"first_response_at",
">=",
"2025-06-01 00:00:00 +00:00"
]
],
"order_by": [
"ai_search_volume,desc"
],
"offset": 0,
"limit": 10
}
]
Scan question fields in the response to identify the top topic clusters, check sources to see which domains are giving your brand visibility, and check the answer fields to understand how your brand is described. Next, you may want to take your top recent mentions, group them by intent, and research related fan-out query opportunities to keep the ball rolling.
If your mentions are declining, focus on investigating what changed. A drop in mentions may indicate that the query landscape has changed, or that AI systems have found clearer, fresher, or more authoritative sources.
Use Top Mentioned Pages to see exact URLs that are being cited for your target topics.
POST https://api.dataforseo.com/v3/ai_optimization/llm_mentions/top_mentioned_pages/live
[
{
"language_code": "en",
"location_code": 2840,
"platform": "google",
"target": [
{
"keyword": "best project management tools",
"search_scope": ["answer"]
}
],
"links_scope": "sources",
"items_list_limit": 10,
"internal_list_limit": 10
}
]
Review the URLs, source domains, and page titles in the response, then use an AI assistant to classify them (e.g., into listicle, review, documentation, community discussion, product page). This will help you determine which type of content is gaining momentum. Perhaps, there’s been a shift toward Reddit/Quora discussions or toward more editorial content from review sites, and you need to double down on your brand presence there.
Together, these steps help you create a practical starting workflow for improving your AI visibility. You can measure trends, investigate changes, and connect visibility gains or losses to specific topics, sources, and AI search behavior patterns. And with the LLM Mentions API, you have the full flexibility to extend this analysis by combining endpoints, filters, locations, and target settings that match your reporting and optimization goals.
Final takeaway
Measuring AI visibility goes beyond simply knowing when a brand appears in AI answers. A brand may have recently acquired new mentions, lost some existing ones, or started appearing in AI answers for higher-demand queries, and all of these changes provide a valuable baseline that teams need to understand before they can plan the next steps.
Historical LLM mentions data gives marketers, SEO and PR teams, and GEO platforms a reliable foundation for exploring that vital AI visibility context.
With just three endpoints in our LLM Mentions API, you can trace changes in AI search performance over time, compare visibility before and after important events, and understand whether growth or decline in mentions is also linked to meaningful differences in AI search demand.
Make historical AI visibility accessible and measurable with our LLM Mentions API, and give your team the dashboards, reports, alerts, and optimization workflows they can act upon.
Try LLM Mentions API and start tracking historical AI visibility today.