Skip to main content
The questions below refer to Conductor’s AI Search Performance features. You can learn about these features in the following articles:

​
AI Search Concepts

A “brand” refers to the presence and positioning of your organization’s identity—or of its products and services—across the AI search landscape. This might include how the brand is cited, mentioned, and positioned across various topics and for various personas.When you add a brand to track in Conductor’s AI Search Performance reporting, you are telling Conductor how to know what to look for and how to identify your presence in AI results. Conductor will not only look for mentions of the brand, but also identify brand products inferred from the brands you’ve tracked. For example, if you track “Nike” as a brand, we’ll also count references to “Air 3” as a related brand mention.The AI Search Performance report provides a view into the presence of each brand you track by monitoring mentions and citations over time, analyzing market share at the topic level, and providing insights into how the brand is surfaced and trusted by AI search systems.
A “topic” refers to a thematic area around which AI search engines organize information. Conductor’s AI Search Performance reporting includes market share analysis based on mentions and citations at the topic level, and further exploration into the prompts that represent each topic.When you add a topic to track in Conductor’s AI Search Performance reporting, you are telling Conductor how to classify the semantically relevant content that it should consider as being related under a single topical area.Conductor designed AI Search Performance to help organizations position their brands at the center of how AI understands and presents information for the topics that matter to your business. This lets you quantify your share of voice for your topics—and to influence how your brands are surfaced, cited, and trusted by AI search systems.
A “prompt” refers to the terms or phrases that users enter into AI search engines. Each prompt you track belongs to a tracked “topic”, and reflects how a user might signal their interest in that topic. Conductor analyzes your tracked prompts to understand how your brand performs for topics in terms of visibility in AI search results.When you track a topic in Conductor’s AI Search Performance reporting, Conductor’s AI engine automatically creates a set of related prompts.The data gathered for each of those prompts appears in AI Search Performance, and provides you with context about the kinds of ways your customers might be searching for information about your topic. You can use insights to track your brands’ performance and discover opportunities for optimization.Conductor will generate up to 100 prompts for a topic. Contact Conductor if you need more than 100 prompts for a topic.
A “persona” refers to an example of the type of person who might be searching for one of your topics.Personas enable us to generate popular prompts for your topics and categorize them. You will then be able to monitor your tracked prompts across those personas, which will enable you to better understand your visibility in terms of the personas most valuable to your business.Don’t already have personas defined at your organization? Start by identifying the key characteristics, goals, and challenges of your target audience based on research and data. Categorize these personas according to their relevance to your business goals, which allows you to generate targeted prompts and monitor visibility effectively.
A “mention” is when Conductor identifies one of your brands as appearing in the text of an AI search result.

​
How does Conductor handle brand mentions if I have multiple domains with different brand names?

Brand configurations, including brand names and associated domains, are managed at the web property level in Conductor’s settings. This allows you to define specific brands for each web property you are tracking in AI Search.For example, if you have two different domains with distinct brand identities, you would set them up as separate web properties with their respective brand names configured. Conductor will then track mentions for those specific configured brands in relation to that web property.

​
How are multiple mentions of the same brand counted for a single AI response?

If a specific brand is mentioned multiple times in a single AI response for a prompt, it is counted as one unique brand mention for that prompt. In other words, Conductor looks at the AI response and asks “Does this brand appear at all in this response?” If the answer is “Yes” then AI Search Performance considers it a mention.However, note that if the same response includes multiple citations of your website, each distinct citation is counted.
A “citation” is when Conductor identifies a link to your website among the sources an AI search engine used to create an AI search result.

​
How are multiple citations of the same website counted for a single AI response?

If a specific website is cited multiple times in a single AI response for a prompt, each link is counted  as a citation for that prompt.However, note that if the same response includes multiple mentions of one of your brands, those mentions will be counted as one, unique mention.
A “website” includes all the web properties you have included in a configured web property group. This means that your citations data will include any property you track with Conductor, including social channels or other off-site content.
“Search Intent” is a classification Conductor gives to prompts in AI Search Performance. This classification describes the underlying purpose or goal of the person performing a search with the prompt. Conductor evaluates each prompt with our AI engine to determine one of the following classifications for search intent:
  • Brand/Service Navigation: User seeks a specific brand, site, or known service. Not all branded prompts are of this type.
  • Comparison: User directly compares alternatives.
  • Education: User seeks to learn or understand.
  • Other: The prompt doesn’t fit neatly into the other categories.
  • Pricing: User seeks cost information.
  • Purchase: User shows intent to buy or take conversion action.
  • Recommendations: User seeks suggestions or options (not direct comparison).
  • Support: User needs help or customer service.
Sentiment score reflects the overall negative, neutral, and positive tone of AI search responses that mention the brands you track. Scores appear between 1 (indicating mostly-negative sentiment) and 10 (indicating mostly-positive sentiment).

​
How is sentiment score calculated?

AI Search Performance calculates sentiment score using a weighted average formula where each of a prompt’s mentions receives a numeric value between 1 and 10, based on whether they are positive, neutral, or negative. Those values are aggregated and then divided by the total number of mentions (for a single prompt or across all the prompts in a topic) to determine the sentiment score.This produces a score between 1 (mostly negative) and 10 (mostly positive) reflecting the overall tone of AI search responses mentioning your brand or competitors.
Conductor’s AI engine assigns common categories for statements that drive positive and negative sentiment in AI search responses. For example, for an online retailer, sentiment categories might include “pricing”, “quality”, or “support”, since these are common areas where content includes a positive or negative sentiment.The possible sentiment categories that Conductor might apply:
  • Health. Example sentiment: “made with nutritious and junk-free ingredients”, “improved skin texture”
  • Quality. Example sentiment: “great sound”
  • Price. Example sentiment: “expensive”, “budget-friendly”
  • User-friendliness. Example sentiment: “Easy to use”, “Confusing Labeling”
  • Popularity. Example sentiment: “two popular choices are…”
  • Ethics. Example sentiment: “cruelty-free”, “more environmentally friendly”
  • Service. Example sentiment: “email support”, “personalized assistance”
  • Experience. Example sentiment: “tasty and organic”, “smooth on the skin”
  • Misc. Catch-all for uncategorized or other miscellaneous sentiments

​
AI Search Data

Conductor collects new data for the AI Search Performance report on a rolling basis. Once collected, it is processed in our data platform, after which the data will appear in the report. This can take as few as five hours, but may take up to a few days.As a result, if you track new topics and prompts toward the end of the week, you may not see data for them by Monday. However, data will appear in reports as soon as it has been collected.
We use official APIs provided by the AI search engines we support to get the content from responses, which provides more accurate, scalable, and richer data (like location-specific results).
AI search engines are still relatively new, and our industry’s understanding of how companies like Google, Microsoft, OpenAI, and Perplexity are crawling and creating AI responses is relatively limited. However, Conductor and its Community of marketing professionals can help guide you as we learn more together.Want to find out how to use Conductor’s products to help understand the AI search landscape and how your site is crawled? Refer to the Optimizing Your Site & Content for AI-Powered Search in the Conductor Community to learn more.
Yes, the AI Search Performance report allows you to see the other brands and sites ranking for your topics and to compare your brands and your domain with your competition.
Because there is currently no definitive or reliable source for AI search volume data—despite other contentions in the market—Conductor does not provide an “AI MSV” metric:
  • Any figures provided by other tools are typically based on sampling or proxy data.
  • Measuring prompt-level search volume may not be meaningful due to their highly specific and personalized nature.
  • A widely accepted method for measuring topic-level volume has not yet been established.
You can learn more about Conductor’s position on AI volumes on our blog.
For “share of model”, you can view share of voice data in the AI Search Performance report to see your market share for your brand and website or your competitors’ brands and websites in each of the AI engines on which you can track topics with Conductor.
  • If you track the same number of topics and prompts across these models, you can compare your presence in each model individually.
  • However, Conductor does not currently provide a single “share of model” metric that aggregates across all models.
Coverage asks, “How often do my brands appear in the prompts we track for a topic?” Share of voice goes further to ask, “Of all the brands mentioned or websites cited in the prompts we track, how often do ours get mentioned or cited?”

​
A note on share of voice for brand mentions

Your share of voice reflects your share of all brand mentions—so if a prompt mentions multiple brands, each is counted, making the total denominator much higher than the number of prompts you are reporting on. Conversely, your percentage of prompts with mentions uses prompts as the denominator. As a result, you’ll find that your share of voice will tend to be much lower than your percentage of prompts with mentions, because it compares your mentions against the full field of all brands mentioned, not just against the number of prompts.

​
Note!

Because your brand share of voice accounts for all the brands you track, any changes you make to add or remove brand names will change the composition of the pie chart.
Conductor does not currently offer insights specifically related to social media engagement related to AI Search performance. However, by analyzing the websites in citations and the AI responses provided in Conductor’s reporting, you can discover social media insights relevant to AI search.
To generate these intent and persona insights, your tracked prompts will need to be returning mentions. If you have no mentions, you won’t see this data.Likewise, if your tracked prompts don’t earn mentions, there will be no signal from which Conductor can determine sentiment.

​
No AI response generated

Conductor successfully submits a tracked prompt to an AI search engine, but no AI response is returned. This is most common for Google AI Overviews, where AI responses are generated only for 50–70% of prompts.

​
No AI response collected

Conductor could not successfully collect a response from the AI search engine for the tracked prompt.

​
AI Search Configuration

Both Standard and Admin user types can add or remove topics and prompts in AI Search Performance reporting.
Conductor lets you choose the tracking cadence per topic so you can match each topic’s importance and pace:
  • Daily — for high-priority or fast-moving areas.
  • Weekly — for active strategic monitoring.
  • Monthly — for broader, lower-frequency coverage.
Topics tracked monthly are reported at monthly granularity only (no daily or weekly views for those topics). The first collection runs within approximately one day of configuration; after that, scheduled collections run at the beginning of each month. Monthly is a good fit for broader strategic areas you want to keep in coverage without over-prioritizing them in your day-to-day workflows.Note that Daily tracking will substantially increase your credit usage for the prompts you track at a daily frequency.

​
How does Conductor determine the topics it generates?

When you track a topic in Conductor’s AI Search Performance, the AI engine automatically generates a set of up to 100 prompts that represent the topic. These prompts are designed to reflect how users might signal their interest in that topic in AI search engines. The data collected from these prompts helps to understand how your brand performs in terms of visibility in AI search results, providing insights into user intents and behaviors. Here’s how our process gets you the best results:

​
1. Domain Context Gathering

We collect essential information about your business environment. Using our proprietary web intelligence data, we gather signals like your organization’s description, your industry, recent news, and important themes in your market to ensure that generated topics align with your real-world business context.

​
2. Keyword Analysis (If Applicable)

If you are already tracking keywords in Conductor Intelligence, we analyze your keywords with the highest search volumes to understand your current SEO strategy, and identify which themes are generating the most search demand.

​
3. Content Analysis (If Applicable)

If Conductor Intelligence has crawled your website content, we analyze your pages using semantic clustering to understand which topics your site already covers well and where your content authority exists.

​
4. Topic Generation

We combine domain context, keyword analysis (when available), and content analysis (when available) to generate relevant topic suggestions. These suggestions are designed to help you defend and expand your organic search reach while focusing on topics that matter to your audience.

​
How does Conductor determine that the prompts it generates are the best prompts to represent the topic?

Your prompts are generated based on your keyword tracking, and consider factors like search intent and your configured personas. This provides a holistic approach for understanding what your organization cares about and the kinds of queries your customers might pose to an AI search engine. The process involves a few key steps:

​
Identifying Topics

  • With Tracked Keywords: The system analyzes your top 1,000 keywords with the highest search volume to suggest relevant topics.
  • Without Tracked Keywords: The system uses your website domain to understand your business and recommend topics based on your content.

​
Generating Prompts

For each topic, the system generates an initial set of 100 prompts.It uses a “natural distribution” approach, using the AI’s training on billions of data points to create a diverse set of questions that reflect how people actually ask questions. The goal is to provide a realistic and broad sample of queries, rather than forcing a rigid structure.

​
Classifying Personas and Intent

After the prompts are generated, the AI analyzes each prompt and assigns personas and a search intent to them.This approach ensures the prompts are not artificially constrained, allowing for a more authentic and natural distribution of questions. This is why the number of prompts per persona or intent may vary—it reflects real-world search patterns.

​
Branded vs. Unbranded Prompts

By default, the system generates a mix of branded and unbranded prompts based on a natural distribution.You can use a toggle to exclude branded queries, which instructs the AI to avoid creating prompts that mention your brand or your competitors’ names.

​
Custom Instructions

You can also provide free-form custom instructions at either the topic level or the prompt level to guide how the generator works. Instructions are plain-English and layered on top of the default generation logic — they don’t replace your keyword data, content analysis, or persona signals. Four common instruction patterns:Establish hard constraints: Words to require or avoid, audience locks (B2B vs. B2C), or fixed splits across intents.Shape the purpose behind prompts: Bias the set toward a specific stance, like “problem-aware” topics or “how-to” prompts.Set context:
  • Provide one to two example topics or prompts in the format you want and the generator will mirror the pattern.
  • Anchor generation to a season, event, or time horizon.
If you leave Custom Instructions blank, Conductor generates suggestions using the default distribution described above.

​
Need more prompts?

If you want more prompts generated for more topic coverage or fewer to account for AI Response Credit budgeting, you can contact your Conductor representative to learn more about your options.
Clear, specific, and in natural language. A few patterns that tend to work well:
  • State the rule as a constraint, not a preference. “Exclude any B2C or retail-oriented topics” beats “we prefer B2B topics.”
  • Provide 1–2 concrete examples of what “good” looks like. “Follow this pattern: [Technology] + [Industry] + [Outcome]” gives the generator a target.
  • Anchor timeframes explicitly. “Emerging trends for the next 6 months” is more useful than “recent topics.”
  • Keep each instruction to one idea. If you have several rules, list them on separate lines.
If the output doesn’t match what you expected, tighten the instruction and regenerate; this flow is designed to be iterative.
Brand configuration, including brand names and associated domains, are managed at the web property level in Conductor’s settings. This allows you to define specific brands for each web property you are tracking in AI Search.For instance, if you have two different domains with distinct brand identities, you would set them up as separate web properties with their respective brand names configured for AI Search Performance.Conductor will then track mentions for those specific configured brands in relation to that web property.
When tracking main brands and sub-brands (like “Nike” vs. “Air Max”), you can differentiate these with your topic configuration. When you add a topic, you can select the brands you want included for the particular topic or set of topics. So, if you know you want to track a topic where one of your sub-brands (like “Air Max”) would likely be relevant to that specific topic or group of topics, you could add that as a brand in the topic configuration process.
Currently, Conductor supports the following AI search engines for AI Search Performance:
  • ChatGPT (Auto). This reflects the default experience when using OpenAI’s consumer ChatGPT product, which may include live web search results.
  • ChatGPT (Search). This reflects the experience when using OpenAI’s consumer ChatGPT product and specifically choosing to use the Web Search mode.
  • Claude Sonnet. This is the model Claude uses by default.
  • Gemini. Note that, depending on the prompt, Gemini strategically uses live search only when it determines that external, current information is necessary to provide an accurate and relevant answer.
  • Google AI Mode. This refers to queries made in Google’s AI Mode.
  • Google AIO. This refers to Google’s AI Overview results.
  • Grok (Auto)
  • Microsoft Copilot
  • Perplexity

​
How does Conductor decide which AI search engines to track?

Conductor evaluates which AI search engines to track based on their overall popular usage and on our customers’ feedback. If there are specific engines that you think we should add to the platform, please let us know: Conductor Support

​
Country-Level Locations

Conductor supports the country-level locations and languages that are available to gather data from the APIs provided by our supported AI search engines. If the search engine provides data for a particular country through their API, then you’ll be able to see data for that location in Conductor. The list of current supported locations is listed below.

​
More-Granular Locations

Specific location tracking is available for prompts tracked on all supported engines except Google AI Mode. Available cities and states reflect the same list Conductor Intelligence uses for keyword tracking. Just select the location and language and you’ll see a link to add more-specific locations:

​
Not seeing a location for which you want to track prompts?

Note that if you do not see a particular location for which you want to track results, you can add the location with the topic you want to track. This will provide our AI engine with some contextual guidance related to the location and help return prompts that are relevant to your location of interest.If there are locations that are not supported that you think should be included, contact Support to request the location be added.
Supported Location (Language)Exceptions
Afghanistan (English)
Afghanistan (Pashto)
Albania (Albanian)
Albania (English)
Algeria (English)
Algeria (French)
Andorra (Catalan)
Angola (English)
Angola (Portuguese)
Antigua and Barbuda (English)
Argentina (English)
Argentina (Spanish)
Armenia (Armenian)
Armenia (English)
Armenia (Russian)
Australia (English)
Austria (English)
Austria (German)
Azerbaijan (Azerbaijani)
Azerbaijan (English)
Azerbaijan (Russian)
Bahamas (English)
Bahrain (Arabic)
Bahrain (English)
Bangladesh (Bengali)
Bangladesh (English)
Belgium (Dutch)
Belgium (English)
Belgium (Flemish)
Belgium (French)
Belgium (German)
Belize (English)
Belize (Spanish)
Benin (French)
Bhutan (Dzongkha)
Bolivia (English)
Bolivia (Spanish)
Bosnia and Herzegovina (Bosnian)
Bosnia and Herzegovina (English)
Botswana (Tswana)
Brazil (English)
Brazil (Portuguese)
Brunei (Bahasa Melayu)
Bulgaria (Bulgarian)
Bulgaria (English)
Burkina Faso (English)
Burkina Faso (French)
Cambodia (English)
Cambodia (Khmer)
Cameroon (English)
Cameroon (French)
Canada (English)
Canada (French)
Cape Verde (Portuguese)
Central African Republic (English)
Central African Republic (Sango)
Chad (Arabic)
Chad (French)
Chile (English)
Chile (Spanish)
China (Chinese (Simplified))Not Supported
ChatGPT (Auto), Gemini, Microsoft Copilot, Perplexity
Colombia (English)
Colombia (Spanish)
Costa Rica (English)
Costa Rica (Spanish)
Cote d’Ivoire (English)
Cote d’Ivoire (French)
Croatia (Croatian)
Croatia (English)
Curacao (English)Not Supported
ChatGPT (Auto)
Cyprus (English)
Cyprus (Greek)
Czech Republic (Czech)
Czech Republic (English)
Czech Republic (Russian)
Democratic Republic of the Congo (English)
Democratic Republic of the Congo (French)
Denmark (Danish)
Denmark (English)
Djibouti (Arabic)
Djibouti (French)
Dominican Republic (English)
Dominican Republic (Spanish)
Ecuador (English)
Ecuador (Spanish)
Egypt (Arabic)
Egypt (English)
El Salvador (English)
El Salvador (Spanish)
Estonia (English)
Estonia (Estonian)
Estonia (Russian)
Ethiopia (Amharic)
Ethiopia (English)
Finland (English)
Finland (Finnish)
France (English)
France (French)
Gabon (French)
Georgia (English)
Georgia (Georgian)
Germany (English)
Germany (German)
Ghana (English)
Greece (English)
Greece (Greek)
Guam (English)Not Supported
ChatGPT (Auto)
Guatemala (English)
Guatemala (Spanish)
Guyana (English)
Haiti (French)
Honduras (English)
Honduras (Spanish)
Hong Kong (English)Not Supported
ChatGPT (Auto), Microsoft Copilot, Perplexity
Hong Kong (Han (Traditional))Not Supported
ChatGPT (Auto), Microsoft Copilot, Perplexity
Hungary (English)
Hungary (Hungarian)
Iceland (Icelandic)
India (English)
India (Hindi)
India (Kannada)
India (Malayalam)
India (Marathi)
India (Tamil)
India (Telugu)
Indonesia (English)
Indonesia (Indonesian)
Iraq (Arabic)
Iraq (English)
Ireland (English)
Israel (English)
Israel (Hebrew)
Italy (English)
Italy (Italian)
Jamaica (English)
Japan (English)
Japan (Japanese)
Jordan (Arabic)
Jordan (English)
Jordan (Hebrew)
Kazakhstan (English)
Kazakhstan (Kazakh)
Kazakhstan (Russian)
Kenya (English)
Kenya (Swahili)
Kuwait (Arabic)
Kuwait (English)
Kyrgyzstan (English)
Kyrgyzstan (Kyrgyz)
Kyrgyzstan (Russian)
Laos (Lao)
Latvia (English)
Latvia (Latvian)
Latvia (Lithuanian)
Latvia (Russian)
Lebanon (Arabic)
Lebanon (English)
Lebanon (French)
Libya (Arabic)
Libya (English)
Liechtenstein (German)
Lithuania (English)
Lithuania (Lithuanian)
Lithuania (Russian)
Luxembourg (English)
Luxembourg (French)
Luxembourg (German)
Madagascar (English)
Madagascar (Malagasy)
Malawi (Chewa)
Malawi (English)
Malaysia (Bahasa Melayu)
Malaysia (Chinese (Simplified))
Malaysia (English)
Maldives (Dhivehi)
Maldives (English)
Mali (English)
Mali (French)
Malta (English)
Malta (Maltese)
Mauritius (English)
Mexico (English)
Mexico (Spanish)
Moldova (English)
Moldova (Romanian)
Moldova (Russian)
Monaco (English)
Mongolia (English)
Mongolia (Mongolian)
Montenegro (Montenegrin)
Morocco (Arabic)
Morocco (English)
Morocco (French)
Mozambique (English)
Mozambique (Portuguese)
Myanmar (Burmese)
Myanmar (English)
Namibia (English)
Nepal (English)
Nepal (Nepali)
Netherlands (Dutch)
Netherlands (English)
New Zealand (English)
Nicaragua (Spanish)
Niger (English)
Niger (French)
Nigeria (English)
North Macedonia (English)
North Macedonia (Macedonian)
Northern Mariana Islands (English)Not Supported
ChatGPT (Auto)
Norway (English)
Norway (Norwegian)
Oman (Arabic)
Pakistan (English)
Pakistan (Urdu)
Palestine (Arabic)
Palestine (English)
Panama (Spanish)
Paraguay (Spanish)
Peru (English)
Peru (Spanish)
Philippines (English)
Philippines (Filipino)
Poland (English)
Poland (Polish)
Portugal (English)
Portugal (Portuguese)
Puerto Rico (Spanish)Not Supported
ChatGPT (Auto)
Qatar (Arabic)
Qatar (English)
Romania (English)
Romania (Romanian)
Rwanda (English)
Rwanda (Kinyarwanda)
Saudi Arabia (Arabic)
Saudi Arabia (English)
Senegal (English)
Senegal (French)
Serbia (English)
Serbia (Serbian)
Sierra Leone (English)
Singapore (Chinese (Simplified))
Singapore (English)
Slovakia (English)
Slovakia (Slovak)
Slovenia (English)
Slovenia (Slovenian)
Somalia (Somali)
South Africa (English)
South Korea (English)
South Korea (Korean)
Spain (English)
Spain (Spanish)
Sri Lanka (English)
Suriname (Dutch)
Sweden (English)
Sweden (Swedish)
Switzerland (English)
Switzerland (French)
Switzerland (German)
Switzerland (Italian)
Taiwan (English)
Taiwan (Han (Traditional))
Tajikistan (English)
Tajikistan (Tajiki)
Tanzania (English)
Tanzania (Swahili)
Thailand (English)
Thailand (Thai)
The Gambia (English)
The Gambia (French)
Togo (French)
Trinidad and Tobago (English)
Tunisia (Arabic)
Tunisia (English)
Tunisia (French)
Turkey (English)
Turkey (Turkish)
Turkmenistan (Russian)
Uganda (English)
Uganda (Swahili)
Ukraine (English)
Ukraine (Russian)
Ukraine (Ukrainian)
United Arab Emirates (Arabic)
United Arab Emirates (English)
United Kingdom (English)
United States (Spanish)
Uruguay (English)
Uruguay (Spanish)
Uzbekistan (English)
Uzbekistan (Russian)
Uzbekistan (Uzbek)
Venezuela (English)
Venezuela (Spanish)
Vietnam (English)
Vietnam (Vietnamese)
Zambia (English)
Zimbabwe (English)
Prompts by themselves can be highly varied and personalized. Generating prompts by topic helps organize this data, allowing you to focus on broader trends.In the Topics tab of the AI Search Performance report, you can filter and select topics to see all associated prompts that have brand mentions or citations.
You can now include or exclude specific prompts in two ways:
  • Custom Instructions (during topic and prompt generation): Use hard constraints like “Exclude competitor names X and Y” or “Always include the word ‘organic’” to shape what the generator produces before you ever review the list.
  • Manual edit (after generation): In the Final Review step, download the prompt list as an XLSX, edit rows directly, and upload it back into Conductor.
For fully custom lists that bypass generation entirely, you can send prompts to your Conductor representative to load on your behalf.
As you create and track prompts for your topics, you can include a minimum of 5 prompts and a maximum of 500 prompts for a given topic.
Yes, Conductor supports tracking citations at a subfolder level in addition to the root domain or subdomain. You can configure different web properties in Conductor to map your brand or domain specifically to a subfolder for tracking purposes.
Topics and prompt groups both organize your prompts, but they do different jobs:
  • Topics organize prompts by theme—they are semantic and thematic.
  • Prompt groups are more flexible. A group can be broader or narrower than a topic, and it can cut across topics. Think of a group as a custom lens that surfaces a specific quality in your data—who created a prompt, the persona or location it targets, the engine you track it on, or the campaign it supports.
Learn how to set up prompt groups

​
AI Response Credits and Credit Usage

Access to data and tracking in the AI Search Performance report is based on a credit system. If your Conductor subscription includes AI Response Credits, your credit usage is determined by the number and types of prompts you track. Each time Conductor prompts an AI search engine for a response, you’ll consume the corresponding amount of credits.The amount of credits you use up for each prompt depends on the AI search engine you are retrieving an AI response from:
AI Search EngineCredit Usage per Prompt
ChatGPT (Auto)1.00 credits
ChatGPT (Search)2.00 credits
Claude Sonnet0.50 credits
Gemini2.00 credits
Google AI Mode0.01 credits
Google AIO0.01 credits
Grok (Auto)1.00 credits
Microsoft Copilot0.20 credits
Perplexity0.20 credits
To get the best and most reliable data for AI responses, Conductor uses API-based methods to pull data directly from search engines, where available. Each search engine has a different cost associated with pulling data through their APIs.We give each search engine their own per-prompt value so that you can decide the tracking strategy that makes the most sense from a data perspective and a budget perspective—so you can make the best strategic decisions for your business.
In the Topics and Prompts activity in the Settings menu, you can use the AI Topic configuration wizard to see how many prompts you can track for your topics based on your configuration. The wizard will indicate the amount of time you have remaining until your usage will exceed your credit budget with your selected configuration.However, for quick estimation purposes, you can use the following calculator:

​
AI Response Credits Estimator

Enter your estimated number of prompts for each AI search engine to calculate your credit usage per week and per year.This calculator reflects weekly tracking for prompts.
No, Conductor will prevent you from adding tracked topics and prompts if you have insufficient credits left in your account.