What is the purpose?
This view enables a lightweight exploration based on a set of generated prompts that cover all key attributes per analysis and common user intentions.
Each prompt simulates real user queries—without mentioning any brands—and is sent to LLMs to identify the online sources they draw upon when forming their answers.
By aggregating these sources, the page reveals the most influential websites and platforms contributing to the LLMs’ understanding of the category landscape.
To ensure diversity and realistic context, prompts are generated across 3 user intentions types:
Information Request: educational or factual prompts emphasizing expert, authoritative sources.
Media & Social Retrieval: prompts based on news, opinion, and social media, capturing public discourse.
Product Purchase: queries reflecting buying intent, surfacing retail and commercial sources.
This approach ensures the analysis reflects a full range of user behaviors—from informational and social interactions to transactional searches.
How the Influence Score Works?
Each source receives an Influence Score, calculated from how frequently it appears and how prominently it is positioned across AI-generated responses.
The formula combines 2 key dimensions:
Frequency (75%) – how often a domain appears in responses.
Position (25%) – how high that domain ranks in the list.
This weighting ensures that repetition—the most consistent signal of visibility—matters most, while still rewarding sources that appear early in AI responses.
Why the Position Matters More at the Top?
We know that people look more at the top of any list. So, the formula uses a special curve that gives extra credit to websites that appear higher, especially in the first few spots—just like how Google search works.
Why This Works Well?
More frequent = more visible and likely more trusted.
Higher position = more attention from users.
Combining both gives a good picture of which websites users are likely to see and trust most in AI results.
Score Interpretation:
High scores (50-100): Domain is frequently cited and appears in prominent positions
Medium scores (20-50): Domain has moderate citation frequency and/or positioning
Lower scores (below 20): Domain has limited presence or appears in less visible positions
The scoring system reflects real user behaviour—sources mentioned first carry significantly more weight than those listed later.
Interface Components:
Top Influential Sources
Visual bar chart showing the three most influential domains with their respective scores, providing immediate insight into market leaders.
Influential Sources Types
Pie chart breakdown showing the distribution of source types (Specialist Blogs, News Media, Retail/Commerce, etc.), helping understand the media landscape for your category.
Influential Sources Changes
Trend visualization showing how source influence has evolved over time, enabling tracking of shifting AI preferences and emerging influential domains.
Detailed Source Table
Comprehensive view showing:
Rank: Position in influence hierarchy
Domain: Source website
Type: Category classification (News Media, Specialist Blog, Branded website, Retail/E-commerce, etc.)
Top Cited URLs: Most frequently referenced pages
Influence Score: Average visibility metric
Search GPT & Gemini: Platform-specific performance indicators
Additionally, in the Sources module, the analysis is not centered on brand mentions. In fact, the brand and competitor analysis parameters are not used at all in this section.
Instead, the data collection prompts draw on:
Analysis category
Analysis attributes
Countries
This means that Sources highlights influential or frequently occurring domains and URLs based purely on categories and attributes. The process works without referencing, searching for, or relying on any specific brand names.