What data is used to generate recommendations?

We gather up to 30 tracked prompts with their citation counts, up to 30 unique citation URLs with positions, up to 25 of your own website pages (title, description, content excerpt), competitor data, and live site metadata. This forms the context for AI-generated recommendations.

How do source references work?

Each recommendation includes a source_refs array like ["P1", "C3", "O2"] linking to the prompts_index, citations_index, and own_content_index in the report. You can see exactly which data points informed each recommendation.

What categories of recommendations are generated?

Recommendations fall into four categories, content_creation (new pages to create), optimization (improvements to existing pages), authority (build authority signals and citations), and technical (infrastructure and technical improvements).

How is the AI analysis performed?

Reports use advanced AI for consistent, focused output. We provide detailed context including your domain, prompts analyzed, citations, and own content to generate actionable recommendations.

Are competitors considered in recommendations?

Yes. Your competitors' domains are included in the context. The AI considers competitive positioning when generating recommendations, helping you identify gaps where competitors are being cited instead of you.

Evidence-Based Strategy

Get evidence-based content recommendations

We analyze your tracked prompts, AI-cited URLs, and your existing content, then use GPT to generate specific, prioritized recommendations with evidence links. Each recommendation references the exact data points that informed it.

Start Free Trial

14-day free trial

Cancel anytime

Features

Data-Driven Content Strategy

Recommendations analyzes your tracked prompts, AI-cited URLs, and your existing content to generate actionable, evidence-based recommendations for improving AI visibility.

Analyze Your Prompt & Citation Data

We gather your tracked prompts (up to 30), cited URLs (up to 30), and your own website pages (up to 25) to understand your current visibility landscape.

GPT-Generated Action Items

Using GPT with your actual data as context, we generate specific, evidence-based recommendations for content creation, optimization, authority building, and technical improvements.

Evidence-Linked Recommendations

Each recommendation includes source_refs linking back to specific prompts (P1, P2...), citations (C1, C2...), and your content (O1, O2...). See exactly why each recommendation was made.

High/Medium/Low Priority

Recommendations are categorized by priority, high (significant impact), medium (valuable improvement), low (nice-to-have). Focus your efforts where they matter most.

Organized by Type

Recommendations fall into four categories, content_creation (new content), optimization (improve existing), authority (build signals), and technical (infrastructure improvements).

Specific Action Steps

Each recommendation includes concrete action_items, specific steps to implement the suggestion. No vague advice, just clear next actions.

How Recommendations Works

Analyze your data and generate evidence-based recommendations

1

Gather Data

We collect your tracked prompts (up to 30), cited URLs (up to 30), your website pages (up to 25), competitor data, and live site metadata.
2

Build Context

Data is indexed (P1, P2... for prompts, C1, C2... for citations, O1, O2... for your content) and sent to GPT with detailed instructions.
3

Generate Recommendations

GPT analyzes the context and generates prioritized recommendations across content creation, optimization, authority, and technical categories.
4

Review & Act

Each recommendation includes priority, action_items, and source_refs linking to the evidence. Focus on high-priority items first.

Recommendations Use Cases

How teams use Recommendations for data-driven strategy

Content Teams

Get specific content briefs based on citation gaps. Know exactly what topics to cover and why, with links to the evidence.

SEO & AEO Teams

See which of your pages could be optimized for better AI citation. Recommendations link to specific prompts where you're not cited.

Strategy Teams

Prioritize content investments based on data. High-priority recommendations show where you'll get the most visibility improvement.

Technical Teams

Get technical recommendations for improving how AI can access and understand your content. Structured data, schema, crawlability.

"LLM Pulse is a pioneering response to the need to measure performance across different LLMs. At TecnoCampus we need maximum transparency, the ability to parameterize results and to carry out visibility audits for our brand, now also in the increasingly established field of AI. LLM Pulse gives us all of that in a single place."

Jordi Perez Rubio

TecnoCampus

Frequently Asked Questions

Technical details about how Recommendations generates actionable insights.

: We gather up to 30 tracked prompts with their citation counts, up to 30 unique citation URLs with positions, up to 25 of your own website pages (title, description, content excerpt), competitor data, and live site metadata. This forms the context for AI-generated recommendations.
: Each recommendation includes a source_refs array like ["P1", "C3", "O2"] linking to the prompts_index, citations_index, and own_content_index in the report. You can see exactly which data points informed each recommendation.
: Recommendations fall into four categories, content_creation (new pages to create), optimization (improvements to existing pages), authority (build authority signals and citations), and technical (infrastructure and technical improvements).
: Reports use advanced AI for consistent, focused output. We provide detailed context including your domain, prompts analyzed, citations, and own content to generate actionable recommendations.
: Yes. Your competitors' domains are included in the context. The AI considers competitive positioning when generating recommendations, helping you identify gaps where competitors are being cited instead of you.

Ready to optimize your content?

Generate a Recommendations report and get specific action items to improve your AI visibility.