> ## Documentation Index > Fetch the complete documentation index at: https://help.semji.com/llms.txt > Use this file to discover all available pages before exploring further. # Tools Guide: Understanding what your agents can use > Understand how the tools work and when to use them AI+ Agents are assistants integrated into the Semji editor to automate SEO and GEO tasks. In addition to ready-made agents, you can create a custom agent by combining several tools according to your goal. Expand a tool to see what it does, when to use it and a concrete example. **What it does** Explores the Google SERP for your main keyword and provides an enriched view of the competition, including notably: * Organic results (Top pages) * PAA (People Also Ask) * Ads **When to use it** * Competitive analysis (who is visible, with which angles) * Identification of the content gap (missing topics / questions) * Study of search intents * Analyze the SERP (PAA, ad messages, highlighted formats) **Use case** Agent Expert News — Objective: identify three topics to investigate within the content, run Google searches, analyze the important URLs, and then recommend updates to incorporate. **What it does** Retrieves the content of a URL in Markdown format (text + structure) to analyze a page in depth. **When to use it** * Detailed analysis of a specific competitor * Content comparison (structure, depth, angles) * Verification of E-E-A-T signals (sources, author, credibility) **Use case** Agent Editorial Assistant TOP3 SEO — Objective: analyze the Top 3 (Google search + content analysis of the URLs) to produce prioritized editorial recommendations. 👉 [Top 3 Hunter agent: align your content with what actually ranks](/en/ai-agents/agent-chasseur-top-3) **What it does** Accesses Semji SEO recommendations from the Focus Keyword analysis. The tool returns, in particular: * Focus Keyword + search volume * Current Content Score * Search intents * Recommended length (vs current length) * Associated terms (required occurrences vs. current + status) * Questions (with status) * Recommended incoming internal links (URL, title, status) * Recommended outbound internal links (URL, title, anchor, status) * Top 10 Google competitors (position, title, URL, page type + detailed outline) * Search intent of the SERP **When to use it** * Complete SEO audit (status report) * Identification of missing intents / questions * Improving the Content Score * Structured comparison with the Top 10 **Use case** Semantic cocoon agent — Objective: generate a complete silo (intent pyramid) + internal linking rules + editorial calendar, based on Semji (SEO) data + SERP + page analysis. 👉 [SEO architect Agent: Building a silo content strategy](/en/ai-agents/seo-architect-agent-building-a-silo-content-strategy) **What it does** Accesses GEO data from the ChatGPT analysis. The tool returns, in particular: * Focus prompt related to the analysis * Full ChatGPT/AI Overview response * Associated terms * Cited sources (URL, title, position) * Mentions (brands / competitors) **When to use it** * GEO audit of a piece of content * Analysis "why I am / am not cited" * Comparative analysis via cited sources * Prioritization of actions to gain citability **Use case** GEO Expert Agent — Objective: diagnose citation/position, compare against higher-ranked sources, and produce prioritized (before/after) recommendations to target top rankings. 👉 [GEO Expert Agent: Optimizing your visibility on ChatGPT and AI](/en/ai-agents/ai-agent-optimize-your-visibility-in-chatgpt-responses) **What it does** Retrieves the complete HTML code of a web page for in-depth technical analysis. Unlike the "Get web page content" tool which returns content in Markdown format, this tool provides access to the full source code, including: * Schema.org tags (JSON-LD structured data) * Technical meta tags (title, description, robots, canonical) * HTML5 semantic structure (article, section, nav, header) * Accessibility attributes (alt, lang, aria) * Tracking scripts (Google Analytics, Facebook Pixel) **When to use it** * Structured data audit for schema.org (Product, Article, FAQPage, VideoObject) * Heading hierarchy verification (H1, H2, H3) * Accessibility analysis (missing alt attributes, semantic structure) * Meta tag configuration check * Tag management plan verification (pixels, analytics scripts) **Use case** 🛒 E-commerce Schema.org Audit Agent — Goal: analyze product pages and recommend structured data optimizations to improve rich snippets. The agent examines the complete HTML and detects: * ✅ A Breadcrumb schema is properly in place * ❌ The Product schema is missing * 💡 Recommendation: add a Product schema with essential properties (name, image, description, price, availability) **What it does** Accesses GEO data from Google AI Overview analysis. This tool helps optimize your content to appear in Google's AI-generated responses, a rapidly growing strategic traffic channel. The tool returns: * The complete AI Overview response generated by Google * Identified related terms * Cited sources (URL, title, position) * Brand mentions present in the response **When to use it** * GEO audit specific to Google AI Overview * "Why am I / am I not cited" analysis in AI Overviews * Comparison with better-positioned sources * Action prioritization to improve citability in Google's generative responses **What it does** Performs a read-only semantic search across your organization's Knowledge Base and returns the most relevant excerpts to use as context for the agent (product documentation, internal references, business documents). This tool never modifies your documents: it only reads them to inform the agent's reasoning. **When to use it** * Ground the agent's answers in your own references (internal guides, talking points, product documentation) instead of relying only on its general knowledge * Ensure consistency with validated internal sources * Reduce the risk of inaccuracies on topics specific to your company or industry This tool can only return results if your organization's Knowledge Base already contains documents. If it is empty or sparsely populated, import your references before enabling this tool on an agent. **What it does** Applies the Brand Voice configured in Semji: the agent adapts its output to your brand's tone, style and vocabulary instead of generating neutral content. **When to use it** * Delegate writing to an agent while keeping a voice consistent with your other content * Avoid systematically reviewing tone and style on agent-generated content * Keep output consistent across multiple contributors or agents Creating and managing your Brand Voice (from URLs, files or text) is done separately, in your organization's configuration. **What it does** Gives the agent direct access to your editorial planning, with the ability to: * List, filter and search content * Create content from scratch, with a focus keyword set at creation * Optimize existing content by importing a URL into a draft * Change a content's workflow status (except the "Published" status) * Assign content to a team member * Set a due date * Move content into a folder * Automatically trigger an SEO analysis when content is created (with the option to skip it) This tool is intentionally non-destructive: the agent cannot delete, publish or duplicate content, nor create, rename or delete folders. It also cannot change the focus keyword of existing content. High-volume operations remain capped (draft creation and multi-edit operations are limited per run): this tool is not designed for unlimited bulk actions. **When to use it** * Let an agent keep your planning up to date as it works (creating content, updating status, assigning, setting due dates) * Automate the import and creation of drafts to optimize * Keep a clear boundary on what the agent can and cannot change in your planning **What it does** Lets the agent set its own work plan before carrying out a complex task: it breaks the task down into steps, then checks off its progress as it executes. You can see at any time where the agent stands and what is left to do. This is not time-based scheduling: this tool does not let you set up a recurring run (e.g. "every Monday") or a delayed one (e.g. "run again in two weeks"). The plan is built and tracked during the agent's execution, not before it. **When to use it** * Hand off a multi-step task (such as content localization) to the agent while keeping visibility into its progress * Track the progress of a long-running task without needing to check in on the agent ### How to combine tools You can use multiple tools together for comprehensive analyses. #### SEO combination example **Search on Google** identifies the Top 3 for your keyword, then **Get the content of a web page** retrieves detailed content from these 3 pages. The agent compares and analyzes according to your instructions. #### SEO + GEO combination example **Retrieve SEO recommendations** analyzes current content optimization, while **Retrieve ChatGPT/GEO recommendations** analyzes visibility in AI responses. The agent provides a complete SEO and GEO audit with priority recommendations. If you only need the list of competitors without analyzing their content, use only "Search on Google". For a complete optimization audit, combine "Retrieve SEO recommendations" and "Retrieve ChatGPT/GEO recommendations". This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.