# AI Recommendation Engine AI Recommendation Engine: Smart, Personalized Guidance, no more blind guessing. This feature combines natural language processing (NLP) with machine learning to understand a user’s needs, level, and preferences through conversation. Unlike traditional static, menu-based recommendations, the AI engine adapts dynamically—it builds on each response, asking follow-up questions and refining suggestions in real time. The result is guidance that feels personal, relevant, and practical. This interactive approach not only enhances the user experience but also significantly boosts conversion rates. #### On the homepage, select *AI Recommendation Engine* from the application menu ⬇️
\ Once inside, go to *Basic Settings* ⬇️ * Keywords to enter/exit the AI Recommendation Engine * Message displayed upon entry into the AI Recommendation Engine * Content displayed upon exit from the AI Recommendation Engine
#### Next, select language model and the language {% hint style="info" %} Model selection usually represents different versions, capabilities, and application scenarios. The choice of model can directly impact the quality of generated content, its language comprehension, creativity, and computational efficiency. {% endhint %}

After completing the settings, move on to *Item Management*—found under the *Database* section in the side toolbar ⬇️
First, add an item
Type in the item name, item URL, item description, and upload an image
{% hint style="info" %} Purpose of Adding Category Tags The goal of setting up category tags is to improve both the accuracy and efficiency of user preference analysis. With these tags in place, the system can analyze behaviors and identify preferences based on their attributes, enabling more precise surveys and targeted communication for different user groups {% endhint %} {% hint style="info" %} Content Hashtag \ This feature automatically creates relevant hashtags based on product information, helping users quickly generate keywords when they’re unsure which tags to use. The result boosts both content visibility and searchability. By clicking the “Generate Content Hashtag” button, the system intelligently analyzes product attributes and suggests recommended tags. If the auto-generated hashtags don’t fully meet your needs, you can manually edit and adjust them to better match your brand voice and marketing goals. {% endhint %}
\ Next, we go to “generate content hashtag”
When a user selects “#Generate Content Hashtag” within an item and clicks “Confirm,” the system automatically generates related hashtags based on the product information. These auto-generated tags are displayed in *gray* to distinguish them from manually edited tags.
In Item Management, there’s also a “#Generate Content Hashtag” button at the top right. When clicked, the system automatically generates relevant hashtags based on *all* product information, applying them across every existing item in the database.
#### Alternative Method - Excel Instead of adding the items manually, you can import them from excel. 1. Click export
2. An excel file will show up in your download. Open the excel file and list your items.
3. After you've finished adding your items, import the excel sheet.
5. The item will look like this:
#### Tag Feature Explained In MantaGO, tags are more than just categories: they help the system “understand” your products. Below are the definitions and use cases of the three types of tags. {% hint style="info" %} The purpose of creating category tags\ Category tags are designed to improve the accuracy and efficiency of user preference analysis.\ \ By assigning category tags, the system can analyze behavior and identify preferences based on tag attributes. This helps conduct more targeted surveys and communication with different user groups in the future. {% endhint %} > #### Item Tags: Basic Categorization > > **Definition:** The most straightforward way to categorize products, mainly used for backend management or basic filtering. > > **Purpose:** To distinguish major product categories. > > **Examples:** Tops, Accessories, Summer Collection, Best Sellers . > > **Recommendation:** This is the most basic setup. It is recommended that each product has at least 1–2 item tags to make future management easier. > #### Hashtags: Social Media & Search Use > > **Definition:** Similar to tags used on social media platforms (Instagram/Threads), used to capture trends or specific keywords. > > **Purpose:** To increase search relevance and make recommendations feel more social and discoverable. > > **Examples:** #OOTD, #MinimalStyle, #PerfectGift, #BudgetFriendly. > > **Recommendation:** You can use more conversational or expressive words to attract consumers’ attention. > #### Auto Tags: AI-Powered Recommendations > > **Definition:** This is the core of the system’s intelligent analysis. AI extracts product characteristics and uses them to match consumer preferences more precisely. > > **Purpose:** These tags affect the accuracy of system recommendations. The system uses them to push products to potential customers who are interested in these characteristics. > > **Examples:** Minimalist Style, Moisture-Wicking, Pastel Colors. > > **Recommendation:** If you’re unsure what to write, we strongly recommend clicking the **“Generate Content Hashtag”** button so AI can identify the key characteristics of your product. #### Step 2: Preference Survey ⬇️ During the preference survey, administrators can design guided questions—such as *“Question 1,” “Question 2,” “Question 3”*—based on product attributes and usage scenarios. These questions help consumers quickly narrow down suitable items according to their needs and preferences. Under each question, you can add and configure multiple answer choices through the “Add Option” feature, allowing the system to recommend corresponding items automatically based on user responses.

⬆️ If administrators have not yet defined specific survey questions, the system also provides an “AI-Generated Questions and Answers” feature. Using semantic analysis and product attribute recognition, it automatically creates guided questions along with corresponding options—helping teams quickly build an effective, structured survey flow.
After the system generates product recommendations, administrators can further refine how each item is displayed. This can be done by clicking “Edit” on the “View Item” page.
The system will then show the available product details—such as summary, price, discount price, size, or brand—in the “Recommendation Preview” section. Any fields left blank will not appear in the preview; only the information you’ve entered will be displayed. Finally, go to “Recommendation Preview Settings” and check the fields you want to display. Once selected, your recommendation display setup is complete.
--- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://blog.mantago.cc/en/applications/ai-recommendation-engine.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.