User interface AI

Models

• Pick the model: Selection of the AI model for the AI search. A model is a pre-trained artificial intelligence that recognizes patterns in the data and thus optimizes the search results.

  
  The following models are available:

  • foxbase: Suitable for smaller data records with up to 1,800 characters. Your data remains with FoxBase and does not reach third-party providers.

  • openai: Supports larger data records with up to 32,000 characters. Processes the data with embedding models from Open AI.

  • amazon: Supports larger data records with up to 32,000 characters. Processes the data with the Amazon Titan Embedding models. Your data remains with FoxBase and does not reach third-party providers.

• Pick embeddings: Selection of embedded columns for the AI search. Embedded columns are the contents of the product data that are to be included in the semantic search model. These columns help the model to understand the context of the products and improve the search results.

• Pick payload columns: Optionally select payload columns if there are fields in the product data that should ONLY be included in the generative model and NOT in the general AI search. The selected columns help the model to generate relevant answers to user queries.

Hybrid search

• Keyword vs Semantic Search: When activated, the platform uses a combination of keyword search and semantic search. The position of the slider defines the percentage of the respective search model.

Generative task

• Model: Selection of the model for the generation of answers by the AI. A generative model is a pre-trained artificial intelligence that independently generates answers based on input data. AI does this by using patterns from large amounts of data. Depending on the size of the data records, we recommend choosing a suitable, small model, as smaller models are faster than the larger ones.
  
  The following models are available:

  • gpt4o-mini: Supports larger data records with up to 128,000 characters. Extremely fast, but slight loss of quality compared to gpt4o. Processes the data with large language models from Open AI in the USA, the data does not remain with FoxBase.

  • gpt4o: Supports larger data records with up to 128,000 characters. More intelligent than gpt4o-mini, but somewhat slower. Processes the data with large language models from Open AI in the USA, the data does not remain with FoxBase.

  • haiku: Supports large data records with up to 200,000 characters. Speed and quality comparable to gpt4o-mini. Your data remains with FoxBase and does not reach third-party providers.

  • sonnet: Supports large data records with up to 200,000 characters. Speed and quality comparable to gpt4o. Your data remains with FoxBase and does not reach third-party providers.

  • gemini: Supports very large data records with more than 1 million characters. This is the fastest model we offer, with very little loss of quality. However, your data will be passed on to Google and Google will use the data to train its own models. We therefore recommend using this model only with dummy data or data that is already publicly available.

• Prompt: Configuration of the prompt. A prompt is an input or instruction that helps the AI to generate a targeted response by defining the tone, type or content of the response.
  
  The following options are available:

  • default: Standard setting. The AI responds like a sales representative.

  • qa: Setting for a question and answer scenario. Selector users can ask questions and the AI answers them.

  • doc: The AI searches the contents of PDF documents. The PDF documents are converted into products via a pipeline and imported into the Workbench. If you are interested in this function, please contact CSM Support.

  • custom: User-defined setting. Add instructions on how the AI should handle search queries. Do you need more complex custom prompts but are not familiar with prompt engineering? Contact your CSM support and we will help you.

💡The prompt has no influence on the search result. It only influences how the AI reacts to the request.

• Context Size: Limits how many products are used for the AI generative task. This is helpful if many products are to be displayed or recommended, but the AI cannot process them all at the same time. The AI therefore only uses the top N products to provide precise answers and recommendations.

Search settings

• Search Similarity Threshold: Determines how similar the results should be to the query. The closer the threshold is to the value 1, the more the result corresponds to the request. A lower value, on the other hand, also provides results that are less accurate but still relevant.

• Reranker: Improves the search results via the semantic search in two steps with subsequent sorting of the products based on their similarity to the query.

Feedback settings

When activated, users can provide feedback on the recommendations. The AI learns which recommendations users prefer depending on the search query.

💡By giving feedback, the platform sends a value to the database. Recommendations with higher values appear higher up on the results page for a subsequent equivalent search query.

• Accept feeedback: Activation of feedback in staging.

• Consider Feedback in Search: Platform accepts feedback for future search queries.

• Feedback Similarity Threshold: Setting the percentage match between feedback and search result.

• Reset feedback: Delete feedback from future search queries.