Understanding BAIA
The Bizzkit AI Assistant (BAIA) is a sophisticated feature within the Bizzkit platform, designed to enhance user interaction and developer possibilities through advanced AI capabilities. BAIA transcends traditional AI by acting as both an information repository and an intelligent assistant. It enables dynamic interactions through natural language processing, semantic functions, and API integrations, allowing both users and developers to efficiently manage tasks and access information within the Bizzkit ecosystem.
BAIA is powered by OpenAI's advanced model and is automatically updated to reflect the latest advancements. This ensures that BAIA delivers the best possible performance at the most cost-effective rate, keeping the experience both cutting-edge and affordable for users.
Click here for a short explanation of terms used in the article
LLM stands for Large Language Model, which is a type of AI model that can understand and generate human-like text. LLMs are trained on vast amounts of text data to learn the patterns and structures of language, enabling them to generate coherent and contextually relevant responses to user queries.
GPT stands for Generative Pre-trained Transformer. It is a type of artificial intelligence model that uses deep learning techniques to generate human-like text. The "Generative" part refers to the model's ability to create or generate text, "Pre-trained" indicates that the model has been trained on a vast amount of data before being fine-tuned for specific tasks, and "Transformer" is the neural network architecture used, which excels at understanding the context of language.
Semantic functions are AI-powered features that provide specialized interactions within an application. These functions enable context-aware actions.
API stands for Application Programming Interface, which allows different software applications to communicate with each other. APIs define the methods and data formats that applications can use to request and exchange information.
RAG Retrieval-Augmented Generation (RAG) works by retrieving relevant information from a local system before calling the AI text generation process. This ensures that BAIA accesses targeted, up-to-date data to generate more accurate and contextually relevant responses. By leveraging RAG locally, BAIA minimizes the need for frequent external API calls, reducing costs while improving performance. This system also simplifies integration for development teams, removing the complexity of building custom retrieval tools, while ensuring security, GDPR compliance, and automatic updates.
Chat interface
BAIA offers a chat interface integrated across all Bizzkit applications, providing users with a seamless and intuitive way to interact with the platform. This interface serves as a central hub for accessing documentation, performing tasks, and receiving guidance, making it easy for users to navigate the Bizzkit ecosystem.
Click here for more information on how to access BAIA in the Bizzkit applications.
Natural language interaction
BAIA is built on the latest advancements in AI technology, allowing users to interact seamlessly through natural language. This means that users can communicate with BAIA just as they would with well-known LLMs (Large Language Models) like ChatGPT, Gemini, or Bing Chat, without needing to use specific commands or technical jargon. Whether asking questions, requesting information, or giving instructions, BAIA can understand, interpret, and respond in a way that feels intuitive and conversational. This enhances the user experience by making interactions faster, more efficient, and more accessible to those without technical expertise.
Users can interact with the model in any language. Whether you're speaking Danish, Swedish, English, or another language, BAIA can understand and respond effectively, making it highly adaptable to diverse user needs worldwide. While you can ask questions in any language, BAIA will respond in English by default unless instructed otherwise.
Trained on Bizzkit documentation
BAIA possesses comprehensive and continuously updated knowledge of Bizzkit Documentation, ensuring that it can provide users with accurate and detailed answers to questions about all Bizzkit products and features. Whether users are looking for guidance on product setup, troubleshooting, or exploring advanced functionalities, BAIA is equipped to deliver relevant insights.
The model is frequently updated with the latest documentation and advancements, so users always receive the most current and relevant information. This guarantees that BAIA stays up-to-date with all the newest Bizzkit features and changes, offering an efficient and reliable experience.
Using API through the chat interface
BAIA can access the Bizzkit APIs, allowing users to interact with BAIA through natural language to query and modify data across various Bizzkit products. This makes complex API interactions much more accessible, as users can describe their needs conversationally, and BAIA translates those requests into API calls. Additionally, leveraging its advanced language model, BAIA can enrich queries by incorporating relevant context or additional data, making the process more powerful and efficient. This capability streamlines workflows, reduces manual effort, and opens up new possibilities for integrating and manipulating data effortlessly.
Privacy and security
All data interactions and conversations are securely contained within the Bizzkit cloud, ensuring full compliance with stringent data privacy and security standards. Crucially, none of the data processed by BAIA is used to train any public AI models. This means that your data remains strictly private and isolated from external systems, preventing any possibility of it being shared or exposed beyond the Bizzkit environment. BAIA operates solely within the confines of the Bizzkit ecosystem, providing users with peace of mind that their data is handled with the highest levels of confidentiality and security, and never utilized for purposes outside of their intended scope.
Additionally, users have the capability to view, resume, or delete their past interactions with BAIA. This functionality promotes accountability, ensures continuity in task execution, and gives users full control over their data within the platform.
BAIA is specifically designed to interact with Bizzkit products in a secure environment, safeguarding user data and operational integrity. Before executing any actions that may pose a risk, BAIA always seeks user approval, and its capabilities are restricted to performing only those actions for which the user has the appropriate permissions. This ensures that BAIA operates safely within the boundaries of the user’s role and access rights, without compromising security.
System recipes
BAIA comes with a collection of predefined system recipes designed to simplify and speed up the most common processes, saving users time and effort. These pre-built prompts allow users to perform tasks efficiently with just a few words, without the need to manually specify every detail. Examples of system recipes include:
- Creating a global list with items in PIM
- Creating a transformation function in DAM
- Validating synonyms in Ecommerce Search
- Listing users in specific roles within User Management
Many more recipes exist, and new ones are frequently added, ensuring that BAIA continuously evolves to support a wide range of tasks and processes.
Custom recipes
In addition to predefined options, users can easily create their own custom recipes directly within the UI, either from scratch or by building on a conversation history. This allows users to automate repetitive tasks and streamline their interactions with BAIA, ensuring that complex processes can be instantly recalled with minimal input. For those who prefer more technical flexibility, custom recipes can also be created and managed through the API.
An example of a custom recipe could be validating a product based on the user's unique attributes and validation rules, ensuring that BAIA is tailored to specific business needs. This capability empowers users to optimize their workflows, reducing manual effort while making BAIA a more powerful, personalized tool.
Favourite recipes
To help organise the recipes, users can pin their favourite recipes and have them at the top of the list. This feature is especially useful for frequently used recipes, ensuring that they are easily accessible and available for quick access.
Semantic functions in applications
Bizzkit products are enhanced with semantic functions, offering a specialized mode of interaction with BAIA tailored to specific tasks. These functions enable users to perform more complex and context-aware actions within the Bizzkit ecosystem, making BAIA an essential tool for streamlined and intelligent workflows.
Examples include AI-generated text in PIM (both in the product details page and in mass edit) and automatic image description in DAM. New semantic functions are added frequently to all products, ensuring that Bizzkit continues to evolve and provide powerful, context-driven solutions for users.
BAIA for developers
BAIA extends its functionality through API access, enabling developers to integrate AI capabilities into their own solutions. By leveraging the BAIA API, developers can create custom functions, interact with semantic functions, and access the full range of BAIA features programmatically. This flexibility allows developers to build powerful, AI-driven applications that enhance user experiences and streamline workflows.
Semantic function endpoints
Developers can integrate BAIA’s AI capabilities into their own solutions through the specific APIs provided by each Bizzkit application. These semantic function endpoints enable developers to access AI-powered features within their applications, such as text generation for specific attributes in PIM or image recognition in DAM.
BAIA's own API provides developers with several endpoints for accessing BAIA's features outside dedicated applications. This includes functionalities like image-to-text conversion, all without needing a DAM.
By leveraging these endpoints, developers can enhance their applications with advanced AI capabilities, providing users with intelligent, context-aware functionality.
Custom functions
BAIA enables developers to create custom functions that integrate AI capabilities into various parts of the Bizzkit ecosystem, eliminating the need to manage semantic memory or handle AI model updates manually. This approach significantly reduces complexity and allows teams to focus on addressing specific business challenges.
While direct interaction with services like OpenAI or Gemini is possible, BAIA custom functions offer several distinct advantages:
- No need for complex boilerplate code to integrate AI – a few simple endpoints in the BAIA API are sufficient.
- No additional subscriptions or API keys required – everything is managed under the existing BAIA license.
- Security, GDPR, and compliance are fully handled by BAIA, ensuring that all interactions meet regulatory standards.
- Automatic updates of AI models, removing the need to monitor or decide on upgrades, as BAIA manages this process seamlessly.
- BAIA uses a RAG implementation, optimizing performance and reducing costs for AI-powered applications.
As examples of custom functions, the following videos from Bizzkit Academy’s training material demonstrate how developers can build solutions on the Bizzkit platform. The first example shows how to create a bot for a shop to answer customer questions using a fact book containing information like opening hours, payment options, and shipping details.
Another example illustrates a backend classification tool. In this case, internal product owners or team members can paste product reviews into the tool, and the AI automatically classifies them based on predefined criteria.
Notice the JSON in the response, making it easy to integrate with other systems or automate processes.
Both examples require very little backend coding since the functions are managed within the BAIA API. All requests are directed to BAIA rather than directly to the AI provider. Behind the scenes, BAIA leverages Retrieval-Augmented Generation (RAG), an architecture that augments a Large Language Model (LLM) with specific data retrieved from internal text. This ensures that responses are more informed and reliable, while also optimizing speed and reducing costs by minimizing unnecessary communication with cloud providers.
By leveraging custom functions, BAIA enhances the ability to automate processes, enrich data, and provide dynamic, context-aware interactions across the Bizzkit platform.