Harness AIDA Support
Overview
In today’s fast-paced world, quick and efficient access to information is more crucial than ever. Technical documentation especially can often feel like navigating a labyrinth of vast knowledge, making it challenging to find answers promptly. At Harness, we understand that navigating through our extensive documentation, blogs, pricing plans and other public resources to find answers can sometimes be a challenge, leading our customers to log support tickets and wait for a response from Harness support. That’s why we’re excited to introduce the Harness AIDA Support, your dedicated assistant ready to provide instant answers to your questions about Harness and its products.
Meet AIDA Support
Harness AIDA Support is designed to streamline your access to information by harnessing the power of AI. It works seamlessly inside the product, offering you a user-friendly interface to ask questions. When you ask a question, AIDA Support quickly scans our vast knowledge base to provide you with relevant and accurate answers, all in real-time. Let’s see it in action.
After logging into your account, click on the AIDA support button located in the bottom right corner of your screen. From there, feel free to ask your question. Within few seconds, you will receive a response, complete with direct links to relevant documentation for easy reference. You can also let us know if the answer was helpful or not. Your input plays a crucial role in fine-tuning and enhancing the accuracy of our assistance.
Before we dive into how it works internally, let’s first understand what embeddings are.
Embeddings
In simple terms, an “embedding” is a fancy word for a vector that represents text. Vector is a list of numbers that help us measure how similar or related different texts are. In order to generate embeddings, embedding models like Google’s text-embedding-gecko or OpenAI's ada models can be used.
Visualizing Embeddings
To better understand embeddings, let’s visualize them. Imagine a vast space where words and phrases are scattered like stars in the night sky. The distance between these points in space signifies the relationship between the texts they represent. Closer points indicate greater similarity, while distant points represent differences.
Consider the example of animals and transportation. Words like “cat,” “dog,” “bird,” and “turkey” are clustered together because they share a common theme — they are all types of animals. In contrast, “car,” “airplane,” and “helicopter” are grouped because they represent modes of transportation. “Soccer,” unrelated to both groups, is positioned far away. This arrangement showcases how embeddings intuitively capture the relatedness of texts.
In essence, embeddings are the secret sauce of natural language processing. They make it possible for computers to understand the nuances, context, and connections within language.
Embeddings have various use-cases -
- Search/Information Retrieval: Embeddings empower search engines to find results that closely match your query. When you search for “puppies,” the engine retrieves content closely related to adorable dogs, not car engines.
- Text classification: Imagine sorting a library of books into different genres. Embeddings assist in classifying texts by their content, allowing systems to categorize them accurately.
- Similarity search: Embeddings simplify the process of finding related texts. Just as similar books are placed near each other in a library, embeddings group related texts for easier discovery.
- Recommender Systems: When you shop online or stream movies, embeddings help recommend products or films similar to what you’ve liked before. They understand your preferences by analyzing your past choices.
- Anomaly Detection: In the world of data analysis, embeddings can identify unusual or unrelated text snippets by measuring how far they deviate from the norm.
Now that we understand embeddings, let’s explore how Harness AIDA Support works internally.
The Blueprint of Harness AIDA Support
- Your Question to the AIDA Support: It all begins when you send a question through the prompt. Your queries are where we start. You can ask anything related to Harness products, documentation, or troubleshooting.
- Turning Your Question into Numbers: To truly understand and process your query, we convert it into an “embedding.” As discussed above, embedding is a numerical representation of your text that helps us make sense of it. We use Google’s
textembedding-gecko@001text-embedding model for this. - Searching the Vector Database: Our support system relies on a vector database that holds embeddings for all of the content of Harness Developer Hub. This database is key to finding the most relevant information related to your query. We’ve integrated Weaviate as our vector database. Weaviate ensures rapid and accurate retrieval of embeddings, making your experience seamless. When you ask about deployment troubleshooting, Weaviate helps identify relevant documents, blogs, or articles related to deployment troubleshooting in our extensive knowledge base.
- Creating Context: We identify vectors within the database that are closest to your query and retrieve them in their original text form. This text is then combined to create context. In essence, context represents text from Harness resources that potentially contains the answer to your query. This step helps ensure that the response provided is not only accurate but also contextually relevant. For instance, the AI identifies documents discussing deployment troubleshooting and combines them to provide a comprehensive context for your query.
- Engaging the Gen AI Models : Now, it’s time to bring in the big guns — the OpenAI’s GPT3.5 model. We provide this model with your query, a tailored prompt, and the context we’ve just created. The prompt guides the model in providing you with an answer in the desired format using the context we’ve prepared.
Here is a visualization showing the above:
Conclusion
The Harness AIDA Support represents a significant leap forward in customer support and access to Harness resources. With the power of GenAI, we’ve harnessed the ability to provide instant, accurate answers to your questions about Harness.
Harness AIDA Support streamlines your experience, ensuring that you have the information you need at your fingertips, making your journey with Harness smoother and more efficient than ever before. We’re committed to providing you with the best support possible, and AIDA Support is a testament to that commitment. Whether you’re a developer seeking technical documentation or a user with product inquiries, AIDA Support is here to empower you with the knowledge you seek, 24/7, without the need to wait for support tickets or responses. Harness the power of AIDA Support to supercharge your interaction with Harness products.
