Do you remember the excitement of holding your first smartphone? The sleek design and its endless possibilities felt like a glimpse into the future. Soon enough, smartphones became an essential commodity worldwide, enhancing productivity and communication. Technology continues to evolve at a rapid pace, and one of the most prominent developments in recent times is Generative AI, which has created numerous new possibilities.

Oracle HeatWave GenAI provides integrated and automated generative AI with in-database large language models (LLMs); an automated, in-database vector store; scale-out vector processing; and the ability to have contextual conversations in natural language—letting you take advantage of generative AI without AI expertise, data movement, or additional cost. You can benefit from generative AI without the complexity of external LLM selection and integration, and without worrying about the availability of LLMs in various data centers. You can use generative AI with your business documents without moving data to a separate vector database and without AI expertise. In addition, you can combine generative AI with other built-in HeatWave capabilities such as AutoML to build richer applications.

Let us look at a few examples of applications you can easily build with HeatWave GenAI, with information on how they’re implemented. You’ll also be able to see demos! 

HeatWave GenAI for e-commerce applications

In the fiercely competitive e-commerce landscape, businesses are constantly on the lookout for innovative strategies to retain customers. HeatWave GenAI enables eCommerce companies to:

• Simplify product review analysis: Picture an e-commerce platform with a vast product range and a multitude of reviews for each item. Customers often find it challenging to sift through all the reviews to make a purchase decision. With HeatWave GenAI, creating a product review summary from all the reviews is as simple as writing a SQL query. This query transforms the unstructured review data into a natural language summary in real time. Any new reviews are automatically integrated into the product review summary, keeping it up to date. HeatWave GenAI provides all the needed components, no need for example to use external LLMs.

Watch the HeatWave GenAI for e-Commerce Demo.

HeatWave GenAI for technical support applications

Technical support teams can struggle to provide quick and efficient assistance to customers due to the large volume of data they must handle. This can cause delays to resolve issues and negatively impact the overall customer experience. By using HeatWave GenAI, technical support teams can quickly generate helpful recommendations and solutions from their extensive knowledge base. This empowers them to easily address complex domain, product, or service-related issues. Consequently, technical support processes can become more efficient, leading to faster problem resolution and increased customer satisfaction.

For instance, ‘Ask MySQL Expert’ (AskME) is an innovative troubleshooting tool built with HeatWave GenAI. It empowers the HeatWave MySQL support team to speed up the process of solving customer issues raised as support requests and serves as the first-line filter for the support team. A wide variety of publicly available documents, such as user guides, white papers, reference manuals, as well as Oracle Restricted documents such as bug details and support documentation, are loaded in object storage, which is referred to as the knowledge base. HeatWave GenAI generates in-database embeddings of the data in the knowledge base and stores them in HeatWave Vector Store. Then, HeatWave GenAI uses in-database LLMs and the vector store to find relevant documents, summarize them, and provide answers. Here are some more details on the process:

• Finding relevant documents: When you ask a question to AskME, HeatWave GenAI leverages the semantic search feature of HeatWave Vector Store to gather all relevant documents from the large knowledge base. This assists the support team in locating relevant information and helps ensure that they have access to a wide range of documents to address the support request effectively.
• Summarizing documents: AskME can generate summaries from the information collected from multiple documents, saving the support team time and effort by eliminating the need to review each individual document. Additionally, it provides references to these documents so that the support team can explore the content in more detail.
• Providing formatted answers: HeatWave GenAI can generate answers in various formats to make it easier to use in any type of documentation. For example, AskME can produce answers in both general-purpose and Oracle Diagnostic Management (ODM) format. ODM-formatted answers can be directly included in troubleshooting documents, saving the support team additional time and effort.

Watch the HeatWave GenAI AskME Demo.

HeatWave GenAI for Healthcare application

Although healthcare organizations have used AI technology for years, Generative AI represents a meaningful new tool that can help bring years of propriety or public data to a clinician’s fingertips in seconds.

Let’s explore a healthcare app that uses HeatWave GenAI to extract relevant information from their propriety content, which can lead to better decision-making, enhanced patient care, and streamlined processes. Enterprises have the option to choose from the following:

• HeatWave GenAI with vector store: Customers looking to create content based on their own proprietary data, such as research papers, patient histories, and clinical records, can utilize HeatWave’s in-database vector store. The process involves uploading unstructured enterprise documents like PDFs, HTML, TXT, PPT, or DOCX files to object storage. HeatWave then parses the text, generates embeddings inside the database, and stores them in HeatWave Vector Store. When a natural language question is asked, embeddings are created for the question, and a similarity search is conducted with the stored embeddings to identify the most relevant content related to the question. This content is then used to improve the prompt given to the LLM, enabling it to provide a more precise, tailored response, effectively leveraging proprietary data to deliver targeted and accurate results. Additionally, references to the sources of the response can be viewed.
• HeatWave GenAI without vector store: Enterprises looking to leverage Generative AI on publicly available data can use HeatWave GenAI’s in-database LLMs such as Mistral-7B-Instruct. These in-database LLMs are smaller parameter, quantized LLMs that run on CPUs. They’re available at no additional cost in all regions where HeatWave is available.

See the HeatWave GenAI for Healthcare Demo.

HeatWave GenAI for personalized meal recommendations

HeatWave enables customers to combine machine learning with Generative AI in a single database. They can leverage insights obtained from HeatWave AutoML to deliver more accurate and personalized responses, resulting in enhanced user experiences and improved decision-making support. Whether it’s suggesting products on e-commerce websites, recommending content on streaming platforms, or recommending dishes via a food delivery app, these personalized recommendations in natural language, based on user preferences, past behavior, and interactions with a platform or service, enhance the user experience by presenting relevant and engaging choices.

Picture a scenario where a user is asking for recommendations for vegan dishes made with tofu. In response, the HeatWave AutoML recommender system suggests top restaurants based on the user’s preferences and past orders. The menus for these top restaurants are then retrieved from the vector store, enabling the LLM to suggest specific dishes to the user, using natural language.

HeatWave GenAI greatly simplifies the development of generative AI applications at lower cost. You don’t need to integrate multiple services, to be an AI expert, or to move data.

Try it out and let us know what applications you’re building with HeatWave GenAI!