Site icon

Solve complex data challenges with CluedIn and Azure OpenAI

This post has been republished via RSS; it originally appeared at: Microsoft Tech Community - Latest Blogs - .

In this guest blog post, Tim Ward, CEO at CluedIn, discusses the benefit of combining graph databases with Microsoft Azure OpenAI and how CluedIn can enhance Azure OpenAI workflows.


In the effort to consolidate data across the enterprise, we often run into situations where standard logic will not suffice. Take the example of splitting an employee’s full name into first and last names. This seems like a simple task that should be easy to solve, but data often contain preferred names, middle names, salutations, and misspellings. With the right enhancements, Microsoft Azure OpenAI can automate processes and address even more complex data challenges.


Making data management easier for people

Azure OpenAI provides an interface to ChatGPT, the conversational large language model (LLM) from OpenAI. With Azure OpenAI, enterprises gain the ability to bridge people and machines, simplifying information access. A person can ask the AI about data without writing code, while the AI can process responses in a structured and autonomous manner, improving processing.


As a result, non-technical users can easily solve data challenges. This is the revolutionary change that everyone should be paying attention to. With Azure OpenAI, enterprises can remove large bodies of code and replace traditional approaches with accelerated implementations.


CluedIn natively powers up master data management on Microsoft Azure

However, transforming data requires adding guardrails to ensure that OpenAI’s ChatGPT produces trustworthy, functional, and practical information. In part, this assurance depends on traditional pillars of data management. It’s important to have reliable data delivery mechanisms (such as Microsoft Azure Data Factory), data governance and management (Microsoft Purview), and easy ways to explore data insights (Microsoft Power BI). So, how do we maximize value from ChatGPT using Azure OpenAI?


Enter CluedIn, a Microsoft Azure-native master data management (MDM) platform that enables business users and non-technical users to effectively transform raw data into information ready for generating insights. CluedIn natively integrates with Azure OpenAI and other Azure components, bringing non-technical users a powerful interface to solve complex data challenges with the simplicity of natural language. CluedIn does this by providing what ChatGPT lacks: context.


CluedIn closes the human-AI gap

By combining its graph database approach with Azure OpenAI, CluedIn provides valuable enhancements to help people ask AI the right questions and enable AI to produce trustworthy answers. CluedIn enhances prompts with context and specificity.

Imagine using the above org chart and asking OpenAI, “Who is the CEO of CluedIn?” I tried asking this question. Where a human might guess that the person on top of the chart is CEO—everyone reports up to this person—OpenAI gave an incorrect answer. There are many explanations for the error, including lack of training data and lack of knowledge. ChatGPT needed a little help.


This is where CluedIn graph database models help. If we have data about the CEO of CluedIn, we can treat him as a node in the network and consider the other people connected directly and indirectly to him. Note that such an approach is not achievable when using non-graph MDM systems. With this network of information, it’s possible to mark up with automated context to ask the same question, like this: “Who is the Australian CEO of the Danish master data management platform called CluedIn, that is also affiliated with Martin Hydahl and Pierre Derval?” Suddenly, the AI can pinpoint the answer.


CluedIn’s integration with Azure OpenAI helps close the gap between AI and humans by offering predictable hooks for humans to be notified and involved when the AI is not confident of its answers. This is much harder to achieve than most would have experienced by playing with LLMs. With CluedIn, users don’t need to be prompt experts.


By combining graph databases and Azure OpenAI, CluedIn enables you to easily build workflows that accurately respond to low-confidence answers from the AI and surface questions to users through Microsoft Teams. These notifications give users the questions asked and the AI answers, enabling people to instruct the system how to proceed. All using conversational language.


CluedIn’s integration with Azure OpenAI gives our customers a powerful interface to easily organize data across your business. If you’re interested in unifying your data and solving complex challenges, you can purchase CluedIn Master Data Management directly from the Microsoft Azure Marketplace. You can also learn more about CluedIn's expertise and partnership with Microsoft and see why Fortune 100 companies have already chosen CluedIn to re-invent master data management.

Exit mobile version