Azure OpenAI Service equips apps with GPT-3

November 23, 2021 - Research V179

By using the OpenAI API in Azure applications, business analysts can expect to improve productivity by 10 to 25 percent by automating SQL query and report generation. This is achieved by leveraging OpenAI’s GPT-3 natural language model for text generation, and derivatives such as Codex for automated coding, and a SQL tool enabling users to query in natural language. As an example, if a user prompted the OpenAI SQL tool by asking “what is the count of employees working in the HR department” the tool would output “SELECT COUNT(*) FROM worker WHERE DEPARTMENT = ‘HR’;”. By delivering GPT-3 and other OpenAI tools to Azure applications, Microsoft brings the natural language models to new roles and workloads, while limiting potentially harmful use with built-in AI-based controls, guardrails, and security to enhance productivity and elevate non-technical users across departments. This access to the OpenAI API also helps maintain Microsoft’s competitiveness with other public cloud providers such as Amazon and Google, each of which offers its own highly trained built-in natural language models.