Microsoft Build 2023: Establishes a framework for creating AI applications and broadens the Copilot ecosystem

Microsoft announced the addition of Power BI, Power Pages, Microsoft Fabric, and Windows to its Copilot ecosystem on 23 May at its annual Build developers conference. Microsoft also introduced new technologies that it claimed would enable programmers to create copilots and cutting-edge AI applications. The usefulness of the copilots is improved by these features, which include plugins that let the copilots communicate with other programmes and services.

The development of software using intelligent copilots is on the rise

According to Kevin Scott, CTO of Microsoft, software without an intelligent copilot helper will become increasingly rare shortly. Applications such as copilots combine AI and big language models to help users with complex cognitive tasks. They can assist in creating a sales pitch to organising a theme-based dinner party. With GitHub Copilot, Microsoft first popularised the idea of copilots. 

For ChatGPT, Microsoft embraced the available plugin standard created by OpenAI, providing compatibility between ChatGPT and Microsoft's copilot services. Microsoft is an OpenAI investor. According to the company, developers may now develop plugins for commercial and consumer platforms like ChatGPT, Bing, Dynamics 365 Copilot, Microsoft 365 Copilot, and Windows Copilot. According to Microsoft, these plugins make it easier for copilots to integrate with external systems, allowing them to retrieve real-time information, including business data, and carry out new computations.

The ability of plugins to function

Copilots and external systems or data sources are connected using plugins. They enable copilots to secure websites or backend systems, access private company data, and enhance their capabilities. For instance, a plugin might allow a copilot to access legal documents stored in a company's database to respond to legal-related queries. The ecosystem of plugins is anticipated to expand quickly and provide copilots with various functionalities.

Copilot development simplification and ethical AI upgrades

Microsoft unveiled new tools to simplify copilot development, including Azure AI Studio and Azure OpenAI Service on data. According to Microsoft, these technologies allow developers to base conversational AI models on personal information and extract organisation-specific insights. Additionally, Microsoft unveiled Azure AI Content Safety, a service intended to identify and filter offensive text and image content. The Azure OpenAI Service integration makes it simple to access content safety features.

According to Sarah Bird, a partner group product manager at Microsoft who leads responsible AI for fundamental technologies, developers must also make sure the copilot produces the desired results and stays away from outputs that are biassed, sexist, racist, hateful, violent, or encourage self-harm.

Extending the copilot framework

Building copilots requires developers to use generative AI technology like GPT-4 via Azure OpenAI Service. A model must be chosen, a meta prompt must be defined, memory must be used, plugins must be used to access data and services, and complex processes must be orchestrated. Microsoft offers technologies like the Semantic Kernel SDK and Azure Machine Learning prompt flow to combine huge language models with conventional programming languages, speed up copilot development, and ensure ethical AI practises.

The plan for Microsoft's copilot platform

Copilots are seen as a new paradigm in AI-powered software and a revolutionary software development method by Microsoft. It includes user experience, architecture, services, safety, security, and product scenarios. The Microsoft ecosystem makes Copilot development possible, which provides Azure and delivers the necessary resources and tools. Microsoft wants to make it possible for developers to iterate quickly and successfully implement their copilot concepts.