Shulou(Shulou.com)11/24 Report--

April 6 news that Microsoft has integrated OpenAI's GPT-4 into the Microsoft 365Productivity Application Suite, which is a very difficult task, but the company has been hoping to complete it as soon as possible. As a result, Microsoft mobilized hundreds of employees to work long hours to brainstorm product solutions and develop three working modes of artificial intelligence assistants based on a unified design framework.

The following is the translation:

"3-16" is a word that has been mentioned countless times in Microsoft conference rooms and corridors in the past few months. It refers to March 16, when Microsoft announced that it would introduce the generative AI model developed with OpenAI into the Microsoft 365Productivity Application Suite. By integrating GPT-4, every productivity application, including Outlook, Word, and Teams, will have a Copilot (intelligent co-pilot) based on generative artificial intelligence. Currently, more than 20 Microsoft customers are testing the technology.

Integrating artificial intelligence assistants into so many applications is a daunting task, but Microsoft hopes to finish it as soon as possible. In November last year, the ChatGPT released by OpenAI caused a global sensation and set off an artificial intelligence competition, with companies racing against the clock to launch new artificial intelligence products and functions to seize the market opportunity. Even Microsoft, which has partnered and invested with OpenAI, is no exception.

Microsoft Design Director Jon Friedman (Jon Friedman) is responsible for Copilot product design in Microsoft 365. He said the project requires the long hours of hundreds of Microsoft employees, including designers, engineers, product managers, marketers, data scientists and ethics teams, and the entire project will last for several months. you even have to work overtime on weekends. Friedman said the plan also requires people to restrain themselves and work together to build such a big project in such a short time that everyone feels they have to let go and stick together.

"it's exciting that we can do something really bold and important together," Friedman said. Although we have a lot of experience in artificial intelligence, this special generative artificial intelligence is more powerful, so I think everyone is working with a learning mentality. "

New user experience Friedman believes that Copilot is a new form of "conversational user experience" that can invoke more rich resources. The Copilot, he says, is an office assistant that represents "a new frontier in user interface design, changing established patterns like the first generation of touch-screen devices". It is reported that Copilot is based on GPT-4 large language model development, pre-trained a large number of Internet content, and can access Microsoft Graph business data, so as to automatically generate e-mail and meeting points and other content.

The challenge in designing a user interface like Copilot is how and when to showcase this new artificial intelligence assistant when people typically work with applications such as Word and Powerpoint.

Friedman said the original design of an artificial intelligence assistant that could be called through a variety of productivity applications was "a vague idea". But as the design team gained a deeper understanding of the application of artificial intelligence assistants in real business, this idea began to become clear. First of all, the design team needs to find specific use cases such as artificial intelligence can significantly save users' time or stimulate creativity in some way, which is the first step in the user experience design process.

The people who know the use cases best include engineers, product managers, designers, and computer scientists for each productivity application. Friedman's design team worked with them. When the Copilot project was launched, he asked all product teams to brainstorm to explore how to use generative artificial intelligence to improve the functionality of productivity applications. Friedman then set up a special horizontal design team to work with all application teams to demonstrate the effectiveness of Copilot in each application.

Friedman said that as various application teams began to develop use cases, horizontal design teams began to notice the commonness between these use cases, namely artificial intelligence use cases associated with multiple applications.

"our discussion was very valuable," Friedman said, recalling the brainstorming process at the time. "We're talking about the specific functional requirements of Copilot, such as how the new generative language model will help us better write email summaries."

As cross-application use cases become clearer, horizontal design teams begin to think that the artificial intelligence helper capabilities of each application do not need to be different.

Friedman said: "because you have a lot of people. Try to look at each scene and be able to make a rough judgment like,'Ah, there's this thing, too.'"

As a result, they began to devise a design framework that allows a general-purpose assistant to work in various applications in several different predictable ways.

A design team led by three levels of Copilot Friedman created an in-depth document library designed to help designers throughout the project create pointcuts for artificial intelligence in a given application. They instruct designers to determine how to invoke Copilot based on different tasks that users may be involved in. "there is an idea that Copilot should be at the right level and do the right job," Friedman said. "

The design framework stipulates that Copilot can be displayed in three ways in the application user interface.

The first is an immersive user experience that allows artificial intelligence assistants to focus on specific business projects rather than on a particular application, so that you can actually extract data or key points from multiple applications to serve the task at hand. For example, Copilot might collect project milestones or risk points from team meetings, slides, or email content, and then summarize and summarize them in the project plan document.

The "immersive" experience model is the most powerful and probably the most influential feature of Copilot in the productivity application suite. Rob Enderle, chief analyst at Enderle Group, a market consultancy, believes it could also help solve a long-standing problem with the Microsoft productivity suite-the lack of tight integration between applications. The reason, Endler said, may be that Microsoft initially bought these applications from other companies and did not share a code base with each other. But Copilot can cover all applications, or at least make users feel that they can work together in some tasks.

Friedman says the second way of presentation is "auxiliary," which means that the Copilot is like the "side fight" of a sidecar tricycle, helping users make the most of the app's functionality in a particular app. For example, in PowerPoint, Copilot can show users how the deep graphics capabilities of an application can be used to describe complex data sets; in Outlook, Copilot can help users understand the most important things in e-mail; and in Word, Copilot can provide feedback on how to better write documents and fit a particular writing style.

In addition, in Copilot's "embedded" presentation, artificial intelligence can exert generative and creative capabilities in applications. For example, Copilot might appear in the pop-up window of an Word document. "it's like a random experience." "when you put your heart and soul into your work, Copilot can help you solve your writing bottleneck, or automatically use text content to help you start making slides," Friedman says. "

Friedman said that Copilot's horizontal design team began to use the concept of "three levels" to describe the work, and members of various application teams gradually accepted the framework.

"We have shared this framework with CEO Nadella and other members of the company's executive team, and it has been accepted by almost everyone," he said. this idea can be applied to three different levels of work. "

The name Copilot is not specifically set up for Microsoft 365. Microsoft's GitHub used this name to refer to programming assistants in 2021, and some of its features also use OpenAI's large language model. Microsoft, on the other hand, is creating a consistent Copilot assistant that can perform some standardized functions in different productivity application suites. For users who have been using productivity applications for a long time, this new job is likely to reduce the strangeness they will encounter when the new version of Microsoft 365 is officially launched. By using a unified assistant, users can use different applications more easily, thus improving their productivity.

As generative artificial intelligence is further integrated into Microsoft's consumer and enterprise products, the Copilot brand and concept may be extended to other Microsoft products such as the Windows operating system and even LinkedIn.

"Nadella likes the name very much because it appropriately describes the work of artificial intelligence assistants," Friedman said. "it exists to help you and guide you in many tasks."

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.