I work in the business of dealing with customer questions on Microsoft 365 all the time (disclosure), either directly or indirectly. This is part of a series of posts where I share them if they can be of help to others. Where I can of course and naturally, not just the questions but the answers too. All questions and answers strive to respect both sides sensitivities (parts will have been redacted and/or anonymised) and the main topic is covered in each post title.
What a whirlwind couple of weeks and months. Since OpenAI and ChatGPT landed properly in the context of its relationship with Microsoft, I and I am sure many of my colleagues at Microsoft have been inundated with questions. This is so fast moving its difficult to keep track. I have already shared several posts under the AI tag about what is being done from public announcements.
This post is a summarisation of things gleaned from public articles, websites and announcements for my own purposes – not unlike what ChatGPT might do 😉 In fact I have compared my input to a ChatGPT response – you be the judge.
Human summarisation – that’s moi 🙋🏻♂️
Please note this is by no means a commitment or definitive list in any way and some of the latter statements are more forward leaning with the earlier ones covering generally available solutions:
- Azure OpenAI Service. The Azure OpenAI Service offers customers access to the latest models from OpenAI (specifically GPT, DALL-E and Codex), and provides an enterprise-grade environment to experiment and deploy these models into production applications.
- Viva Sales generative AI-powered experiences. Suggests email content to sellers harnessing the power of GPT-3, combined with data from Office and CRM for an intelligent, contextual response. It increases seller productivity by reducing the time they spend managing e-mail to spend more time with customers. It helps sellers write better emails, faster based on the meeting. Works for both Salesforce and D365 customers
- Teams Premium. Intelligent recap and Live translation leverage GPT-3.5 to drive productivity and focus both during and after meetings. It make meetings more productive and impactful with virtual assistant, suggestion on action items & owners, AI-notes and personalized timeline markers.
- GitHub Co-Pilot. It uses the OpenAI Codex to suggest code and entire functions in real-time, right from your editor. It increases developer productivity by reducing time spent on creating boilerplate and repetitive code patterns, and creates more time for what matters.
- Microsoft Intelligent Data Platform. Azure SQL, Cosmos DB, Synapse, Purview and PowerBI combine to accelerate innovation and achieve agility. With PowerBI on the front end for example, you can find insights, make decisions and take actions through AI driven natural language report generation which saves time building complex report queries. Easy to use AI analysis capabilities like anomaly detection and AI powered insights and text summarisation is also included.
- Power Platform. OpenAI components will soon permeate all Power Platform modules as in the example above with PowerBI. For example, in Power Automate you will be able to describe what you want in a sentence, and an AI-based co-pilot will build a flow in seconds (built on GPT-3 + Codex).
- Office Productivity. This covers the scope of some of the traditional productivity tools like Word, PowerPoint, Excel plus many other smaller workloads in the M365 suite. As you can imagine, they will be enhanced through AI to harness the power of GPT-3, combined with data from the Microsoft Graph for an intelligent, contextual response in content creation, data driven or productivity tasks. Contextual ranking will resurface historical and relevant data in-the-moment to complete a task.
- Business Applications like Dynamics 365. In addition to the core sales function mentioned earlier, there will be a continuous focus on proactively anticipating & solving customer needs, provide real-time agent assist, power one-to-many customer engagements and continuously learn and improve with the support of AI. Support agents will be AI assisted and be supervisors of AI Agents.
ChatGPT’s take
First the question: What has been announced by Microsoft in terms of what OpenAI and ChatGPT elements are already available in its technologies and what is still to come?
And now the answers as I tried with both the Bing search application and then through the ChatGPT service itself.
Bing
Microsoft has confirmed that it has invested a total of $10 billion into OpenAI, the maker of ChatGPT1, a revolutionary chatbot that can generate natural and engaging conversations. Microsoft and OpenAI have also built a new kind of supercomputer specifically for ChatGPT and other generative AI technologies1. Microsoft plans to offer these systems to its own customers through Azure OpenAI Service, which is now generally available2. ChatGPT is based on GPT-3.5, a large language model that was trained on an Azure AI supercomputing infrastructure3. ChatGPT can sometimes write plausible-sounding but incorrect or nonsensical answers, so it still has some limitations3.
ChatGPT has many applications across different domains and industries. Some of them are:
- Code-Writer: ChatGPT can help you write code for a particular problem by generating snippets of code based on your natural language description1.
- Story-Writing Application: ChatGPT can help you write stories by generating creative and coherent narratives based on your prompts or genre preferences1.
- Debugging Code: ChatGPT can help you resolve errors in your code by suggesting fixes or explaining the causes of bugs1.
- Explaining Difficult Concepts: ChatGPT can help you understand complex topics by simplifying them or providing examples in natural language1.
- Diet Planning: ChatGPT can help you plan your meals by generating personalized and healthy diet recommendations based on your goals, preferences, and allergies1.
- Business and Marketing Ideation: ChatGPT can help you generate creative ideas and strategies for your business or marketing campaigns by providing relevant information and suggestions2.
- Content Creation: ChatGPT can help you create high-quality content for email or social media marketing by generating catchy headlines, engaging copy, and call-to-actions3.
- E-Learning: ChatGPT can help you educate and train students on various topics by providing interactive lessons, quizzes, feedback, and guidance3.
These are just some of the examples of how ChatGPT can be used. There are many more possibilities that are yet to be explored.
ChatGPT Service
Microsoft has announced several collaborations with OpenAI, including the integration of OpenAI’s natural language processing technology into Microsoft’s products and services. One of the most prominent examples of this integration is the use of OpenAI’s GPT-3 language model in Microsoft’s Power Platform, which allows users to generate natural language descriptions of data.
Additionally, Microsoft has also integrated OpenAI’s language model into its Azure Cognitive Services, allowing developers to use it for a variety of natural language processing tasks, such as sentiment analysis and language translation.
In terms of what is still to come, Microsoft and OpenAI have announced plans to continue their collaboration and develop new technologies together. One of the most exciting developments is the creation of a new supercomputer for OpenAI, which will be hosted on Microsoft’s Azure cloud platform. This supercomputer will be used to train even larger language models and accelerate the development of new AI applications.
Overall, Microsoft’s collaboration with OpenAI has already produced significant advancements in natural language processing technology, and we can expect even more exciting developments in the future.