There has been a rush to incorporate the power of generative AI in everything from search engines like Bing to customer relationship management products such as Salesforce since the unexpected success of OpenAI’s ChatGPT. Now Microsoft, a major partner of OpenAI is bringing that same technology to its Office 365 productivity suite with ‘Copilot for Work’. The tool will see generative AI added to Word, Excel, Powerpoint and Outlook as well as the launch of a new Business Chat tool.

Microsoft will use the Copilot brand to large language model tools across its product range from 365 to GitHub (Photo: Microsoft)
Microsoft will use the Copilot brand to bring large language model tools across its product range from 365 to GitHub. (Photo: Microsoft)

Microsoft invested $10bn in OpenAI and as part of that deal has access to its large language models. To date, the company has incorporated this into a number of products including Bing Search, its CRM platform Dynamics 365 and integration with the Azure cloud service. In Bing there is an AI chatbot tool, similar to ChatGPT and powered by GPT-4, that delivers more context-aware results.

Copilot for Work was revealed today in an online event hosted by CEO Satya Nadella and Jared Spataro, corporate vice president for modern work and business applications. Nadella described it as a “whole new way of working”, moving AI from “autopilot to copilot”.

He said: “We believe it will unlock a new wave of productivity growth with Copilots freeing us to rediscover the joy of creation. Just as we can’t imagine computing today without keyboard, mouse or multitouch, going forward we won’t be able to imagine computing without natural language prompts and human copilots.”

Spataro described it as a new era of computing. “Work is meant to be more than just a job. At its best work not only feeds our families, it feeds our souls but today the reality often falls short of that promise as we spend too much on the drudgery of work,” he said. “The tools that should make us productive often get in the way.”

“To reconnect with the soul of work we need a whole new way of working. We are announcing Microsoft 365 Copilot. It combines the power of large language models with company data and 365 apps,” he said, with the aim of turning “your words into the most powerful productivity tool”.

The AI has been fully integrated into Microsoft 365, including being embedded into everyday apps like Word, Excel, PowerPoint, Outlook and Teams. including allowing users to add content to existing documents, summarise texts, write an entire document and even convert a Word document with descriptive terms into a PowerPoint document. In Excel it can analyse and explore data, offering natural language questions over the data or creating formulas from natural language. This could also include what-if scenarios to suggest new formulas.

Each product gets a Copilot update

Copilot in Word can be used to write, edit, summarise and create content “alongside the user”, acting as a copilot rather than an autopilot. “With only a brief prompt, Copilot in Word will create a first draft for you, bringing in information from across your organisation as needed. Copilot can add content to existing documents, summarise text, and rewrite sections or the entire document to make it more concise. You can even get suggested tones—from professional to passionate and casual to thankful – to help you strike the right note.”

In Excel, it can answer questions on the data set in natural language, reducing the need for complex formulas and reveal correlations, propose what-if scenarios and even suggest formulas based on questions you’ve asked. It can identify trends, create visualisations and ask for recommendations to drive different outcomes.

The PowerPoint version of Copilot has been designed as a “storytelling partner” that can transform a written document into decks complete with speaker notes and sources. It can also do this from a simple prompt or outline and condense a lengthy presentation at the click of a button including retiming animations.

Outlook will be capable of better quality email triage including summarising lengthy, convoluted email threads, even if they contain multiple people and voices with different viewpoints. It can highlight open questions yet to be answered and respond to an email with a quick prompt or turn notes into a professional message. It can use content and data from across Microsoft 365.

Microsoft has also unveiled a new product called Business Chat. It works across the LLM, the Microsoft 365 apps, and user data from calendars, emails, chats, documents, meetings and contacts to give natural language search access. “You can give it natural language prompts like “tell my team how we updated the product strategy” and it will generate a status update based on the morning’s meetings, emails, and chat threads,” Microsoft revealed.

This is being branded an entirely new product category within Microsoft 365 that the company says will allow users to “spend less time focused on the tools and more time focused on the most important work. Potential prompts include asking it to summarise all of the chats, emails and documents on a particular customer escalation, or when the next milestone on a project will be hit and any risks identified. It can then use the various 365 apps to carry out tasks on those findings.

The products are being initially tested among 20 customers including eight in Fortune 500 enterprises with plans to expand in the coming months. Specific details on pricing will come over the next few months.

Microsoft Copilot for Work follows Google’s AI Workspace tools

This follows the news that Google plans to incorporate its own massive 540-billion parameter PaLM large language model into Workspace, initially starting with Docs and Gmail and offering users the ability to generate text from a simple prompt. There are already some AI features in Workspace including Gmail’s Smart compose and summaries in Docs but to this point, it hasn’t had the full generative capabilities seen in tools like ChatGPT or Wordtune Spices from AI21 Labs.

Google says there will be a new set of features that can produce content after typing in a topic. This will generate a draft instantly that can be adapted as needed. These features will be gradually rolled out to a select group of testers in the coming weeks but it isn’t clear when they will be widely available.

Forrester Analyst Rowan Curran said embedding generative AI into everyday office and productivity tools has the potential to “significantly change people’s workflows in whatever role they have” as having the ability to generate a summarisation of a white paper into a blog post within a core productivity app reduces friction around integrating different tools.

“However, these models are not nearly as powerful in their foundational version as when they have been fine-tuned on a company’s specific data – and so when the major productivity suite providers start enabling this capability is when we may start to see an acceleration of the use of these capabilities, if they don’t immediately take off in their initial version. The impacts might not transform the workplace tomorrow, but the wheel has started rolling forward on this.”

Tech giants must address issues of AI trust and security

“Twenty years ago, in response to the security debacle that was Windows XP, Bill Gates pledged in a memo that Microsoft would put security first and not rush features and products to market,” said AJ Grotto, former White House Director of Cybersecurity Policy. “Has Microsoft rescinded this memo? Sure seems like it with its rush to integrate ChatGPT – a tool with known problems – into Office 365.”

Marshall Choy, SVP of Product at SambaNova Systems says this is evidence of ChatGPT shifting from being a curiosity to the technology behind it becoming a real-world application that millions of people will soon be taking advantage of to find productivity improvements, but warned the one-size-fits-all approach isn’t ideal for all situations.

“Rather than using generative AI to simply improve business processes such as creating content, what happens when the models are at the very heart of businesses? These types of foundation models are where it gets interesting,” says Choy. “Compared with one-size-fits-all-models where your data needs to be shipped off to a third-party API, customised owned models bring huge efficiency savings that give businesses a clear competitive advantage in their lines of work.”

Read more: Microsoft is now a cybersecurity titan. That could be a problem