At Microsoft, our internal communications teams are adopting a growth mindset and encouraging a transparent culture of learning to use AI more effectively for creative work. They’re integrating AI into their daily tasks and collaborating between teams to save time and enhance content quality in our internal communications. 

“Our team is fully dedicated to using AI as a partner for content creation,” says Diana McCarty, the principal group content program manager for our IT Communications team, which sits in Microsoft Digital, the company’s IT organization. “Our team culture is to learn and grow and try new things. We’re constantly pushing what we’re capable of, and we push the capability of the tools we use. So, we decided that we’re going to go learn everything that we can about AI as individuals, bring it back to the team and collectively grow together.”

McCarty’s team consists of six content program managers who run 17 internal programs at Microsoft, spanning Microsoft 365 Copilot, Microsoft Teams, and network infrastructure. The team supports internal communications and helps ensure that Microsoft’s 336,000 employees, vendors, and contractors have the information and encouragement they need to successfully adopt and embrace technology change.

AI tools and processes are part of approximately three-quarters of their daily content production processes and a driver for creativity across the team.

“We’re marketers, too,” says Sarah Lundy, a senior content program manager. “Our creative marketing campaigns inject a little sunshine into employees’ days and drive understanding and action at the same time.”

Why we’re prioritizing AI in communications

AI tools like Microsoft 365 Copilot have become a game-changer for the IT Communications team and communications teams across the company.

“I can’t even imagine a world where I don’t have Copilot to run things against,” says Allison Michels, a senior program manager in the Microsoft Viva product group who works closely with the IT Communications team. “Whether it’s creating webinars, writing blogs, or responding to emails, AI has become an integral part of my daily communication tasks.”

AI handles repetitive tasks, giving communicators more time for strategic and creative work. AI also helps the team analyze data quickly and accurately, providing insights that guide decisions and strategies.

“There’s no fear of AI taking our jobs,” says Laura Oxford, a program content manager on the IT Communications team. “We found the more we used AI and played around with it, the less worried we became about it replacing roles. In fact, AI is enhancing the human capabilities of our team rather than replacing them. It can’t replace the unique value that human communicators bring to the table.”

The team has observed significant enhancements since prioritizing AI, including boosted efficiency, increased creativity, and greater insight into their daily work. Embracing AI has helped the team stay ahead and deliver more impactful content.

AI skills are critical in the modern workplace

The IT Communications team has also been using AI to become better communicators by identifying new opportunities for learning, growth, and skill enhancement. Using Copilot to compose and critique their content helps identify their strengths and weaknesses, providing personalized recommendations for improvement. This included asking AI questions such as:

• “Can you help me improve this text for clarity?”
• “Can you help me rephrase this text to avoid passive voice?”
• “Please proofread the following message and highlight any changes you make.”

AI skills are a catalyst for career growth, and those skills are quickly becoming the standard. According to Microsoft’s 2024 Work Trend Index Annual Report, 66% of leaders surveyed said they would not hire someone without AI skills.

AI enables continuous learning by allowing the team to quickly research and summarize the latest industry trends and technologies.

“We knew really early on that we had to learn about this and get behind it,” Oxford says. “The more we used it and played around with it, the more excited we became about the future of communication with AI.”

The team isn’t using Copilot for content creation alone. They’ve also been using AI in the flow of their Microsoft 365 modern work practices, where it simplifies meetings, enhances email and calendar management, and enhances networking opportunities by creating valuable opportunities for deeper collaboration and career building. The team has even used Copilot to showcase their skills and achievements to the rest of the organization, increasing their visibility, sharing their expertise, and growing Copilot and AI adoption across Microsoft.

Building AI into the flow of work

Successfully adopting Copilot into the flow of work didn’t happen overnight. In fact, according to research, it can’t.

In a survey of Copilot users across different functions and industries, employees were asked about AI’s impact on various areas, including productivity, work enjoyment, work-life balance, and the ability to attend fewer meetings. 

The findings presented what the researchers called “the 11-by-11 tipping point,” which shows the time it takes for individuals and an organization to start building a habit of using Copilot consistently. A time savings of just 11 minutes a day was the magic number where users started to see value from AI. Additionally, they also found that 11 weeks is the breakthrough moment when most people say Copilot is consistently improving productivity, work enjoyment, work-life balance, and the ability to attend fewer meetings. Thus, the 11-by-11 tipping point. 

The team’s experiences support these findings. They built their skills with Copilot through a structured and collaborative approach. Initially, there were challenges.

Some team members were concerned about the implications of AI on their roles and the broader impact on the workforce. There was a significant amount of anxiety about what AI could mean for their jobs and the future.

Another challenge was the need for effective prompting skills. The team quickly realized that the quality of AI-generated content depended heavily on how well they could prompt Copilot, making prompting an important skill for effectively using Copilot.

Additionally, the team encountered a learning curve integrating Copilot into their existing workflows. Some team members struggled initially to find successful and productive ways to integrate the tool into their daily rhythms and work patterns.

These challenges all had a common solution-set: awareness, education, and expertise-building. They gathered for training sessions to help everyone understand the capabilities and limitations of Copilot and found the knowledge—and people—they needed to overcome the obstacles. This approach helped demystify the technology and alleviate fears, allowing team members to see AI as a valuable tool rather than a threat.

Lundy was instrumental in building and growing AI and Copilot adoption on the team and even deeper into other parts of the organization. “In the beginning, it was really informal,” she says, “A few of us set up a little virtual team where we would get together every week and share what we were learning, try things, and experiment. It was very organic.”

Lundy’s approach grew from there. Team leaders, recognizing the potential impact of AI on communications development, gave Lundy the opportunity to develop an AI learning program for the team. The program’s goal was to advance the use of AI in communications, helping the team to improve efficiency, content quality, and create longevity for the team as communicators.

She created a survey for the team asking questions about concerns, current AI usage, aspirations for using AI, comfort level with AI, learning topics, and more. Using the data she collected, she set up a schedule of learning topics and started identifying potential presenters.

“We had biweekly meetings for about a year with different AI related topics,” Lundy says. “In some cases, people on the team were already working on something, so they would put together a little presentation. In other cases, I’d find someone who was on the product group for Microsoft Excel or Microsoft Designer, for example, and have them come to our team and do a presentation.”

The team also shared their experiences, prompts, and other resources in a Microsoft Teams channel that’s still very active.

The effort grew more when some team members began providing prompting training sessions for thousands of employees across Microsoft. This training became the foundation for how the team—and others at Microsoft—approached AI, emphasizing the importance of experimenting, sharing learnings, and collaborating.

Practical applications of AI in communications

So how do they do it? What does AI usage look like in our IT Communications team?

“We’re thinking about how we can use AI for every single communication that goes out from our team,” says Eva Etchells, a senior content program manager on the team. “Using AI to generate ideas and options that we can choose from and refine is ingrained into our processes. We’d be much less effective—and less creative—without it.”

Injecting creativity and efficiency into subject lines

In the early days, the team started using AI to generate subject lines for emails.

“Subject lines are extremely important for communications,” Oxford says. “The quality and creativity in a subject line can be the difference between someone opening a communication item or simply leaving it unread. If we can increase open rates, we’re getting that information out to more people.”

Initially, they wrote their own subject lines and then used Copilot to generate additional options. This A/B approach allowed them to compare the performance of AI-generated subject lines with their own. The results were impressive, with AI-written subject lines performing equally well, and sometimes even better, in terms of how many emails were opened.

Using Copilot to create subject lines not only injects creativity into communications, it also increases the efficiency of the team and removes some of the drudgery from content creation. Copilot’s ability to quickly generate multiple subject line options saved them a significant amount of time. Instead of spending time brainstorming and refining subject lines, they could rely on Copilot to provide a list of options, which they could then fine-tune as needed. It enabled the team to create more engaging content faster and provided them with time to focus on the bigger picture, including strategy and creativity.

Creating great imagery with Copilot and Microsoft Designer

Both Copilot and Designer offer AI-based image creation capabilities that the team has been using to create visuals for email messages, Teams channel posts, PowerPoint presentations, and Viva Engage posts.

“Copilot Chat and Designer have been absolute game-changers for us,” Etchells says. “They’ve given us the tools we need to create compelling visuals in our communications. The images look great, they’re consistent, and we can quickly create assets that support the story we’re trying to tell.”

The team found that they could use Designer to create a variety of assets that looked cohesive and were part of the same design family. Designer’s capability to generate multiple assets for different formats and functions was particularly useful for their campaigns. By using detailed prompts, they could describe the style, colors, and subjects they wanted in their images, ensuring that the generated content aligned with the company’s brand.

For example, this prompt produced the butterfly image that follows. The team used PowerPoint to add the text, logo, and blur.

Here’s the prompt they used: An urban brick wall painted white with rustic and edgy, simple, and bold street art painted on it. Include large butterfly wings with interesting patterns and shapes flowing out of it across the wall. The colors should be bright and bold: purple, hot pink, acid green, yellow, orange, blues, and teal, with a few black accents.

The example images provided by Designer, accompanied by the prompts that produced the images, were a big hit with the team, along with Designer’s post-creation editing tools and simple-to-use interface.

The team quickly found that image creation—just like the other AI processes they’ve adopted—is highly dependent on prompt quality. They’ve developed the skills and knowledge to learn how to guide the AI in terms of style, color, and composition. By specifying these details, they ensure that the generated images align with brand guidelines and requirements.

Using Copilot to perform sentiment analysis

The team is using AI for sentiment analysis to gain valuable insights into their communications. They collect feedback from various sources, such as Viva Engage, internal portals, and feedback forms and run the data through Copilot to analyze the sentiment behind the responses.

“Typically, the big issue with sentiment analysis is the amount of data that we have to go through,” Oxford says. “Copilot has given us the ability to take that data and apply a human-like perspective and decision-making capability to analyze sentiment. Taking all that qualitative information and turning it into quantitative information is huge because it’s a heavy lift from a human process perspective, but Copilot does it in a few seconds.”

Here are few examples of how the team is prompting Copilot to analyze sentiment:

• Thematic analysis: “Analyze the following feedback and provide the top 5 themes, by volume.”
• Granular analysis: “I’d like a detailed sentiment analysis of this text. Can you categorize it as very positive, positive, neutral, negative, or very negative?”
• Emotion detection: “What specific emotions are expressed in this text? Can you identify emotions like happiness, sadness, anger, etc.?” or “Can you score this based on the 9 emotions?”
• Net sentiment score: “Using the net sentiment analysis technique, analyze the sentiment of the following text. Net sentiment is calculated as the difference between the percentage of positive responses and the percentage of negative responses.”
• Specific analysis: “Analyze the email content below to evaluate its clarity and overall sentiment. Provide feedback and suggestions on how to improve the clarity of key points and maintain a positive tone. Also provide me with the possible sentiment that an employee might feel when reading this email based on the 9 different emotions.”

The team is also using sentiment analysis to improve their communication strategies. AI helps them create more effective and engaging content by tracking the sentiment of responses to different types of communications and enabling them to adjust their approach based on the results.

Sentiment analysis also finds its way into their reporting. By pulling the quantifiable results from sentiment analysis into reports, they’re presenting clear and actionable insights to leaders and stakeholders. Using sentiment analysis ensures that their communications are aligned with the company’s goals and the audience’s needs and preferences.

Fostering a culture of learning and growth

Learning and growing is a constant with AI tools and the team is continuing to champion effective AI usage throughout Microsoft. AI adoption practices are springing up across the company.  A learning series called “First Fridays” showcases product and communications leaders in monthly sessions, happening on the first Friday of each month, to help communicators at Microsoft learn more about how to:

• Use AI and Copilot in their daily work.
• Gain insights into how other communications leaders and organizations are growing AI skills.
• Partner with Microsoft Digital to communicate, adopt, and use AI.
• Understand current trends, tools, and innovations in the communications field.
• Understand how AI is changing the way people work.

A communications-focused prompt library in Copilot Labs has been created to help others in the company use Copilot effectively in communications. It offers practical suggestions to try in the context of the Copilot interface in Microsoft 365 apps. Copilot Labs enables users to share these prompts with others, increasing the reach of the communications team’s guidance and expertise.

Looking forward

As AI technology continues to evolve, the team knows that it will play an even more integral role in enhancing the efficiency, creativity, and effectiveness of internal communications.

“We’ll always be learning and growing,” McCarty says. “We have a forward-looking culture and that allows our team—and our company—to pivot pretty quickly. We’re excited about how the evolution and growth of AI capabilities will move our team forward and give us the opportunity to collectively learn and grow together.”

The team is expecting AI to make a significant impact in the automation of routine tasks in the near future with Copilot Agents. By automating repetitive and time-consuming tasks, AI will free up communicators to focus on more strategic and creative aspects of their work. This shift will allow the team to dedicate more time to crafting compelling messages and developing innovative communication strategies. The team is already using agents to quickly create communications plans based on their established template. For example, the team has created an Agent to draft blog posts for certain communications channels, aligned to the messaging, tone, and other aspects established for the program.

Copilot’s data analysis capabilities are continually improving, with deeper integration into data analysis tools such as Microsoft Excel and Microsoft Power BI. With more advanced AI tools, the team will be able to better process and assess large volumes of data to gain valuable insights into employee sentiment, engagement levels, and communication effectiveness.

The future of using AI on the internal communications team and across Microsoft is bright and full of potential. As AI technology continues to advance in the future, its role in shaping the future of internal communications will only become more significant, and that’s a future that the IT Communications team looks forward to.

Looking to inject AI into your team’s content creation? Try these suggestions:

• Embrace AI as a partner. Integrate AI tools like Microsoft 365 Copilot into daily tasks to enhance content quality and save time.
• Foster a learning culture. Encourage a transparent culture of learning and growth within your team to effectively use AI for creative work.
• Prioritize effective prompting. Develop effective prompting skills to maximize the quality of AI-generated content.
• Use AI for sentiment analysis. Use AI to understand the sentiment of posts and responses, which helps tailor communications to better meet audience needs.
• Take advantage of AI to boost creativity. Use AI tools like Microsoft Designer to create compelling visuals and enhance the overall quality of communications.
• Focus on practical applications. Apply AI in practical ways, such as generating subject lines, managing repetitive tasks, and brainstorming sessions.
• Create agents for process improvement. Agents, created in minutes with the agent builder in Copilot Chat, can automate tasks and processes that you do often, from creating communications plans to writing blogs and promotional content.

Start your Microsoft 365 Copilot journey.

• Learn how we’re transforming our internal communications at Microsoft with Viva Amplify and Viva Engage.
• Check out how our IT Communications Team is using Microsoft Designer to revolutionize the way they create visual content.
• Find out how our Copilot Champs Community is leading our effort to get more out of Microsoft 365 Copilot internally here at Microsoft.
• Learn more about how you can use Microsoft 365 Copilot to transform the way you work.

• Want more information? Email us and include a link to this story and we’ll get back to you.

The post Elevating internal communications at Microsoft with AI appeared first on Inside Track Blog.

Are you interested in creating cool, custom visuals by typing in a few quick prompts? Let us show you how we do it with AI at Microsoft Digital, the company’s internal IT organization.

Sarah Lundy and Eva Etchells are senior content program managers and part of the Microsoft Digital IT Communications team. They ensure that Microsoft’s 336,000 employees and vendors have the information and encouragement they need to successfully adopt and embrace technology change.

“As a comms team, we typically don’t have access to a graphic designer,” says Lundy, a senior content program manager. “Prior to AI, we were doing our own visuals by taking brand assets, stock imagery, and brand templates, and trying to do our best.”

Enter Microsoft 365 Copilot Chat and Microsoft Designer. Copilot Chat is our AI-powered tool that integrates with your work data and applications. It helps you manage tasks, find information, and create content by leveraging data from emails, files, meetings, and more. Copilot Chat is accessible through the Microsoft 365 app, Teams, Outlook, and the web, and it provides a conversational interface to interact with your data. Microsoft Designer is our AI-powered graphic design tool that helps you create, design, and edit images based on your descriptions.

Both Copilot Chat and Designer offer AI-based image creation capabilities that Lundy and Etchells have been using to create visuals for their team.

“Copilot Chat and Designer have been absolute game-changers for us,” Etchells says. “They’ve given us the tools we need to create compelling visuals in our communications. The images look great, they’re consistent, and we can quickly create assets that support the story we’re trying to tell.”

Making the most of AI image creation

So how can you and your team use AI to create amazing visuals? Here are a few tips from our Microsoft Digital IT Communications team.

Learn as you go

The IT Communications team started their journey with image creation in Copilot Chat, using the conversational engine to generate images in the context of the other AI-fueled work they were doing.

As Designer matured since it launched in 2022, the team began to favor its larger and more detailed image creation toolset.

“Designer provided a combination of flexibility and inspiration that changed the way we approached AI image creation,” Lundy says.

The team found that Designer allowed them to create a variety of assets that looked cohesive and were part of the same design family. Designer’s capability to generate multiple assets for different formats and functions was particularly useful for their campaigns. By using detailed prompts, they could describe the style, colors, and subjects they wanted in their images, ensuring that the generated content aligned with their brand guidelines.

Example images, accompanied by the prompts that produced the images, were a big hit with the team, along with Designer’s post-creation editing tools and simple to use interface.

Use examples for inspiration

There are tools you can use to help you get started.

“Starting with a blank canvas is one of the most difficult things for me,” Etchells says. “I love the way Designer provides multiple ways to get started with an image.”

You can use the example content from any of the image categories in Designer to inspire your next creation. By hovering over any of the example images, you can see the prompt used to generate the image. Example prompts can provide new approaches for describing colors, shapes, framing, art styles, and more in your visuals.

“I can take bits and pieces from this prompt and then create my own prompt,” Etchells says. “That’s where it started to get really powerful, because it was difficult to start with nothing. Seeing the prompts used to create other images I liked unlocked a new level of effectiveness. Seeing how the prompt was written and then being able to change it to suit our needs was a game-changer for me.”

Designer’s large library of image types including clipart, emojis, icons, monograms, avatars, stickers, and wallpapers offers a wide variety of approaches to creating visuals for any use case.

Reuse prompts for themed results

“Reusing the elements of a prompt to create another image with a similar theme or look is a great way to create consistency,” Lundy says. “We take a prompt that we’ve had success with and adjust the subject to get a bunch of assets that look like they’re part of the same design family.”

Reusing prompts helps create a starting point for Copilot Chat or Designer to create visual elements consistently.

For example, for a recent Copilot Agent campaign (here’s where to go learn more about Copilot Agents and to try building one yourself), the team used a secret agent theme. The following prompt was used to create the image on the left that follows:

In the style of a chunky block print vintage spy movie poster that is a bit kitschy, a silhouette of a friendly, likeable basset hound as a secret agent in a fedora and trench coat looking out from one side of the image. There is a city skyline in the background. Pastel theme blues, purples, pink colors, with rich layers.

And another, with a couple changes, represented by the image on the right:

In the style of a simple chunky block print vintage spy movie poster that is a bit kitschy, a silhouette of a striped, friendly cat holding an envelope as a secret agent in a fedora and trench coat looking out from one side of the image. There is a city skyline in the background. Pastel theme blues, purples, pink colors, with rich layers.

The result is two images, built for the same campaign, using a common theme.

Lean into your strengths

The best way to get the most out of Copilot Chat and Designer is to start with what you know best.

“AI does some aspects of image creation so well, but it does have limitations,” Lundy says.

The prompt-based image creation process allows for verbose descriptions and inclusion of many details.  The team learned that the details they include and how they prompt the AI for image creation are important. Treating AI like a graphic designer works much better than treating it like a graphic design tool.

“We try to explain the image like you would to a person,” Etchells says. “Providing a descriptive, informative, and thematic prompt can produce amazing and creative results. However, trying to be too precise and telling the AI exactly how you want the image created in terms of shapes, positioning, and measured dimensions is typically less effective.”

Here’s a list of tips that will help you use the strengths of AI image creation and avoid its weaknesses.

• Use AI to create ideas, concepts, and inspiration for your final work. AI results offer creativity and new ways of approaching visuals.
• Use AI to create specific scenes or images that are easy to describe.
• Use descriptions of style and subject matter. You can get specific in the elements you want included in the image, but you can’t control how Designer brings them all together.
• Use other tools, like PowerPoint, to add and manipulate text. AI isn’t great at creating complex text in images. It will often misspell words!

Hone your prompting skills

Like all your interactions with AI and Copilot, it’s all about what you ask for and how you ask for it.

“The skill to develop in all this is prompting,” Etchells says. “It’s all about knowing what to tell the AI and how to phrase your prompt to get the results you’re looking for.”

Prompting skills are essential when working with AI tools to create images because they serve as the foundation for generating high-quality and relevant visuals. With a detailed and well-thought-out prompt, the AI can better understand your vision and produce images that closely match your expectations.

“You really start to learn how to guide the AI in terms of style, color, and composition,” Lundy says. “By specifying these details, you can ensure that the generated images align with your brand and requirements. For example, if you need an image with a specific color palette or a particular artistic style, including these details in your prompt will help the AI create something that fits your needs.”

Hone your prompting skills for image creation with these tips:

• Clearly describe the style, colors, and subjects you want in your image. The more descriptive your prompt, the better the AI can understand and generate the desired result.
• Be patient and refine your prompt through multiple iterations to achieve the desired result.
• Focus on what you want in the image. Negative prompts or telling AI what you don’t want can sometimes lead to unwanted elements being included.
• Maintain a list of images and styles you like, including the prompts that you used to generate them. This can serve as inspiration or thematic basis for future prompts.

Be bri11iant today

The IT Communications team launched the “Be bri11iant today” campaign to help drive daily usage of Microsoft 365 Copilot internally at Microsoft.

“Our research shows that saving 11 minutes a day is the sweet spot where you start to see Copilot becoming a habit, a part of your daily workflow,” Etchells says.

The team wrote a blog post and created a series of promotions that used images like the one shown below.

Here is the prompt used to generate the Alpine scene:

Create an image using dimensional felt craft, side flat view, of a mountain lake scene at sunrise a chair and table with an open laptop and a cup of steaming coffee on it, surrounded by evergreen trees, ferns, mushrooms, rocks, and a canoe on the shore, with fluffy clouds, a rainbow-colored hot air balloon, and birds in the sky. Create the image in a children’s book style, light pastel theme, blues, purples, pink colors, with rich layers, high detail, 8k, high quality.

Here’s the prompt used to create 11 icon:

Create an image using dimensional felt craft, side flat view, of a tablet with the number 11 and evergreen trees, ferns, mushrooms. Create the image in a children’s book style, pastel blues, pastel purples, pastel oranges, and pinks, with rich layers, high detail, 8k, high quality.”

The team then animated the image in Microsoft PowerPoint. These were the steps for that:

1. Create all the visual elements in Copilot using a consistent prompt for style, colors, etc.
2. Insert them into PowerPoint:
   • Remove backgrounds and add drop shadows in PowerPoint
   • Add text in PowerPoint
   • Add animations in PowerPoint
3. Save the slides as a GIF.

Here are some tips for getting started with creating images in Copilot Chat and Designer at your company:

• Use examples for inspiration. Drawing from a collection of past successful images can spark new ideas and provide a template for what you aim to achieve.
• Reuse prompts for themed results. By consistently using similar prompts, you can maintain a cohesive style and theme across multiple projects.
• Lean into the strengths of AI image generation. Use the advanced capabilities of AI to create complex and detailed visuals that might be difficult to achieve manually.
• Hone your prompting skills. Continuously refine your ability to craft precise and effective prompts to enhance the quality and accuracy of generated images.

Get started with Microsoft Designer at your company.

• Learn more about Microsoft Designer.
• Go here to get answers on commonly asked questions about Microsoft Designer.
• Learn how to get started with Microsoft 365 Copilot based on our internal experience here at Microsoft.
• Check out our Top 10 tips for using Microsoft 365 Copilot every day.
• Find out how we’re getting more out of Microsoft 365 Copilot with purposeful prompts.

• Want more information? Email us and include a link to this story and we’ll get back to you.

The post Transforming communications with AI: How we’re using Microsoft Designer to revolutionize our visual content appeared first on Inside Track Blog.