Rethinking content creation Azure AI-powered track story bot

At Microsoft Digital, the company’s IT organization, innovation is the fuel that drives us.

To stay ahead, we are continually exploring ways to use technology to promote efficiency and value. One way we’re innovating is to use generator AI to accelerate content development for Inside Track blogs.

Using generative AI, we set out on a journey that revolutionizes how we interact with our customers through compelling stories and case studies, even though humans are always in the loop. In fact, this article was written with the help of artificial intelligence.

Vision and challenge

The vision for this project was clear. Use AI to accelerate content creation, reduce turnaround times, and enable Microsoft Digital team members to share stories with customers quickly. However, the path to realizing this vision was not without its challenges.

“We were using AI in our childhood with limited budgets and low resources to challenge unknown territory,” said the leading product manager of the frictionless device team that led the initiative. says Dwight D. Jones Sr.

To overcome these challenges, the team created a pitch deck based on their vision, built detailed technical specifications, and worked with stakeholders to build support for the project. They founded a virtual team with Microsoft Digital Inside Track Team, Frictionless Devices Team, and employee productivity engineering teams. With approval from the Microsoft Digital Leadership, they began to proof of a small concept that proved successful. In the pilot, the team was able to use an AI-powered interview bot to create a great first draft of blog posts, saving a lot of time per story.

Accelerate publications and reduce costs

Senior Director Keith Boyd told more stories with lower overall costs and highlighted the double challenge of accelerating publications.

“The key technology that powers bots is the openness of the navy blue ure,” Boyd says. “We’re already helping us by enabling Microsoft Digital subject experts to share their expertise on their own schedules.”

By using bots, the team wants to increase the number of stories in the pipeline and reduce the time it takes to publish.

“Our main indicator is when it’s public,” Boyd says. “Does the bot help you move from a six-week cycle to create a story?

Streamline the process

Revanth Chandra Pydimarri, Senior Product Manager for Inside Track Content Bot V-Team, highlighted the need to streamline the drafting and publishing process.

“We were wasting resources through inefficient operational processes. We reached out to experts on related subjects, got content from them, wrote stories, and finally made it public. It took a lot of time,” says Pydimarri.

The team turns to Microsoft Azure AI and builds bots that use generator AI features including searched generation (RAG) and Azure Openai.

“We trained the bots in the style of an internal track,” Pidimali says. The results were impressive, saving at least 5 hours of time per story while maintaining publishing standards.

Capture input and maintain style

Urvi Sengar focused on training bots on how to write stories in the same look and feel as human-created stories, and how to use BOT to capture input from small businesses and turn them into content . Bots are now able to emulate the storytelling style of Inside Track teams by being trained using a comprehensive dataset of previously published stories and writing guidance provided by the editorial team.

“BOT’s sophisticated content summary and curation techniques allow us to quickly draft high-quality stories and maintain the subtle details and tones that small businesses offer,” says the Microsoft Digital team Sengar, a senior software engineer, said:

Despite the challenge of transferring subject expertise and repeatedly improving the process based on user feedback, the project has been well received by stakeholders.

“Bots can now generate 70% or more drafts, significantly reducing the time they need to create stories,” Senger says.

The team is adding the ability for BOT to verbally interview SMEs by integrating Voicerag using Azure AI Search for Audio to Text Fafability. This allows bots to accurately and efficiently transcribe audio interviews, streamline the content collection phase, and create richer, multidimensional stories through a more natural interview process.

Combine multiple perspectives

Lukas Velush, managing editor at Inside Track, used AI-powered bots to interview multiple SMEs and addressed the challenge of combining the interviews into a single, coherent story.

“Our challenge was to continue talking about subject matter experts across Microsoft Digital despite small teams and budgets being reduced,” Velush said.

This downward pressure on resources is the inspiration for bots and can help them create stories at a faster pace at a lower cost.

The bots are designed to emulate human writers and interview small businesses about IT jobs, challenges and how they overcome them.

“One of the biggest challenges we’re still working on is to make the bots get to interview multiple people for the story, then weaving in what we have to say in a consistent story.” says Velush.

The current solution is to interview each person individually with a bot interview and combine the resulting story into one comprehensive story.

The initial results are promising, with stories written in bots costing 50% less than stories written by humans and completed at least 30% faster.

“It’s been an incredible journey. I learned about AI and humans and what it takes to report and write stories like this,” Velush says. “I think Bots are great helpers who can work with our human writers.”

The future of AI-driven content creation

As Microsoft Digital innovates and pushes boundaries, Inside Track Bot changes the way stories are told as evidence of the power of AI. Using AI, teams have shown that they can accelerate their content creation process, allowing them to deliver high quality, engaging content faster and more efficiently than before.

In the future, the team plans to further enhance the bot’s functionality. Jones shared that the team plans to use AI to create multiple types of content in a single entry, including articles, blog posts, PowerPoint decks, and white papers. It also plans to generate content in multiple languages ​​to support readers with accessibility needs.

Boyd emphasized the possibility that AI can enhance human abilities.

“One goal is to reassure other content teams that they’re going to increase their capabilities and increase productivity, rather than replacing AI-powered bots here,” he says.

This is just the beginning.

“This is a V1,” says Pydimarri. “We also plan to add more features and expand functionality to meet other team requirements.”

Future extensions will focus on expanding bot features and capabilities. “This will continue to meet the evolving needs of the inside track team and other users,” Senger says.

I’m looking for the future

Demonstrate your team’s innovative spirit and commitment to customer engagement by using AI to accelerate Microsoft Digital’s content creation.

“This project is just the beginning,” says Jones. “We are excited by the potential to revolutionize how AI engages with our customers and look forward to where this journey will take us.”

Internal trackbots demonstrate the potential for AI to streamline processes, improve efficiency and provide a high-quality, engaging storyline that resonates with the audience. This is a great example of digital transformational action that will pave the way for future innovation.

I encourage you to take part in this journey as you continue to push the boundaries of what is possible with AI. Find out the capabilities of tools and technologies available through Microsoft Azure AI in your own content creation process and see first-hand how you can improve productivity and storytelling quality.

We use AI to accelerate content generation with Microsoft. Here are some highlights from our journey:

AI-powered content creation: Microsoft Digital uses generated AI to automate content creation for Inside Track blogs, significantly reducing turnaround time. Pilot Success: The AI-powered chatbot has been successful with the pilot, creating a great first draft of blog posts, saving a considerable amount of time per story. Efficiency and Cost Reduction: The initiative aims to help accelerate publications by telling more stories, more stories at overall cost. Streamlined Process: AI BOT eliminates the need to interview subject experts (SMEs) when convenient, adjusting the schedule, saving at least 5 hours per story. Maintaining Quality: The bot is trained to emulate the storytelling style of internal track teams, capturing the subtle details and tones that small businesses offer. Added audio: Integrates Azure AI Speech Studio for audio-to-text functionality. This makes Inside Track Content Bot a more natural experience for small businesses. Future Extensions: We are expanding the capabilities of BOT to support other Microsoft digital teams.