From Hot Wheels to Content Management: How Brands Are Using Microsoft AI to Be More Productive and Imaginative

For example, TaylorMade Golf Company turned to Microsoft Syntex for a comprehensive document management system to organize and secure emails, attachments, and other documents for intellectual property and patent filings. . At the time, the company’s lawyers manually managed this content, spending hours filing and moving documents for sharing and further processing.

With Microsoft Syntex, these documents are automatically categorized, tagged and filtered more securely and make them easy to find through search instead of having to dig through a traditional file and folder system. TaylorMade is also exploring ways to use Microsoft Syntex to automatically process orders, receipts and other transactional documents for Accounts Payable and Finance teams.

Other customers use Microsoft Syntex for contract management and assembly, Teper noted. Although each contract may have unique elements, they are built with common clauses regarding financial terms, change control, schedule, etc. Rather than writing these common clauses from scratch every time, people can use Syntex to assemble them from various documents and then introduce changes.

“They need AI and machine learning to spot, ‘Hey, this paragraph is very different from our standard terms. That might need some extra monitoring,’ he said.

“If you’re trying to read a 100-page contract and look up the thing that’s changed significantly, that’s a lot of work versus the AI ​​helping with that,” he added. “And then there is the workflow around these contracts: who approves them? Where are they stored? How to find them later? A lot of that is metadata.

When DALL∙E 2 becomes personal

The availability of DALL∙E 2 in Azure OpenAI Service sparked a series of explorations at RTL Deutschland, Germany’s largest private media company, on how to generate personalized images based on customer interests. For example, in RTL’s Data, Research and AI Competence Center, data scientists are testing various strategies to improve user experience through generative imaging.

RTL Deutschland’s RTL+ streaming service is expanding to offer on-demand access to millions of videos, music albums, podcasts, audiobooks and e-magazines. The platform relies heavily on images to grab people’s attention, said Marc Egger, senior vice president of data products and technology for RTL’s data team.

“Even if you have the perfect recommendation, you still don’t know if the user will click on it, because the user uses visual cues to decide if they are interested in consuming something. So artwork is very important and you have to have the right artwork for the right person,” he said.

Imagine a romantic comedy about a professional footballer who is transferred to Paris and falls in love with a French sports journalist. A sports fan might be more inclined to watch the movie if there is an image of a football game. Someone who loves romance novels or travel might be more interested in an image of the couple kissing under the Eiffel Tower.

Combining the power of DALL∙E 2 with metadata about the type of content a user has interacted with in the past provides the ability to offer personalized imagery at a scale previously inconceivable, Egger said.

“If you have millions of users and millions of assets, you have the problem that you just can’t scale it — the workforce doesn’t exist,” he said. “You would never have enough designers to create all the custom images you want. So it’s an enabling technology to do things you couldn’t do otherwise. »

The Egger team is also thinking about how to use DALL∙E 2 in Azure OpenAI Service to create visuals for content that currently lacks images, like podcast episodes and audiobook scenes. For example, a podcast episode’s metadata could be used to generate a single image to accompany it, rather than repeating the same generic podcast image over and over.

RTL Deutchland, Germany’s largest privately owned cross-media company, is exploring how to use DALL∙E 2 in Azure OpenAI Service to engage people browsing its RTL+ streaming service. One idea is to use DALL∙E 2 to generate unique images to illustrate individual podcast episodes, rather than relying on the same podcast cover.

Along the same lines, someone listening to an audiobook on their phone is usually looking at the same book cover for each chapter. DALL∙E 2 could be used to generate a unique image to accompany each scene in each chapter.

Using DALL∙E 2 through Azure OpenAI Service, Egger added, provides access to other Azure services and tools in one place, allowing his team to work efficiently and seamlessly. “As with all other software-as-a-service products, we can be sure that if we need massive amounts of images created by DALL∙E, we’re not worried about having them online.”

Appropriate and responsible use of DALL∙E 2

No AI technology has generated as much excitement as systems like DALL∙E 2 that can generate images from natural language descriptions, according to Sarah Bird, Microsoft Group Senior Project Manager for Azure AI.

“People love images, and for someone like me who isn’t visually artistic at all, I’m able to do something much more beautiful than I could ever do using other visual tools” , she said of DALL∙E 2. “It gives humans a new tool to express themselves creatively and communicate in a compelling, fun and engaging way.

His team is focused on developing tools and techniques that guide people towards the appropriate and responsible use of AI tools such as DALL∙E 2 in Azure AI and that limit their use in ways that could cause damage.

To help prevent DALL∙E 2 from providing inappropriate outputs in Azure OpenAI Service, OpenAI removed the most explicit sexual and violent content from the dataset used to train the model, and Azure AI deployed filters to reject prompts that violate the content policy.

Additionally, the team has incorporated techniques that prevent DALL∙E 2 from creating images of celebrities as well as commonly used objects to try to trick the system into generating sexual or violent content. On the output side, the team has added templates that remove AI-generated images that appear to contain adult, gory, and other types of inappropriate content.