D-ID enables 100x growth and scales its digital avatars globally with Azure AI Services

Gil Perry grew up around computers and graphic user interfaces (UIs). He studied computer science at Tel Aviv University and got into generative AI as early as 2014. “Inspired by both a personal pain point and curiosity, I invented one of the first algorithms in the world to protect people from the misuse of face-recognition technology,” shares Gil Perry, now the Chief Executive Officer and Co-founder at D-ID. 

D-ID, founded in 2017 as a privacy solution, soon became something bigger. Perry and his team realized that the same underlying technology could be repurposed—not to hide faces, but to bring them to life. “We saw that voice cloning and digital humans would soon be able to move, speak, and interact—and imagined them becoming the new interface for AI,” Perry adds. “With our strong roots in privacy and security, and guidance from top experts, we knew we could lead this space in a responsible way.”

Digital humans as the UI for AI

D-ID is designed to let companies turn text, images, video, and voice into interactive, hyper-realistic avatars. The visual component can be anything—from a CEO’s digital twin to a fictional animated character. “Essentially, we’re revolutionizing AI-powered interactions by making digital humans more lifelike and engaging,” shares Perry. These visual interfaces are transforming industries like sales, marketing, and internal communications, helping companies scale interactions. 

Because D-ID works with digital identities, it’s critical to prevent abuse and build trust. Originally built on another platform, their solution needed stronger security, scalability, and integration. That’s why the company chose Microsoft—a reliable technology partner with a strong commitment to security. “Microsoft's strong commitment to ethical and responsible AI practices was a key factor for us," Perry recalls. "Microsoft’s leadership in generative AI, combined with its deep integration with OpenAI, made it a natural fit.

Once the partnership kicked off, support from senior Microsoft executives sealed the deal for D-ID. “We experienced something rare with Microsoft: genuine, high-level executive engagement. They saw our vision early on and made it clear this was a strategic partnership—something we hadn’t encountered with any other provider.”

Built highly securely from input to output

As the first step, D-ID migrated its solution to Microsoft Azure. “Azure provides seamless AI integration, fast, secure, and scalable—which accelerates our time to market and optimizes performance,” Perry explains. 

The company then introduced Microsoft solutions across every stage of its workflow to improve security further. Eli Cohen, Vice President of Product at D-ID, explains the details: “We take voice as an input and translate it to text using Azure AI Services’ Speech-to-Text service.” Once the platform understands what was said, it refines the answers using Azure Open AI Service. The platform also uses Microsoft security and moderation tools to help flag inappropriate content.

The next step is bringing that response to life. “The text then goes again through Azure’s Text-to-Speech service in order to get it back as voice again,” adds Cohen. Once that’s done, video synthesis generates the avatar’s visuals. For speed, D-ID uses Azure Kubernetes Services to accelerate processing for real-time, high-quality AI avatars.

Finally, Azure API Management helps D-ID ensure that everything integrates smoothly with their customers’ applications. By offering their solution through the Azure Marketplace, D-ID makes it easy for a wide range of users to get started—whether it’s enterprises looking for a digital human interface, teams producing personalized or multilingual video content at scale, or developers building large language model (LLM) powered applications who want to add a visual layer with just a few lines of code. As Perry puts it, “Customers can easily access our API and upgrade their product with a more human, engaging interface.”

Avatars available around the clock

One of the most common use cases for D-ID's avatars is customer support. “Instead of browsing a website, clicking through pages, or watching static videos, users can simply speak with an avatar and instantly get what they need,” Cohen reflects.” It transforms the interface into something intuitive and human. It also scales effortlessly—avatars are available 24/7.” By now, companies have created over 150,000 D-ID Visual agents, which have in turn sent 1,800,000 messages and spent 340,000 minutes interacting with customers. 

These avatars can also qualify leads in real time and in different languages—capturing names, emails, phone numbers, and pushing them into a customer relationship management (CRM) software. D-ID’s other common applications are also in investor relations and public appearances. Speakers can prepare questions and answers in advance, then hand it off to their avatar to represent them by responding naturally and engaging the audience. “Before, I would travel and pitch to 100 investors. Now, with my avatar, I can reach 1,000, and do so more efficiently,” Perry shares.” Sometimes, using an avatar can be more convenient for certain tasks. We've seen high-level executives create avatars for interviews, public appearances, and even book launches when they don't have the time or energy to attend in person."

Consumer brands are also experimenting with where avatars live—from conferences and drive-throughs to kiosks, airplanes, and ATMs. “With advancements in AI and digital avatars, we might soon see avatars integrated into various everyday items, including barcodes on bottles,” Perry remarks. “PepsiCo created a D-ID avatar that was trained as a hydration consultant and showcased it during Cannes Festival of Innovation.”

Supporting expression through avatars

D-ID can also be used as a voice for people with impairments. “Hundreds of millions of people experience speech disabilities,” shares Perry. In many cases, these individuals use assistive technologies, like eye-tracking or finger-based input, to communicate. But these tools are often slow and cumbersome. “With our solution, they can have a visual, speaking avatar that represents them, letting them reclaim their voice and visibility—restoring self-expression, confidence, and inclusion,” Cohen adds. 

The same avatars can also help in sectors where the availability of professional services is limited. “In industries where there is a shortage of skilled personnel, such as teachers or therapists in certain geographies, a D-ID visual agent can be trained to meet those needs. It can even interact in multiple languages and offer support and engagement to underserved communities,” Cohen elaborates.

For instance, D-ID is being used as the first line of support for the American Alzheimer’s Foundation. People can go on their website and interact with D-ID-powered visual agent to get basic support. “If, on the other hand, they need professional help, the AI agent routes them to the human agents,” shares Cohen.

In line with the company’s vision of using technology for good, D-ID has developed Sophia—the world’s first interactive chatbot designed to assist victims of domestic violence—in partnership with Microsoft. Sophia provides multilingual support, offering critical information, guidance, and emotional assistance to survivors worldwide. The chatbot ensures anonymity, enabling individuals in distress to seek help and access resources without fear of exposure. “By combining AI-driven empathy with Microsoft’s secure and scalable cloud infrastructure, D-ID and Microsoft are helping create a safer and more supportive digital environment for those in need,” shares Perry.

Discover more about D-ID on Facebook, Instagram, LinkedIn, X/Twitter, and YouTube.