Digital Avatar Technology Enters Mainstream PR and Journalism

The Rise of Interactive AI Avatars

Synthesia, a video-generation startup valued at $4 billion, has pushed digital avatar technology beyond novelty into practical business applications. The company, which crossed $100 million in annual recurring revenue last year, operates alongside competitors like D-ID, HeyGen, and Colossyan in a rapidly expanding market. Synthesia's newest offerings include an interactive avatar of Alexandru Voica, the company's head of corporate affairs, trained to answer press questions about the startup's capabilities. The technology has evolved from simple script-reading avatars to sophisticated interactive systems capable of understanding and responding to user queries in real time.

How the Technology Works

Synthesia's interactive avatar system combines four distinct AI components into a seamless experience. Voice-to-text models convert spoken questions into written text, which an agentic language model then interprets and determines appropriate responses to. A text-to-voice system converts those responses back into audio, while Synthesia's proprietary video model animates the avatar's facial movements and lip-sync. The company offers flexibility in its tech stack, allowing enterprises to select voice and language models from providers like Cartesia, ElevenLabs, Google, or OpenAI. Customers can choose their own cloud hosting or use Synthesia's infrastructure.

Synthesia operates three distinct product tiers. The entry-level video-creation platform lets users script responses for avatars to deliver verbatim. The Sessions platform enables interactive experiences like surveys and roleplay simulations where employees can practice skills like sales pitches against responsive AI avatars. The API platform allows developers to integrate Synthesia's video and voice models with other services to build custom applications.

A Journalist's Experience with Digital Twins

A senior venture capital reporter at TechCrunch became the first journalist to receive a custom interactive avatar after visiting Synthesia's new New York office. The creation process required approximately two minutes of voice recording and numerous photographs taken in a mini studio setup. The company built four avatars: two personal versions that read scripts verbatim, with and without glasses, and two interactive versions capable of understanding and responding to questions, also in glasses and non-glasses variants.

The interactive avatars were trained on a single article about venture-backed startup fraud. When tested, the avatars demonstrated deterministic behavior—they would only answer questions related to the trained content and redirect all other inquiries back to the original story. When asked off-topic questions like where the journalist previously worked or which New York neighborhood they lived in, the avatars consistently steered conversations back to the fraud story. Non-technical observers found the personal avatars more convincing than the interactive versions, noting the interactive models' voices didn't quite capture the original's tone and facial likenesses appeared slightly off-model. However, the personal avatar successfully replicated the journalist's voice tone, even avoiding hoarseness that was present in the source recording.

Implications for Journalism and Corporate Communication

The technology raises substantial questions about the future of journalism and media. Some industry investors dismissed the notion of avatar-presented news entirely, while others acknowledged potential augmentation uses. The core tension centers on trust—a fundamental pillar of journalism that artificial systems appear incapable of replicating. A journalist's primary value proposition involves direct human connection, original research, and relationship-building with sources; these functions seem difficult to delegate to deterministic AI systems.

Beyond journalism, avatar technology presents compelling use cases for corporate communications. Business leaders could deploy digital versions of themselves to handle routine inquiries, eliminating post-vacation email backlogs and providing always-available employee support. However, the technology also raises concerns about uncanny valley effects and potential psychological discomfort when systems behave unpredictably.

Looking Forward

The journalist predicted that Generation Z—those who grew up with science fiction depicting AI—will likely find persistent discomfort with digital twins, even as the technology becomes technically polished. The experience of interacting with these avatars feels inherently sci-fi rather than naturalistic. Yet the reporter noted that avatars feel less disturbing than physical humanoid robots, partly because users can simply log off if interactions become unsettling. The technology's deterministic nature provides a degree of safety that more autonomous AI systems might not offer. As Synthesia and competitors continue refining avatar technology, corporate adoption will likely accelerate, particularly for customer service, employee training, and internal communications roles.