On June 15, 2025, Chinese entrepreneur Luo Yonghao made global headlines by hosting a six-hour livestream that was entirely fronted by his AI twin avatar. I am amazed and impressed by this that it went well without a hitch for 6 hours! The virtual event was powered by Baidu’s ERNIE large language model. It was streamed via the e-commerce platform Youxuan.

The results? Over 13 million viewers and generated approximately $7.6 million USD in sales. For marketers both online and offline, and corporate communication professionals, this event is a game changer!

This post will explore what I see as the implications for the future of marketing and sales from Luo Yonghao’s digital avatar event success. This might be a bridge too far for small companies but mid to large corporations can get involved and test and adopt this technology to elevate engagement, scale communications, and innovate more engaging customer experiences.

This post is available as a podcast:

What Happened: A Breakdown of the Event

Luo Yonghao’s AI avatar interacted, reacted, and appeared to host the event in real time. This is the biggest element of change in my opinion. Key highlights:

• Powered by AI: Built on Baidu’s ERNIE model, the avatar had access to a knowledge base of 13,000 items and delivered over 97,000 words of dialogue.
  
• Human-like Movements: Over 8,000 coordinated avatar gestures made the AI appear lifelike and expressive. When I have made avatars I had very limited movement compared to what I see in that event.
  
• Sales Impact: The livestream outperformed Luo’s previous human-hosted events, setting new records in viewership and revenue. This is very impressive. Who said avatars could not help to generate revenue?

Why This Matters to Marketers

The implications for digital marketing are big. Now we know that avatars can:

• Scale content without burnout: A digital avatar can work 24/7, across time zones. Wow!
• Ensure brand consistency: Scripted yet adaptable avatars can ensure on-message communications.
• Cut through digital fatigue: Human-like avatars are still seen as magical, and can boost engagement rates.
• Enhance personalisation: AI avatars can theoretically respond to individuals. I tried this out a few weeks ago at HeyGen. It is fun to talk to an Avatar and it answer my questions. I use HeyGen to make my AI avatars.

For mid to large corporations, this opens the door to:

• Hosting multiple product launches simultaneously.
• Creating regionalized brand ambassadors without extra headcount.
• Running evergreen training or onboarding sessions.

Potential Applications for Corporate Marketing & Communications

Here are realistic use cases where AI avatars could be implemented in the next year or two:

• Product Demonstrations: Avatars can walk users through complex products in real time, explaining features dynamically based on customer questions that staff have picked out.
• Event Hosting: Think webinars, internal town halls, or hybrid conference Master of Ceremonies.
• Customer Support: A branded avatar at the front end of a chatbot, creating a more immersive experience is a real possibility.
• Retail & E-commerce: In-store screens or virtual storefronts featuring avatar hosts to guide purchases or explain promotions. Sounds useful?
• Training & HR Onboarding: AI avatars delivering consistent onboarding or learning and development sessions across global offices.

It is not about replacing people. More about reducing strain and pressure. Plus increasing availability and brand reach.

Real-Time Responsiveness: Where Are We Now?

Luo’s avatar felt live, but it was not fully unscripted. It required:

• 97,000 words of scripted dialogue
• 8,300 programmed avatar gestures

That’s a massive investment in planning and production.

Limitations Today:

• Real-time interaction is still limited by AI latency and unpredictable user input.
• Natural voice and facial sync often lags or feels robotic.
• Manual scripting and motion planning are not scalable.

What’s Realistic in 2025:

• Semi-live formats: AI avatars can use adaptive scripts with logic branches for predictable paths (e.g., FAQs, guided tutorials).
• Real-time Q&A with limited scope: Trained on a curated knowledge base, avatars can answer a wide array of preset customer queries.
• Voice-to-animation sync tools: These have improved, enabling better lip-sync and tone matching in avatars.

Overcoming the Scalability Barrier with AI Avatars

To move from scripted mega-productions to scalable everyday use, companies must:

• Use modular scripts: Break scripts into topic blocks, reusing them across contexts.
• Leverage real-time Large Language Model (LLMs): Here we are talking about tools like ChatGPT, Claude, or ERNIE to dynamically generate responses within safe parameters.
• Adopt avatar animation platforms: Use platforms like
  – Synthesia (studio-quality videos with AI avatars and voiceovers in 140+ languages).
  – Hour One (create, manage and streamline cinematic AI avatar videos), or
  – D-ID (They say you can “create visual AI agents to engage your audience)
  for automating avatar creation and updates.
• Focus on MVP (Minimum Viable Product) avatars: Not all avatars need to be hyper-realistic. Start with branded, stylized characters that are less uncanny but more efficient. This is how I learnt 😉

Future AI will increasingly support:

• Contextual memory
• Real-time animation generation
• Emotionally aware voice modulation

Once these mature, the need for thousands of gestures and word-for-word scripting will fade which I am looking forward to.

Strategic Considerations for Implementation of AI Avatars

Before jumping in, consider these planning steps:

• Define Use Cases: Start with high-impact, low-risk applications (e.g., training videos, FAQ avatars).
• Choose the Right Avatar Type: Photorealistic? Stylized? Match it to your brand tone, and budget.
• Pilot & Iterate: Test with internal audiences or small customer segments. Always a good idea.
• Integrate with your digital landscape: Link your avatar to CRM, product catalogs, and social media marketing tools and apps for example.
• Measure ROI: Use metrics like time on page, engagement, conversions, and support resolution time.

Ethical and Brand Risks

AI avatars must be handled with care:

• Transparency: Always disclose that it’s an AI, not a human.
• Bias & Accuracy: Easily done by mistake and there are high profile cases where bias crept in. Test thoroughly to avoid misinformation.
• Brand Fit: Always a concern with the brand team, bless them.
• Security: Ensure avatars can’t be hijacked or spoofed. Check all set ups with your IT experts.

AI Avatars key takeaway: Your Move, Marketers!

Luo Yonghao’s avatar livestream is a glimpse into the near-future of marketing. While full real-time, unscripted avatars may still be a few years away. This post has highlighted use cases for piloting AI Avatars beyond the use of AI for content generation. The opportunity is there for mid to large organizations looking to scale and grow communications. Key point is experimentation with engaging with audiences in novel new ways.

Start small. Experiment (MVPs) and remember—in the age of AI avatars, the face of your brand could soon be virtual, but the impact will be very real. That is the aim!

How do you feel about adopting AI Avatars?