Gemini 3.8 Live Avatar: Is It Available in Spain?

Google has just put a face on its enterprise agents. Gemini 3.8 Live with Live Avatar combines low-latency conversation, voice and video to create avatars that can speak and react live. It sounds like a digital shop assistant from the future. The problem is that some very down-to-earth questions still need answers: what it costs, how it performs and who can sign up for it from Spain.

TL;DR: The no-nonsense summary

• What it is: an enterprise model that generates voice, text and a video avatar during a live conversation.
• Real availability: Google confirms general availability in Gemini Enterprise and connection points in the US and EU.
• Spain remains unclear: European infrastructure exists, but there is no specific confirmation that Spanish organisations can buy or access it.
• The fine print: continuous sessions last only a few minutes, while custom avatars require access to an approved list.

Verdict: there is serious technology behind the avatar, but a flashy demo does not yet make it a profitable commercial channel.

What is Gemini 3.8 Live with Live Avatar?

Gemini 3.8 Live with Live Avatar is a Gemini Enterprise capability for building audiovisual agents that converse through a synchronised avatar. According to the announcement, it is far removed from the usual filter that lets any Gemini app user build a virtual presenter in five minutes.

Google officially unveiled it on 24 September 2026. The system continuously receives audio, images, video and text, and can return text, voice and video in near real time.

The avatar adjusts lip movements, expressions and speaking turns as it chats. The agent can also query data and run tools in the background without interrupting the interaction. That last part matters more than the pretty face: it lets the conversation get things done, rather than simply recite a script with a catalogue smile.

There are also pre-built avatars. Creating a custom one from reference images requires explicit inclusion on an enterprise-approved list. Google also says that generated audio and video include SynthID to make them easier to identify as synthetic content.

Is Gemini 3.8 Live available in Spain and the EU?

The European Union has confirmed technical availability. Spain is another matter: there is no specific commercial confirmation. Google Cloud classifies Gemini 3.8 Live with Live Avatar as generally available in Gemini Enterprise and documents multi-regional connection points in both the United States and the EU.

This is not just semantics. Having a European connection point shows the region is covered, but it does not clarify plan, contract, account or billing-country requirements. The Google Cloud post confirms general availability, but it does not allow us to claim that every Spanish company can enable it today.

Two exaggerations need stopping immediately. The stage is clear: Google says it is ready for enterprise production. What we cannot do is slap a “available in Spain” label on it. EU confirmed. Spain awaits the contractual small print.

There is more fog. The sources provide no specific pricing for video or provisioned capacity. Nor do they offer independent measurements of latency or quality. There is also a documentation discrepancy: announcements mention 97 languages, while the general Gemini Live API documentation lists 24. Selling language coverage before resolving that would be getting ahead of ourselves.

What can Live Avatar do for marketing and commerce?

Google suggests shopping assistants, ecommerce support, brand representatives, customer service, interactive tours and agents for tasks such as hotel check-in. These are official use cases, not proof that they generate more sales, cut costs or improve satisfaction.

The most obvious example is an online store where a user shows a product on camera, asks about an alternative and receives a recommendation explained by an avatar. The agent could check inventory or product information while keeping the conversation going. On paper, it makes sense.

It also fits with Google’s agentic commerce, where AI stops merely answering questions and participates in tasks within the buying process. The avatar adds visual presence. It does not automatically add judgment, trust or a compelling commercial proposition. Don’t confuse packaging with strategy.

The trap is measuring the project by the number of conversations started. An avatar may spark curiosity while harming the business if it gives poor recommendations, drags out a simple request or puts customers in a kind of video call they never asked for.

A useful test compares query resolution, abandonment, assisted conversion and cost per conversation against traditional chat. If nobody defines that comparison before launching the talking doll, we will get PowerPoint innovation. Very futuristic. Very little business value.

The limits the demo does not show

The official Gemini 3.8 Audio model card limits continuous Live Avatar interactions to a few minutes. That works for a brief recommendation or a specific query, but raises doubts over longer journeys, complex sales and multi-step support sessions.

That duration changes the entire design. Teams will need to decide which task the avatar resolves before the session runs out, what data it needs and how it hands the conversation to a person when things go sideways. Because they will. Customers have that irritating habit of not following the prepared demo.

Then there is the brand question. A virtual representative can repeat an approved tone and be available at scale, but it does not turn a bland company into a brand with personality. If product knowledge is mediocre, the avatar simply puts synchronised lips on a mediocre answer.

Gemini 3.8 Live with Live Avatar deserves a controlled trial in commerce or customer service, especially where a visual query adds genuine value. Until we know pricing, Spanish access and independent performance, treating it as a mass replacement for salespeople or human agents is selling smoke with excellent lip sync.