Gemini Omni: Realistic AI videos and implications for SMEs

What has changed with Gemini Omni

On May 23, 2026, Google made Gemini Omni available, its AI model defined as anything-to-anything . In practice, the system accepts any combination of inputs — text, image, audio, video — and returns outputs in any format. Thus, the barrier between different content modalities has effectively disappeared.

According to hands-on published by The Verge , the model is able to generate realistic videos starting from static images with a surprisingly reduced technical effort. In fact, tests conducted by the editorial team show convincing results even on complex subjects, without requiring advanced post-production skills.

This marks a major shift from earlier models. Plus, the quality of the generated video blows past what we saw just months ago with similar tools. As a result, the line between real and synthetic content is getting tougher and tougher to spot with the naked eye.

Multimodal architecture: how it really works

Gemini Omni is not just a small upgrade. Instead, it is a deep revamp of the model's architecture. Google built a native multimodal system, where different input and output modes are baked right into the training stage.

In short, the model doesn't convert text into images and then images into video through separate pipelines. Because of this, the process is smoother and gives you outputs with fewer visual glitches. This setup also cuts down the overall wait time when creating stuff.

To delve deeper into the technical evolution of multimodal models, it is useful to consult the analyses of MIT Technology Review , which has documented the trajectory of these systems in recent years. Similarly, reports from Gartner on generative AI provide a framework for enterprise adoption expectations.

Immediate impact on digital content production

For Italian SMEs, the first practical implication concerns content production. Today, generating a credible promotional video requires fewer resources than it did last year. Therefore, production costs for visual campaigns are potentially reduced in a significant way.

However, this democratization brings a specific risk. In fact, if the access threshold is lowered for everyone, competitors - and malicious actors - can also produce high-quality content with ease. Consequently, brand differentiation can no longer be based solely on the technical quality of the video.

We at SHM Studio we notice that the most forward-thinking companies are already shifting their focus to narrative consistency and perceived authenticity. In other words, value is moving from production to strategy. For this reason, a digital marketing strategy well-structured becomes even more relevant in this context.

The deepfake node: a concrete reputational risk

Gemini Omni amplifies an issue already present in the digital landscape: corporate deepfakes. Therefore, this topic must be tackled clearly, without alarmism but also without underestimating it.

A synthetic video showing a company executive making statements they never made can now be produced quickly. Furthermore, the quality achieved by Gemini Omni makes it harder to recognize at first glance. Therefore, SMEs must consider this scenario in their reputational risk management.

Mitigation measures include active monitoring of brand mentions, definition of internal verification protocols, and transparent communication with stakeholders and customers. In particular, those operating in regulated sectors - finance, health, legal - must pay attention to regulatory implications. The organic visibility of the brand on Google can also act as a reputation shield, since a solid positioning makes the viral spread of fake content much harder.

What nobody tells you: the problem of scalable trust

There's one thing people rarely talk about when discussing these tools. The real issue isn't just one fake piece of content. Instead, it's how trust in pretty much anything you see keeps slowly chipping away.

When the public can no longer distinguish the real from the synthetic, the adaptive response is generalized distrust. Therefore, even authentic content is viewed with suspicion. This mechanism disproportionately affects smaller brands, which lack the necessary notoriety to be considered credible by default.

For this reason, investing in authoritative copywriting and in verifiable content becomes a strategic lever. Similarly, presence on channels like Linkedin — where the professional context reduces the virality of manipulated content — offers an extra reputation safeguard.

What to do now: operational guidance for SMEs

First off, it helps to map out your most vulnerable digital touchpoints. Video channels in particular—YouTube, Instagram Reels, TikTok—are where impersonation risks run highest. Because of this, you should double-check that you have verified control over your official accounts.

Subsequently, it is advisable to define an internal policy on the use of generative AI. This policy should establish which tools are approved, in which contexts, and with which review processes. Furthermore, it should include guidelines on disclosure, i.e., when and how to communicate that content has been produced with AI support.

On the opportunities front, SMEs can leverage Gemini Omni and similar tools to speed up low-risk content production: explainer videos, product tutorials, animated presentations. However, it is advisable to maintain human supervision over the creative process. I AI services that we offer integrate this supervision in a structured way.

Finally, those managing ad campaigns should assess the impact of these tools on how ads are perceived. google ads campaigns that use generative video assets might benefit from lower production costs, but they require careful verification of brand consistency.

Prospects: where we're going in the 2027-2028 biennium

Projections for the next two years indicate a further acceleration. According to McKinsey Global Institute , the adoption of generative AI tools in marketing and communications functions will grow steadily through 2028. Therefore, those who do not develop internal skills today risk falling structurally behind.

On the other hand, companies that integrate these tools with clear governance will enjoy real competitive perks: faster production, tailored content, and lower operating costs. So, the question isn't whether to use generative AI, but how to do it responsibly.

For Italian SMEs, the most effective path involves specialized consulting that evaluates the company's specific context. We at SHM Studio we support companies in this process, from strategy definition to tool selection, up to internal team training. Those who want to learn more can visit the section web services , explore our analysis on the blog or contact us directly from the contact page .