GEO Strategy

Gemini Embedding 2 - What multimodal AI means for your employer brand

By 25 March 2026No Comments

The day AI learned to watch

Imagine this: a software engineer asks Gemini: “Which tech company in Amsterdam suits me if I value autonomy and an informal culture?”

Until last week, AI based that answer solely on text. Your careers page, your Glassdoor reviews, your LinkedIn posts. That employer branding video you spent €15,000 on? Invisible. The podcast in which your CEO talks about your culture? Not considered. The photo shoot of your workplace? Irrelevant.

That's over.

On March 10, 2026, the article Google Gemini Embedding 2 launched. It explains an AI model that for the first time understands text, video, images and audio simultaneously. Not as separate files, but as one coherent message.

The question is: what does AI see when it looks at your employer brand looks?

The gap between what candidates saw and what AI understood

Until March 2026, AI systems only analysed textual content from employers. Employer branding videos, podcasts and photo reports were invisible to ChatGPT and Gemini.

A candidate orienting on a new employer asks ChatGPT or Gemini, “Is company X a good employer for someone with my profile?” The answer given by AI increasingly determines the initial view. Not your website. Not your job posting. The AI's answer.

Until now, that response built exclusively on text. Your employer video on YouTube, your photo report of the workplace, the podcast in which your CEO talks about your culture: to AI, that was invisible unless there was an extensive transcript or caption. This is exactly why Generative Engine Optimization has so far focused on clear texts, Q&A structure and factuality.

What specifically will change with Gemini Embedding 2?

Gemini now understands video, audio and images without the need for text. From now on, a video in which employees talk about career opportunities has direct semantic value. A podcast about your culture counts. Footage from the workplace is read as part of the employer story.

Employers who invested in employer branding videos, photography and podcasts for years, but never set that content up for AI readability, are now at real risk. AI has long ignored that content. Soon it will actively interpret that content, right or wrong.

Why this is urgent now

The numbers are clear:

  • 38% of job seekers under 30 use AI for employer research (LinkedIn Global Talent Trends, 2025)
  • GEO-optimised content is 30-115% cited more often in AI responses than non-optimised content (Princeton University & Georgia Tech, 2024)
  • 72% of candidates have positive experiences with AI-driven orientation (Tidio, 2025)

Increasingly, the answer given by AI determines the first view. Not your website. Not your job posting. The AI's answer.

And that answer just got a lot richer.

What are the three immediate risks to your employer brand?

  1. Inconsistency between image and text becomes apparent. Does your website say “informal culture with short lines” but does the imagery show formal meeting rooms and suits? AI recognises this contradiction and weighs it up in the image it paints to candidates.
  2. Untitled video does not count or counts incorrectly. An employer branding video without clear description, chapter titles or spoken key messages has no semantic weight. AI will not extract useful information from it.
  3. Competitors who do optimise win the equation. Candidates who ask AI to compare employers will get an answer based on what AI understands. Those who are more findable and understandable to AI win that comparison. Every time.

How do you recognise whether your employer brand is ready for multimodal AI?

Test whether your employer brand is ready for multimodal AI with 3 questions.

Question What you check
Are the titles and descriptions of your videos written in such a way that someone reading them would without looking understands who you are? Textual findability
Does your video/podcast contain explicit, spoken statements about culture, working conditions and growth path? Audio extraction by AI
Does the visual material on each channel match the text you use? Cross-modal consistency

Three times “no”? That's a serious signal.

What is the difference between SEO and GEO for employer branding?

Search Engine Optimisation (SEO) ensures that your pages rank high in Google. Generative Engine Optimisation (GEO) ensures that AI systems such as ChatGPT, Perplexity and Gemini represent your organisation correctly, completely and positively in their responses. GEO does not drive clicks. GEO drives on the image AI forms before a candidate clicks on a link at all. With the introduction of Gemini Embedding 2, that playing field expands from text to all the content an employer publishes.

What about GDPR for European employers?

For publicly accessible content, careers site, YouTube channel and podcast, Gemini does not process personal data of the organisation and is outside the GDPR scope. Organisations that want to use the Gemini Embedding 2 API themselves to process candidate data can do so GDPR-compliant via Google Cloud with EU data location in region europe-west4 (Netherlands), but that requires processor agreements and a deliberate technical setup.

What does AI Rebels do for employers who want to be at the forefront?

AI Rebels actively monitors how AI systems position employer brands and optimises employers' entire digital presence for AI visibility. Nigaio Wijnen and Bryan Peereboom guide organisations in setting up their employer brand for the new reality of multimodal AI, from technical audit to ongoing reputation management. Watch the packages and approach or read more at the knowledge page. We will soon be hosting a webinar on what multimodal AI means for your employer brand and what you can do today. Keep an eye on our knowledge page for the announcement.

The bottomline

Gemini Embedding 2 changes the rules of the game. Your employer brand is no longer judged by what you writes, but on what you shows And whether the two match up.

And this is just the beginning. If Gemini understands video and audio today, tomorrow it will understand subtleties: body language in interviews, energy in the workplace, the tone in which your CEO speaks.

Want to know how Gemini now positions your organisation against competing employers? Schedule a demo.

Written by Nigaio Wines and Bryan Peereboom - AI Rebels, 25 March 2026

Sources: