How We Scan 15 Years Of Creator Content In 48 Hours: Building Secure On Google Gemini

About two years ago, we had a problem.

Secure, our award-winning brand safety and creator vetting platform, wasn't scaling the way we needed it to. We're talking about brands that work with hundreds of thousands of creators, and they need to know that every single one of those creators is safe for their brand. Not in a few weeks. Not tomorrow. Right now.

But it was bigger than just speed. We could see where influencer marketing was heading toward AI that could actually understand content at scale, not just count likes and followers. Secure needed to go beyond basic text scanning. It needed to analyze video frame by frame, understand context across languages, detect subtle brand mentions, and predict creator fit. We had ideas for what we wanted to build, but our infrastructure couldn't support it.

Our old system couldn't handle it. Analyzing historical content took forever. Our engineering team was stretched thin just keeping things running. And the cloud bills kept getting bigger while service quality stayed the same or got worse.

We knew that if we wanted to make Secure what we envisioned, a definitive brand safety and creator intelligence platform, we needed to make a change.

So we started looking at alternatives. After testing everything on the market, we moved to Google Cloud earlier this year. It's probably the best infrastructure decision we've made as a company.

What Actually Changed

The big headline number is that our infrastructure costs dropped 92%. That's not a typo. We're also running with 70% fewer engineering resources dedicated to infrastructure, which means our team can focus on actually building new features instead of keeping the lights on.

But the cost savings aren't even the most important part. What matters is what Secure can now do for our clients.

Secure monitors 86,000 creator channels every day.

It looks at everything: text, video, images, and audio across all the major social platforms. When we switched to Google Cloud and started using Gemini for the AI analysis, we didn't just get faster performance. We have capabilities that didn't exist before.

Here's a real example: one of our clients used over 100,000 creators in a single year. Secure analyzed all their historical video content with Gemini.

If someone had to do that manually, it would take 36 years. Gemini did it in 48 hours.

But it's not just speed. Secure can now catch things that human reviewers would miss.

Subtle brand mentions buried in hours of content, tonal shifts that indicate a creator's values might not align with a brand, and patterns across thousands of videos that predict future performance. The AI automatically detects unpaid brand mentions and logos, identifies past sponsorships, and classifies content by topic, style, and age rating.

For brand safety, Secure scans for over 10 categories of concerns: hate speech, drug use, and other sensitive content, with customizable risk tolerance levels. It works across multiple languages and can even perform OCR on text in videos. Brands can set their own parameters based on their specific values and risk tolerance.

"We work with some of the world's largest brands, and they want a deep background of their creators' social activity," says Nick Spiro, our Chief Commercial Officer. "But when a marketer wants to run a campaign, they don't always have weeks to look at brand safety. Our old technology couldn't handle that speed or volume."

Why Gemini Specifically

When we were testing AI models for Secure, Gemini was just better than the competition. Other models had higher false-positive and false-negative rates. They'd flag things that were fine or miss things that weren't. When you're protecting major brands, you can't mess around with that kind of accuracy problem.

But here's what really matters for Secure: Gemini can actually understand video content at scale, unlike anything else. We're not just scanning text or doing basic image recognition. We're analyzing every frame of video, understanding context, detecting paid partnerships, and determining whether a creator's content style actually matches what a brand is going for.

"The key for us is being able to combine this social intelligence insight with scanning every frame of video with Gemini," Nick says. "We don't think anyone else can really do that right now."

This capability is what makes Secure different from any other creator vetting tool. We can now detect subtle brand mentions that a human might miss while watching at regular speed. We can identify content themes and patterns across years of creator output. We can predict which creators will resonate with specific audiences based on actual content analysis, not just surface-level metrics like follower count or engagement rate.

The multimodal analysis is huge. Secure doesn't just read captions, it watches videos, listens to audio, reads text on screen, and understands how all of that fits together. That's the kind of AI work that changes what's possible in creator vetting and brand safety.

The Rest of the Stack

We use a bunch of Google Cloud tools beyond just Gemini:

BigQuery is the backbone of our data infrastructure. When advertisers want to see how their campaigns are performing, we need to pull and analyze data fast. BigQuery makes that easy.

.          

We use Vertex AI for our machine learning workflows. It gives us the flexibility to build custom solutions for specific client needs.

And honestly, Google Workspace has been great for internal collaboration. We've now got Gemini integrated into Sheets, Meet, Docs, and Gmail. Nick put it well: "I'd have happily used Excel my whole career, but the collaboration just wasn't there - there were too many user barriers and pain points."

What We're Building Next for Secure

Here's where it gets interesting. The Google Cloud infrastructure isn't just making Secure faster and cheaper; it's enabling us to build capabilities into Secure that weren't possible before.

We're adding AI that can auto-generate campaign briefs by analyzing a brand's goals, past successful campaigns, and matching them to creator content at scale. Instead of spending hours briefing creators, Secure will generate briefs that already understand the creator's style and the brand's needs.

We're building smarter creator discovery features that go way beyond follower counts and engagement rates. Secure will soon identify creators poised to trend in specific niches, detect authentic brand affinity before a creator has ever worked with a brand, and predict campaign performance based on historical content analysis.

The discoverability piece is huge. Right now, if a brand wants to find creators who naturally talk about sustainability in an authentic way, they're usually relying on keyword searches and manual vetting. Secure will be able to watch thousands of hours of content and identify creators who genuinely care about these topics, even if they never use the obvious hashtags.

We're also training models within Secure to understand cultural context and regional nuances at scale. A creator who's massive in Brazil might have completely different content patterns than someone big in Japan. Secure is learning to understand those differences and help brands navigate them.

None of this would be possible on our old infrastructure. Google Cloud provides the compute power to experiment with new models for Secure, the data infrastructure to train them effectively, and the scalability to deploy them to our entire client base.

"Being able to scale with a reliable partner like Google Cloud is incredibly valuable for us because we can keep exceeding customer expectations while continuing to grow," Nick says. "It really has become a beautiful story between our two companies."

The influencer marketing industry is moving toward AI-powered decision-making. Secure is about making sure we're not just keeping up with that shift. We're leading it.

If You're Thinking About Migrating

Look, every company's situation is different. But if you're in a similar spot to where we were, growing fast, cloud costs getting out of hand, tech not keeping up with your AI ambitions, it's worth looking at your options.

For us, Google Cloud wasn't just cheaper. It was faster, more accurate, and it gave us the foundation to make Secure what we knew it could be. In influencer marketing, where everything moves at internet speed and AI is reshaping how campaigns work, having the right infrastructure isn't optional.

The migration itself took some work, obviously. But we're seeing the benefits every single day in what Secure can deliver to clients and how efficiently we can operate. More importantly, we're now able to move as fast as our product ideas, not as fast as our infrastructure allows.

If you want to see what Secure can do for your brand safety and creator vetting, check it out at viralnation.com/solution/secure or reach out to learn how we're using AI to transform influencer marketing.