How We Used AI to Recover 7,000 Missing Meta Descriptions in 48 Hours

From SEO Setback to 7,000 Optimized Pages in 48 Hours: A Real-World AI Case Study

What happens when a large e-commerce brand relaunches 7,000 category pages… and forgets to include meta descriptions?

That’s the situation we found ourselves in recently. One taxonomy update. Thousands of critical SEO entry points suddenly blank. And no one had looped in the SEO team.

But instead of panic, we leaned into process—and AI.

The Problem: A Taxonomy Relaunch Without SEO Input

We work with a large e-commerce brand that manages:

1. Hundreds of thousands of SKUs
2. Over 7,000 dynamic category pages
3. A complex PIM (Product Information Management) system powering site content

When they rolled out a massive taxonomy restructure and relaunched all category pages, they missed a critical step:

Meta descriptions were left empty across every single one of them.

Historically, fixing this would take:

1. Weeks of copywriting labor
2. Generic formulas that underperform in SERPs
3. Or… nothing—letting Google auto-generate snippets (and SEO suffer)

Not anymore.

Our Solution: AI That Thinks in Workflows, Not Just Prompts

Here’s how we solved it in under 48 hours using Gryffin:

1. Exported all 7,000 category pages
2. Pulled live data including category names, sample products, and tags.
3. Created a smart AI template
4. Used Gryffin to combine:
5. Category context (e.g. “Men’s Trail Running Shoes”)
6. Product examples
7. Search-intent-informed language
8. …into custom meta descriptions that were informative, on-brand, and SEO-ready.
9. Ran the data through Gryffin’s bulk prompt engine
10. Unlike basic AI tools, Gryffin let us:
11. Import CSV data
12. Run dynamic prompts across rows
13. Format output for PIM-ready integration
14. Cleaned, formatted, and delivered
15. In two days, we handed the dev team a plug-and-play CSV—fully populated with optimized metadata.

Why This Would’ve Been Impossible Without AI (and Gryffin)

Could ChatGPT help here? To a degree—but not at scale.You’d still be stuck copying, pasting, and formatting thousands of prompts.

Gryffin made it repeatable, customizable, and scalable—with workflows built specifically for SEO and e-commerce use cases.

This is exactly the kind of AI-powered transformation teams need right now—not theoretical. Not future-state. It’s real, repeatable, and revenue-impacting.

The Bigger Picture: This Is the New SEO Playbook

If this had happened five years ago, the only real options would have been:

1. Let Google guess
2. Use a one-size-fits-all formula
3. Spend weeks of time (and thousands in cost) rewriting manually

Today?

You can turn a multi-thousand-page SEO emergency into a data-driven, brand-consistent recovery plan—powered by AI.

Takeaway for SEO Teams

AI isn’t here to replace strategic thinking. It’s here to make your playbook scalable.

1. Don’t just generate content—engineer structured, brand-aligned SEO assets.
2. Don’t just automate output—automate quality at scale.
3. And don’t wait for disaster to test your AI workflows. Build them now.

Want to see how we structured the prompts, workflows, and outputs inside Gryffin?

Happy to share how we’re using AI to solve real SEO problems every day.