Uploaded October 2025 | Updated September 2026, 8 hours ago
I visited Pandora's booth at Dreamforce 2025 and learned something fascinating about using AI agents to replicate in-store magic online.
Pandora is the world's largest jewelry brand with thousands of product options. Their challenge? In physical stores, sales associates ask questions, understand emotions, and help find the perfect gift. Online? That personal touch has been nearly impossible to replicate. Until now.#Salesforcepartnership #Sponsored
🎁 MEET GEMMA - PANDORA'S AI GIFT ADVISOR
Gemma is Pandora's new Agentforce agent, currently live in Australia, helping customers find the perfect gift online.
The challenge Luciano Rodembusch, General Manager for North America at Pandora, described is unique:
"Pandora has more than thousands of options and many of the charms and jewelry don't have literal meaning. They have an emotional meaning. To really understand that part, you need the agent to go through thousands and thousands of interactions and a lot of learning to narrow down those options."
This isn't about product specs. It's about understanding EMOTIONS at scale.
💡 HOW IT WORKS
When you visit Pandora online, Gemma asks questions like a sales associate would:
Who is the gift for?
What's the occasion?
What feeling do you want to convey?
Then uses AI to navigate thousands of options with emotional meanings to find the right match.
🎯 PANDORA'S TWO-AGENT STRATEGY
1️⃣ Clara - Customer service agent (already live)
2️⃣ Gemma - Gift advisor agent (live in Australia, expanding)
🤔 THE TRUST QUESTION
I asked: "Do you trust those agents to deliver a similar service to your in-store experience?"
Luciano's honest answer: "It's a beginning of the journey. We're just testing one market. So far, we have very good results and very good learning."
That's the right approach: Start in one market. Learn. Iterate. Scale.
⚡ WHY THIS MATTERS
Most e-commerce agents handle transactions or answer FAQs.
Gemma is doing something harder: understanding emotional context to recommend products with emotional meaning.
That's a different level of complexity. You can't just match keywords to a catalog. You need to understand:
Relationships
Occasions
Sentiments
Personal preferences
Cultural context
And do it through natural conversation, learning from thousands of interactions.
💬 MY TAKEAWAY
This is what retail's digital labor economy looks like:
Not replacing the human touch in stores. Scaling it to millions of online customers who couldn't get that personalized service before.
Pandora isn't trying to eliminate their sales associates. They're giving online customers the same thoughtful, emotional, personalized guidance that in-store customers get.
That's the promise of the Agentic Enterprise in retail: human-level service at digital scale.
I visited Pandora's booth at Dreamforce 2025 and learned something fascinating about using AI agents to replicate in-store magic online.
Pandora is the world's largest jewelry brand with thousands of product options. Their challenge? In physical stores, sales associates ask questions, understand emotions, and help find the perfect gift. Online? That personal touch has been nearly impossible to replicate. Until now.#Salesforcepartnership #Sponsored
🎁 MEET GEMMA - PANDORA'S AI GIFT ADVISOR
Gemma is Pandora's new Agentforce agent, currently live in Australia, helping customers find the perfect gift online.
The challenge Luciano Rodembusch, General Manager for North America at Pandora, described is unique:
"Pandora has more than thousands of options and many of the charms and jewelry don't have literal meaning. They have an emotional meaning. To really understand that part, you need the agent to go through thousands and thousands of interactions and a lot of learning to narrow down those options."
This isn't about product specs. It's about understanding EMOTIONS at scale.
💡 HOW IT WORKS
When you visit Pandora online, Gemma asks questions like a sales associate would:
Who is the gift for?
What's the occasion?
What feeling do you want to convey?
Then uses AI to navigate thousands of options with emotional meanings to find the right match.
🎯 PANDORA'S TWO-AGENT STRATEGY
1️⃣ Clara - Customer service agent (already live)
2️⃣ Gemma - Gift advisor agent (live in Australia, expanding)
🤔 THE TRUST QUESTION
I asked: "Do you trust those agents to deliver a similar service to your in-store experience?"
Luciano's honest answer: "It's a beginning of the journey. We're just testing one market. So far, we have very good results and very good learning."
That's the right approach: Start in one market. Learn. Iterate. Scale.
⚡ WHY THIS MATTERS
Most e-commerce agents handle transactions or answer FAQs.
Gemma is doing something harder: understanding emotional context to recommend products with emotional meaning.
That's a different level of complexity. You can't just match keywords to a catalog. You need to understand:
Relationships
Occasions
Sentiments
Personal preferences
Cultural context
And do it through natural conversation, learning from thousands of interactions.
💬 MY TAKEAWAY
This is what retail's digital labor economy looks like:
Not replacing the human touch in stores. Scaling it to millions of online customers who couldn't get that personalized service before.
Pandora isn't trying to eliminate their sales associates. They're giving online customers the same thoughtful, emotional, personalized guidance that in-store customers get.
That's the promise of the Agentic Enterprise in retail: human-level service at digital scale.
