Pandora is resolving 60% of customer service cases with AI. PacSun is selling products directly through ChatGPT. Different use cases, same outcome: both brands are proving that AI commerce is already here. What they're doing is visible. What's making it possible is where the story gets interesting, and that's what this blog breaks down.
Pandora Proves the Model: Service and Sales on One Platform
Pandora runs two Salesforce Agentforce agents. Clara, the service agent, automates routine requests like order tracking and FAQs, and now resolves 60% of cases without a human, a shift that also lifted Net Promoter Score by 10%. Gemma, the personal shopper agent, helps customers choose jewelry by asking about occasion, recipient, and budget, the way an in-store associate would.
Neither agent started from zero. Pandora already runs 22% of its global sales through a single, unified Salesforce Commerce Cloud instance. The agent layer sits on top of that existing foundation.
Source: Salesforce Customer Stories - Pandora
Implement AI-ready commerce with Tru's Salesforce Services and build the foundation AI depends on.
PacSun Proves the Channel: Selling Inside the Conversation
PacSun took the opposite angle: instead of upgrading the in-store-like experience, it's meeting Gen Z and Gen Alpha where they already spend time. Its product catalog now syncs directly into ChatGPT through Salesforce's Agentforce Commerce and OpenAI integration, no middleware, no custom build, reaching general availability in July 2026. The value here is discovery moving upstream of the storefront entirely.
For that sync to work, PacSun's catalog had to be clean, structured, and pulled from one system. That's the quiet prerequisite behind every flashy AI commerce headline, and it's exactly the gap Salesforce points to: AI-referred traffic already converts at 8x the rate of social, and retailers running their own shopper agents grew sales 59% faster than those still on the sidelines during last year's holiday season.
Source: Salesforce Newsroom – Agentforce Commerce Release
The Blueprint Behind Successful AI Commerce
Strip away the brand names, and three patterns hold across every serious deployment right now:
One platform, not five. Every agent sits on top of unified commerce infrastructure, not a patchwork of legacy systems duct-taped together.
Clean, connectable data. Catalogs, pricing, and inventory have to live in one place before any AI channel can reliably surface them.
AI as a layer, not a rebuild. The winners aren't ripping out their stack to "do AI." They're adding an intelligent layer on top of infrastructure that already works.
According to Salesforce Newsroom, AI influenced 20% of global online sales last holiday season, worth $262 billion in commerce, a scale no brand can afford to treat as a side experiment anymore.
Tru Builds the Foundation AI Commerce Runs On
This is where Tru fits, not the AI models, but everything underneath them:
Commerce platform implementation. Salesforce Commerce Cloud, Shopify, and headless builds that an agent layer actually connects to.
Catalog architecture and data governance. The structured, connectable product data that makes a brand AI-channel-ready in the first place.
AI readiness assessments. Digital maturity workshops that identify exactly where a brand's infrastructure has gaps, before they spend a dollar on an agent with nothing solid to sit on.
Pandora and PacSun didn't win because they picked the right AI vendor. They won because the groundwork was already there when the AI arrived.
Ready to launch AI-powered commerce? Tru's Agentic Commerce Solutions help bring AI shopping experiences to life.
The Foundation Comes First
Build the foundation before you build the agent. Most brands aren't missing an AI strategy; they're missing the foundation an AI strategy needs to run on. The right time to fix that is before the gap between "AI-ready" brands and everyone else gets any wider.
