Bringing In-Store Warmth Into Digital Channels With AI Agents

For retailers, the challenge is no longer simply being present across channels. It is creating digital experiences that feel as helpful, personal and brand-right as the best moments in store. Customers do not separate service from shopping, or commerce from care. They expect fast answers, relevant guidance and continuity across the entire journey, whether they are asking where an order is or looking for the right gift.

Pandora’s work with Publicis Sapient and Salesforce shows what that can look like in practice. Together, the teams used Agentforce to design AI-powered service and shopping experiences that extend the warmth and expertise of Pandora’s in-store interactions into digital channels. The result was not AI deployed as a cold automation layer, but AI shaped around brand voice, customer context and measurable business value.

A More Human Model for Digital Service

Pandora had a clear customer experience challenge. In store, shoppers benefit from rich, story-driven conversations with associates who help them find meaningful pieces. Online, the experience was under pressure from high volumes of customer inquiries across email, web and chat. Routine questions created bottlenecks, especially during peak periods, and slowed response times for customers who needed support.

The opportunity was bigger than deflecting contact. Pandora wanted to bring more of its in-store warmth and personalization into digital channels while making service faster and more scalable. That meant designing interactions that felt conversational and context-aware, not scripted and mechanical.

Publicis Sapient helped shape that strategy and bring it to life. Working with Salesforce, the team developed a roadmap to scale personalized service globally through intelligent agents that could support both service and shopping journeys.

Two Agents, Two High-Value Moments

Clara was developed as a customer service agent focused on routine questions that often create friction in digital support. She handles inquiries such as order status and jewelry care FAQs in conversational language. Just as importantly, Clara is connected to the systems needed to make those conversations useful. By integrating with Service Cloud and Commerce Cloud, and drawing on Pandora’s IBM Sterling order management environment, Clara can provide real-time order updates rather than generic responses. That gives customers fast answers to one of the most common post-purchase questions: where is my order?

Gemma was designed for a different moment in the journey: inspiration and decision-making. Acting as a personal shopper agent, Gemma helps customers choose jewelry based on the occasion, the recipient and the shopper’s budget. To do that well, the experience needed more than a product catalog. Gemma draws on data from Commerce Cloud, Bloomreach and Pandora’s UX research through Data Cloud to deliver recommendations that feel more tailored and relevant.

Together, Clara and Gemma show an important shift in how retailers can think about AI agents. One supports service efficiency and post-purchase confidence. The other supports product discovery and conversion. Both are grounded in the same principle: use AI to extend brand expertise into digital moments where customers want help, not just speed.

Designing Beyond Automation

What makes this work notable is not only the technology stack, but the experience design behind it. Publicis Sapient helped Pandora move away from rigid scripts toward more dynamic, agentic interactions. That distinction matters.

Many digital service experiences still feel transactional and brittle. They can answer a narrow question, but they do not adapt well to intent, context or emotion. In categories like jewelry, where purchases are often tied to milestones, gifting and self-expression, that limitation becomes even more visible. Customers are not only completing tasks. They are choosing something meaningful.

A more effective AI experience therefore has to reflect the brand’s personality and the customer’s situation. For Pandora, that meant designing agents that could respond in a more human way, using conversational language and context to make interactions feel supportive rather than procedural. It also meant recognizing that AI should strengthen the role of human specialists, not erase it. When routine questions are handled autonomously, service teams can focus on the higher-value interactions where empathy, judgment and deeper brand knowledge matter most.

Built on the Right Retail Foundation

Agent experiences like Clara and Gemma do not succeed in isolation. They depend on connected systems, accessible data and operational foundations that support accurate, useful responses.

That is where Pandora’s broader transformation becomes relevant. Publicis Sapient’s work with the brand extended beyond AI into commerce modernization and omnichannel operations. The teams helped modernize Pandora’s Salesforce Commerce environment with a common codebase, shared functionality and common storefront implementation across e-commerce platforms. In APAC, unique custom code was decoupled and rebuilt to reduce redundancy and improve maintainability. That work accelerated release velocity from several months to a matter of weeks, cut migration time in half and required no additional downtime after the first successful migration.

Publicis Sapient also helped strengthen Pandora’s omnichannel fulfillment capabilities through IBM Sterling Order Management on Cloud. That work improved real-time inventory visibility, supported ship-to-home, optimized returns and refunds and enabled services such as Click-and-Collect and Store Fulfillment. It also created better tools for handling post-purchase inquiries.

This matters because trusted AI depends on trusted operations. An agent can only deliver a strong answer about an order if order data is visible in real time. A personal shopping agent can only feel relevant if it has access to the right commerce and customer insight. In retail, experience design and operational design are inseparable.

Measurable Impact Across Service and Commerce

The business reported 60% autonomous case deflection, helping reduce pressure on service teams and allowing specialists to focus on more valuable interactions. It also reported a 10% uplift in Net Promoter Score driven by agent-first service, pointing to a customer experience gain rather than a simple efficiency play. And with 22% of total sales handled online through Commerce Cloud, the digital channel is clearly a meaningful part of the broader retail business.

These results reinforce an important lesson: AI agents can create value across both cost-to-serve and revenue-supporting journeys. They can reduce friction in service, improve confidence after purchase, guide product selection and strengthen digital commerce performance. But they do that best when they are designed around customer needs and brand experience, not just containment targets.

What Retail Leaders Should Take Away

Pandora’s Agentforce journey offers a practical blueprint for retailers that want digital interactions to feel more personal without sacrificing scale.

First, start with the moments that matter most. Routine service questions and high-intent shopping decisions are both strong candidates for AI support because they sit at the intersection of customer need and operational pressure.

Second, design for warmth and context, not only automation. Customers can tell when an interaction has been reduced to a script. The goal should be to create experiences that feel natural, informed and aligned to the brand.

Third, connect AI to the systems that make it useful. Service Cloud, Commerce Cloud, order management and customer data all play a role in turning an AI interaction into a meaningful answer.

Finally, treat AI agents as part of a broader transformation. Strong digital experiences are built on agile commerce foundations, connected fulfillment operations and cross-functional ways of working. When those elements come together, AI can become more than a support feature. It can become a new expression of the brand.

For Pandora, that meant bringing more of the expertise, care and guidance of the store experience into digital channels. For retailers more broadly, it points to a bigger opportunity: using AI to make digital commerce feel less distant and more human.