Ecommerce stores have come a long way in terms of search and filtering, yet one fundamental problem remains: shoppers do not think in categories and attributes. They think in needs. In a physical store, a sales representative can step in, ask a few questions, and help the shopper reach a decision faster. On the web, that guidance is usually missing, and the shopper is left to navigate the journey alone.
I need something warm for a weekend trip. I’m looking for a gift for someone who already has everything.
Standard navigation was never built for that kind of input – and that gap is exactly where the Levi9 team decided to build something.
Starting from a real platform shift
The idea came from watching where Salesforce was heading. Agentforce had been positioned as the platform’s strategic AI direction for commerce, and the team wanted to explore what that actually means in practice – not as a concept, but as a working storefront experience. Guided Shopping stood out as the right use case: it sits at the intersection of customer discovery and purchase intent.
The team already had strong roots in the Salesforce Commerce Cloud ecosystem – attending webinars, following platform updates, tracking new capabilities as they appeared. When Agentforce and Guided Shopping for B2C storefronts surfaced as viable options, the decision to invest time into building it themselves followed naturally.
Building inside the ecosystem, not around it
One of the early decisions was to use Salesforce’s own AI layer rather than connecting external tools. Agentforce provides a native way to tie AI to B2C Commerce data, storefront actions, and the broader Salesforce platform – which matters when the goal is a solution that fits how Salesforce Commerce Cloud customers already operate.
Under the hood, the solution uses Einstein generative AI to handle natural-language interaction with shoppers. The agent understands what a customer is looking for, grounds its responses in real storefront data – product catalogs, customer records – and guides the shopper through discovery and purchase. In practice, that means the Guided Shopping agent can do more than recommend products: it can answer common shopper questions about materials, return policies, and order details, while also helping customers take action directly in the storefront, including adding products to the cart. The result is a more intuitive starting point, especially for first-time or undecided shoppers who benefit most from a conversational approach.
The project required a specific combination of expertise: Salesforce Commerce Cloud knowledge, Salesforce Platform and Data Cloud (now Data 360) experience, and solution architecture. Understanding storefronts, APIs, and customer journeys was not enough on its own – the team also needed to know how emerging AI capabilities map onto enterprise systems.
Where theory meets real implementation
Like many new platform capabilities, Agentforce and Guided Shopping are promising – and demanding. Making the experience work smoothly across real storefront architectures was the central technical challenge. The team entered the project as Salesforce Commerce Cloud specialists, but the integration required Salesforce Platform and Data Cloud knowledge they had to develop under a tight deadline.
Salesforce’s own support teams were still getting up to speed on this feature, which meant the team had to navigate implementation decisions with limited external guidance. They worked through the available documentation, mapped out what Salesforce said the platform offered versus what was actually configurable, and built from there.
The project also gave the team room to experiment with new development approaches – including vibe coding, an emerging method of AI-assisted development – and to refine how they design instructions for AI agents to better interpret customer needs.
What this kind of work actually looks like
The broader point this project makes is about how AI fits into enterprise software. It is not a standalone feature dropped into an existing product. In this case, it connected AI, commerce logic, integration work, and storefront engineering into one coherent implementation. That is the kind of practical, platform-grounded AI work that Salesforce Commerce Cloud customers are increasingly asking for – and the kind Levi9 is now positioned to deliver.
***This article is part of the AI9 series, where we walk the talk on AI innovation.***
Interested in bringing AI-driven shopping experiences to your Salesforce Commerce Cloud storefront? Contact Levi9 to connect with our Salesforce team and explore what Agentforce can do for your customers.
In this article:
Levi9 Serbia
