Leading flooring retailer answers 83% of employee product questions with its AI agent

This retailers existing knowledge base held the information new and experienced colleagues needed, but having the information and being able to find it as they worked with customers were two different problems.

• New colleagues faced a steep, high-stakes learning curve. Flooring is complex and personal, and sales colleagues need deep product knowledge to earn a customer’s trust quickly in the middle of a live sales conversation, not after a lookup.
• The range of products worked against speed. New sales people often struggled to absorb the breadth of products well enough to guide a homeowner confidently, especially soon after joining.
• Experienced colleagues had knowledge but not always the habit of using self-service tools. Long-tenured staff carried years of institutional knowledge, but when processes changed, they didn’t consistently turn to the tools built to help them adapt.
• The knowledge base required knowing how to search. Like many knowledge repositories, the answers existed somewhere in the system, but employees had to already know the right terms to find them, which meant the knowledge was often unavailable at the exact moment someone needed it on the shop floor.
• These gaps repeated at scale. With hundreds of employees spread across 175-plus locations, any friction in onboarding or knowledge access multiplied across the business rather than staying contained to one store.

Their objectives were specific: accelerate onboarding, increase retention of new colleagues, and foster a genuine self-service culture, all in service of the ultimate goal of great conversations with customers. The team knew that to get there, the tool itself needed to feel different from a standard search engine or chatbot, or colleagues simply wouldn’t adopt it.

Quiq helped them built a fun, quirky, brand-aligned AI agent with its own personality, to do that work. The build included:

• Pulling knowledge directly from their learning and development platform, so AI agents answers stayed grounded in the same material the training team already produced.
• Optimizing that knowledge specifically for AI consumption, rather than repurposing content written only for human readers browsing a page.
• Automating new knowledge ingestion, so updates to policy or product information flow into the AI agents answers without the team manually rebuilding content, unlike the previous system.
• Giving their AI a distinct personality, aligned to their brand, designed to make colleagues want to return to it rather than default to asking a manager.
• Serving three distinct audiences from one platform:  website visitors, contact center agents, and store retail staff, each drawing on the same underlying knowledge but controlled for the right context and channel.

Under the hood, when someone asks the AI agent a question, multiple language models work in parallel to classify the topic, identify key attributes, and extract the specific entities in the question. Then the AI agent then queries all the resources, including its knowledge base, product catalog, order lookup, and account information, to gather what’s actually needed to answer. Before any response is generated, guardrails keep it tied to company resources, templated language, and trusted links, and every answer passes through accuracy checks and hallucination detection before it reaches the person who asked.

Rather than producing static content and hoping colleagues found it, the Training team now has a tool that puts the right answer in front of employees at the moment they need it, whether they’re new to the business or years into their tenure. Internally, colleagues have responded well to working alongside the tool, and the team has already identified where to expand its knowledge next based on what people are actually asking.

In its early stages, the AI agent is already showing the Training team exactly where employee knowledge gaps sit and how well the AI agent closes them.

• 83% of product questions were answered accurately, with the remaining 17% being addressed by adding more knowledge to the AI agents resources.
• 55% of all inquiries were product-related, giving the team a clear signal to prioritize expanding product knowledge categories next.
• Knowledge updates now flow from the Learning & Development platform into the AI agent automatically, reducing the manual content-maintenance burden the previous system required.

This retailer and Quiq are already planning the next phase: feeding the AI agent more detailed product knowledge, teaching it systems, process, and policy knowledge beyond onboarding, and having their own team learn to use Quiq’s AI Studio to keep improving it directly.