Video: QVC Increases Conversion and Engagement With Bluecore’s Shopper-Facing Agent, alby

Brands and retailers have been talking about – and envisioning – the future of shopping with AI for years. With generative and agentic AI, that future is already here – and at QVC, these new technologies are driving results. QVC is testing AI with curiosity and intentionality, with the aim of discovering how agentic AI will resonate with their typical customer – a woman who is 50 or over. 

Onstage at eTail West in Palm Springs, Marie Fleisher, Director of Digital Optimization and Operations at QVC, and Max Bennett, Bluecore Chief Product Officer and co-founder of alby, showed how QVC is using agentic and conversational AI to solve challenges that benefit the customer – and the company. 

The results are clear: On QVC’s product detail pages, alby’s AI shopping agent drove a 1.5% increase in conversion rates and 6.2% increase in engagement rates.

Here are four challenges that QVC was able to address with alby.

Challenge #1: Surfacing instant, friendly expertise everywhere

QVC’s streaming and television presence creates a unique relationship between the brand and its customers. Shoppers call in and ask questions, and hosts answer them live. Shoppers typically don’t get that kind of instant gratification online.

Solution: To help recreate that sense of in-the-moment, readily accessible information, QVC put alby on all its product detail pages, providing suggested questions for every item. For a wheeled piece of luggage, those questions might be “Does it have a bottom handle?” and “How do you clean it?” The AI shopping agent also answers questions posed by shoppers (perhaps, “Will it fit in an overhead bin on an international flight?”). The quick, accurate answers mimic the experience of asking a question to a host and getting a friendly, helpful reply.

Challenge #2: Eliminating product information overload

QVC has a large volume of information on each of its products, often including video and transcripts. But real estate on product detail pages is highly valuable, and each customer may need different data to feel comfortable purchasing an item. That makes it difficult to figure out which information should be most prominently displayed. 

Solution: Alby’s platform is uniquely able to digest all of that helpful information and content and make it easily accessible to the customer.

On the product detail page for a stand mixer, for example, shoppers will see a variety of attachments that go with the mixer. So alby’s pre-populated questions include those specific to the attachments, such as, “How do you use the wire whisk?” along with more general questions such as, “Does it come with a warranty?” 

The AI shopping agent can explain that the wire whisk is great for making whipped cream at home, and then go a step further by supplying a recipe that uses the whisk. The combination gives the customer a much more holistic idea of how that stand mixer might fit into their kitchen and their life.

The health and beauty category is another good example of the utility of agentic AI. Every customer's skin, and their response to makeup and skincare, is different. So alby can answer questions such as “Is this good for oily skin?” and “Is this foundation buildable?” in addition to providing information about ingredients.

Challenge #3: Making sense of product reviews

Reviews and other user-generated content are hugely important to shoppers, but the volume can become nonsensical, with thousands of reviews on some product detail pages.

Solution: alby summarizes user generated content in an easy-to-read user interface. These summaries balance the pros and cons of a product, faithfully reflecting customers’ opinions. For a skin moisturizer, alby notes that it’s “generally well-received for its hydrating and non-greasy formula… it is praised for its ability to improve the appearance of dry patches and wrinkles.” But alby also notes that some people report irritated skin, “possibly due to the glycolic acid content,” and that others find the fragrance too strong.

Shoppers read reviews because they expect the unvarnished truth. If an agent doesn’t respect the integrity of the reviews, then it’s not helpful to the shopper or the retailer.

Challenge # 4: Understanding where shoppers get stuck

With hundreds of thousands of items, it’s impossible for QVC to be deeply familiar with the content available for each product, making it challenging to provide an excellent content experience for every customer. And it’s hard to know how that experience could be improved.

Solution: alby’s insights dashboard helps brands discover their most relevant content. With a particular pair of lab-grown diamond studs retailing for about $4,000 – a high consideration purchase – the insights dashboard showed that customers were asking if the earrings were hypoallergenic. 

This sort of data mining helped the QVC team better understand enhancements they could make to product descriptions, improve product assortments, and even help make on-air presentations more compelling by including key selling points that might otherwise get overlooked.

Preparing for success with AI agents

Fleisher and Bennett said that alby generated positive results for QVC across almost every product category and on all platform types. They also noted that those gains didn’t come on day one. The placement of alby is important – if someone is trying to add an item to their cart, you don’t want to distract them.  QVC continuously tests and refines its agent configuration to create a meaningful impact for shoppers.

QVC also built their own guardrails for alby. They created an internal task force that included members of the legal team, and came up with boundaries for questions alby would and would not answer. Alby is designed to make it easy to embed these guardrails, upload information, and fine-tune the system, further ensuring a high-quality and on-brand customer experience that both QVC and their shoppers can trust.

To get started with a free alby proof of concept, visit us here.