75% of Our Site's Customer Queries Handled Without a Human

Most ecommerce websites assume customers already know what they want to buy. They search for a product, browse categories, apply filters, compare options, and eventually make a purchase. But what happens when customers know what they need but don't know which product is right for them?

A customer planning a trip might need a waterproof jacket but be unsure about which material or features to look for. Someone setting up a home office might need a sofa that fits a specific space and supports long hours of video calls, without knowing which product specifications matter most. In both cases, the challenge isn't simply finding a product, it's figuring out which product best meets their needs.

This is where the Velocity’s AI Shopping Assistant makes a difference. By allowing customers to describe their requirements in natural language, it helps them discover relevant products, compare options, and get answers to pre-purchase questions without navigating endless categories and filters. It transforms ecommerce search from a process customers must figure themselves into a guided shopping experience that makes buying easier.

Why Is Ecommerce Search Bigger Than Finding a Product?

The traditional ecommerce search is best suited when the shopper is familiar with the product structure of the retailer.

The typical journey looks like this:

Search → Category → Filters → Product pages → Compare → Decide → Buy

There's one issue with that trip. It puts an awful lot of burden on the customer.

The shopper must be aware of the keywords that are required, know the product jargon, know the correct filters and know the differences between products.

How the Velocity’s AI Shopping Assistant changes the journey

Conversational product discovery

Instead of relying on exact keywords or navigating multiple categories, customers can describe what they need in plain English. 

• Search naturally: Customers can explain their needs without knowing the exact product name or technical terminology. 
• Discover relevant products: The assistant uses information from the store's catalogue to identify suitable options. 
• Simplify product discovery: Customers can narrow down their choices through an interactive conversation rather than browsing endless search results. 

Recommendations based on requirements

Product recommendations become more useful when they reflect what customers actually need, rather than simply promoting popular products. 

• Understand customer preferences: The assistant considers requirements such as budget, size, intended use, material, compatibility, and performance. 
• Personalise recommendations: Customers can communicate multiple requirements through a conversation instead of manually applying individual filters. 
• Match needs with products: The assistant helps customers identify products in the catalogue that align with their stated requirements, making the decision-making process easier. 

Questions about a product before purchase

Customers may have more questions about the products they want to purchase. With the use of our AI Shopping Assistant we can help answer those questions. Using product information it has access to.

• Product Specifications: Allow customers to ask questions about the products specification.
• What's the difference between products: Allow customers to ask what differences there are between products.
• Will this product fit my needs: Customers can ask if the product will suit there needs.
• Decrease amount of customer support enquiries: Instead of looking through product pages, faq's or google and contacting an agent for basic questions.

Where conversational shopping has the strongest business case

Not all ecommerce sites require an AI shopping assistant.

Traditional navigation might be enough if a customer is well acquainted with a product they seek and buys it after just a single search.

The opportunity increases when there's a choice to be made as to which product to purchase.

Fashion and apparel

The customer might require guidance with regard to fit, material, event, weather and style. These preferences may not be easily represented by regular filters.

Electronics

There are many technical specifications to think about when purchasing a product. The customer may know what they want to accomplish but may not know what specifications or products fall under that category.

Home and furniture

Customers may have multiple restrictions to consider, such as dimensions, room type, style, material and functionality. A chat can offer a more organic place to start than sifting through numerous combinations of filters. 

Sports and outdoor

There are many products to choose from depending on the sport or outdoor activity you want to do or your surroundings. Customers can describe what they are looking for without knowing the product and have it recommend to them.

B2B ecommerce

Conversational discovery can be particularly effective for B2B buyers who are familiar with the application or job they are looking to solve, but not exactly sure of the SKU or technical specification.

Across all these sectors, the common problem is the same:

The customer needs assistance in deciding on a product.

How Do You Measure an AI Shopping Assistant’s Effectiveness?

Just because you have an AI Chatbot with thousands of conversations doesnt make it a good chatbot.

You need to look at what is happening from these conversations and measure them pre and post.

Query resolution rate

How many customer questions get answered successfully without any human intervention? This is one of the more obvious ways of measuring automation.

Human escalation rate

When will you still need a human involved? As shown above there are many instances where your product data can be improved on or where a human handoff may be beneficial.

Product discovery and engagement

Assisted shoppers can check if they are able to locate product pages and compare fewer irrelevant products or move faster towards the purchasing process.

Conversion Rate and Average Order Value

Your Shopping Assistant should be creating value to your customers commercial journey. You can measure if your improving on product discovery by measuring Assisted Conversion and AOV.

Customer experience

Other factors such as response time and customer feedback are important, too. If an assistant can answer the question in a timely fashion, but gives poor recommendations, it isn’t solving the problem.

What matters is the impact on the business, not AI activity.

What this case study tells us about ecommerce AI

The most interesting thing about the 75% answer is not that AI can answer customer questions.

Of course, modern AI can do that!

But the more significant lesson is that product discovery can itself be taken into conversation.

Not all customers are interested in conversing with a chatbot just for the sake of it. They don't want to learn all the parts of the retailer's catalogue to find the right product.

This alters the way businesses ought to consider AI in ecommerce.

The question isn’t ‘Where can we add a chatbot?’

It’s:

“Where are your customers having a problem picking a product?”

If you have many customers asking questions that can be automated such as what product they should purchase. That is an opportunity that can be calculated if its by your customers. Helping them choose a product that may be similar to another product.

Conclusion: The best ecommerce AI removes buying friction

The Velocity's AI Shopping Assistant managed 75% of customer inquiries without an agent's intervention. That is a good operational outcome, but more important is the shift in customer interactions with the catalogue.

With traditional ecommerce search, the customer is expected to know what he/she wants to find.

With conversational shopping, they can tell you what they need to know first.

As product catalogues grow, products become more specialised, and customers grow more at ease with AI in product research, that distinction becomes even more crucial. Even with the traditional site search, there is already evidence of poor performance from independent ecommerce research, and recent industry data reveals a rapidly increasing level of AI-driven shopping behaviour. The 75% number reflects the capacity to achieve when these pieces are combined.

Better search boxes might not be the future of ecommerce search. It could be a better conversation.