Voice Commerce

Voice Commerce translates as “voice trading.” It means shopping where you don’t type or click, but speak. For example, you say “Order me a pack of coffee” into a small speaker in your living room, and the order goes out. Such devices listen for a wake word like “Alexa” or “Hey Google” and only react afterward. The process is the same as with a normal online shop: select a product, pay, have it delivered. Only the interaction happens via voice instead of a screen.

Why retailers are betting on voice

Speaking is faster than typing. A person speaks around 150 words per minute, but only types about 40 on a phone. For retailers, this means less friction between the desire to buy and the order itself. And the fewer hurdles there are, the more often purchases happen. This is especially an advantage for things people buy again and again.

There’s also an economic point that is often overlooked. A voice assistant doesn’t read out a results list with twenty offers. It usually names a single product. Whoever gets that one spot essentially captures the entire sale. For brands, this is fierce competition, and for the operators of the assistants, it’s a position of considerable power.

Still, it’s important to put this in perspective: Voice Commerce is so far a niche. The vast majority of online retail continues to run through apps and websites. Voice orders work well for simple repeat purchases, but poorly for anything you’d want to see beforehand.

From spoken sentence to order

The process happens in several steps. First, the device converts the audio recording into text. This step is called speech recognition. Then a program tries to understand what is meant: Is this an order, a question, or a cancellation? The important details are also extracted from the sentence, namely product, quantity, and brand.

The system then checks the retailer’s inventory. It verifies whether the product exists and what it costs. If information is missing, the assistant asks a follow-up question. In the end, the response is converted back into spoken language, usually using an artificially generated voice. Only after confirmation is the order actually placed.

Newer systems use large language models for the understanding part. These are programs that have learned from huge amounts of text how language is structured. This means even imprecise sentences like “the detergent from last time, but the big pack” work. Older systems, by contrast, required commands given almost word-for-word.

Where voice ordering happens

The best known examples are smart speakers from Amazon, Google, and Apple. They sit in many kitchens and living rooms. Typical purchases are detergent, pet food, or beverages—things that need no explanation. Food delivery services and taxi apps can also be controlled by voice.

A second area is the car. There, you can’t take your hands off the wheel, so voice is the natural way to interact. Several manufacturers are now building payment functions for fuel or parking directly into the vehicle. In call centers, too, voice systems are taking over simple orders and reorders.

In business news, the term usually comes up in connection with data protection and market power. A device waiting for a wake word is, technically speaking, listening at all times. Consumer advocates therefore ask which recordings are being stored. A second point of contention is preselection: when the assistant suggests only one product, the platform operator has a say in which brand even gets a chance.