To sell on Alexa, you need to rethink your entire product strategy | Artificial intelligence

The voice revolution has yet to make itself heard but when it does, expect products to change in response. Considerably.

However, one of the problems with voice applications today is that they both expect artificial intelligence (AI) to be more advanced than it is, and they assume products needn’t change to fit voice.

You can’t sell that on voice

Take commerce, for example. Given that Amazon dominates the voice-activated assistant market with its Echo, one would expect the Echo to be a significant driver of commerce. Nope. As covered in TechRepublic before, just 2% of Echo users have used their Alexa-powered Echo for shopping and, worse yet, 90% of these opt never to do so again.

SEE: IT leader’s guide to the future of artificial intelligence (Tech Pro Research)

Surely one of the reasons behind this dearth of voice-driven shopping is the absence of product information. In person or online, you can see the product (or a representation thereof) and have a good sense of what you’re buying, before you buy. You can read online reviews to see what others have thought. With voice, you get none of this, and are basically shouting commands into a vast warehouse in the sky, praying that the right product will miraculously emerge from the exercise.

It’s hard to imagine anything the leading voice companies (Amazon, Google, and Apple) could do to change this experience. Instead, we may have to see the product companies change to fit into the narrow range of possibilities afforded by voice.

Fitting products into the voice age

Rakesh Agrawal hits this theme perfectly on Twitter, noting that the “Big challenge with voice is to get you to the product you want without false positives.” Of course, it’s not just what consumers want, but also what brands want them to want. Or, as Agrawal has tweeted: “People are more likely to say ‘Buy me a pack of 12 AA batteries’ than ‘Buy me a pack of 12 Duracell AA batteries.’” Arguably, both would satisfy the consumer’s need, but those two voice commands would work out very differently for companies that have spent decades building their brand.

So, what’s a vendor to do? Simplify.

SEE: How we learned to talk to computers, and how they learned to answer back (cover story PDF) (TechRepublic)

As Scott Wessman pointed out, “If transacting via voice becomes common, it could shape the products that get sold. Product designers might make sure their products are not so option-heavy they can’t be said verbally.”

More about artificial intelligence

What would this mean in practice? For a consumer goods company like Tide, it would mean presenting far fewer options, running contrary to the received wisdom of the last few years. As Agrawal put it, “Tide could simplify from 150 SKUs to 5….That would be good.”

Even Apple, traditionally so focused on selling a few, good products, has started to slip into the industry trend of “more choice being more good.” In a voice commerce world, however, the less a consumer has to say to get the product she wants, the better. And so we’re likely to see more companies offer less choice in order to sell more.

Even if AI were more advanced and could better “intuit” what we want, vendors would still need to do this. No one really wants to have a 30-minute discussion with a robot to settle on the just right crayon set. We’ve got better things to do, and if voice isn’t speeding up the purchasing process, it won’t be an integral part of it.

Also see