This article is more than 1 year old
Billions of 'custobots' are coming online. Marketers may need to learn SEO for AI
You can't take a machine to lunch or put it to sleep with PowerPoint, but you can try to shape its worldview
Billions of new buyers of goods and services will emerge in coming years, but salespeople won't be able to woo them all with slideware, gifts, football tickets, or lunch – because many of these new customers are machines.
So say Gartner distinguished VP analyst Don Scheibenreif, and Mark Raskino, a fellow in in the analyst firm's CEO and Digital Business Leadership research team, in a recent book titled When Machines Become Customers.
As explained to The Register by Scheibenreif, machines-as-customers – or "custobots" as he likes to call them – are easily understood. They are like HP's Instant Ink subscription service, which sees printers advise the vendor when they're low on consumables and place an order for fresh supplies.
The analyst believes bilions of custobots will come online in coming years, and that they will be immune to marketing's appeal to emotions.
"Say, for example, I'm a manufacturer and I'm selling my widgets to a retail store. What if that retail store decides to outsource its negotiation function to a machine?"
- Google Chrome Privacy Sandbox open to all: Now websites can tap into your habits directly for ads
- 'Millions' of spammy emails with no opt-out? That'll cost you $650K, Experian
- Google rebrands 'android' as 'Android' to remove any doubt about its affiliations
- Infosys launches 'sonic identity' – an aural logo to 'reinforce brand purpose'
A former marketer, Scheibenreif admitted that in his former life – like all marketers – he would create fear of missing out. "We [marketers] use a really cool spokesperson, or we have something else that grabs your attention. But if I'm trying to sell that product to a machine, the machine doesn't care if Beyoncé is my spokesperson."
Marketers and salespeople will therefore need to find different ways to convey information to those machines, which crave data to inform their decisions.
Scheibenreif said he's aware of an Italian vendor creating a digital marketplace for machines to shop at – so that industrial machinery that has been programmed to order commodity consumables such as lubricating oil could advertise its needs to the marketplace and then assess bids in a kind of entirely digital reverse auction.
Another way machines become customers involves consumers creating bots to do their shopping – perhaps by having AI turn a list of meals into a shopping list. Scheibenreif said that process could be informed by a human advising a personal model of their preferred brand of commodities like pasta, but that generative AI could reach its own conclusions about brands individuals might find appealing.
Marketers fear this scenario, he said, because they will "need to figure out what the algorithm is to get into your shopping basket" – meaning how to convince the bot that one product is somehow more suited to your requirements than another. "Because if I can't figure that out, and if [the] product is not differentiated enough, then it's going to be a race to the bottom on price."
"Marketers that I talked to about this are not happy because some of their differentiated stuff is really not differentiated enough. It's all the same. It's all marketing."
One tactic marketers may need is to influence the corpus of AI models. The aim would be to ensure the bots are keenly aware of the detail of their products – an effort akin to search engine optimization, but with AI models as the target.
If products rank highly enough, they could attract the attention of a bot consulting a generative AI for advice on the best pasta to go with your intended dinner, or the best oil to keep a physical robot in robust condition. ®