Are killer bots about to do away with smartphone apps?
Is the smartphone app doomed?
To look at the stats you wouldn't think so: Apple has two million of them in its App Store and Google Play has a few hundred thousand more than that. Total app downloads have passed the 150 billion mark.
But some are wondering whether apps are about to be replaced by something smaller, smarter and faster.
These programs, thanks to AI [artificial intelligence] software in the cloud, can chat to humans via text, extract the meaning and then act on it.
They are little digital helpers.
Any time you see a live chat box open up on a retailer's website, or order a taxi or flowers through chat platforms such as WeChat and Facebook Messenger, you're most likely talking to a bot.
Despite the vast choice of apps open to us, the average number we use is 27, according to research by Nielsen. This hasn't changed for years.
And the problem with apps - and their seemingly endless updates - is that they eat up our smartphone storage capacity alarmingly quickly.
Developers often get a raw deal as well. One estimate suggests that 94% of the cash generated by apps in Apple's App Store goes to just 1% of publishers, and those firms also get 70% of all downloads.
"One of the worst things about the App Store is the App Store itself, because it's such a walled garden," says Ted Nash of Tapdaq, who was a veteran app developer while still a teenager.
Apple's oversight of all apps slows down development and forces programmers to include specific chunks of code that look after adverts, usage statistics and other metrics, he says.
Add to this the trouble of making apps work across lots of different devices and keeping up with changes to Apple software, and it's no wonder some people are disillusioned, he says.
So is app fatigue setting in?
The joy of text
"Apps used to be the big thing," says Kriti Sharma, head of mobile development at accounting software firm Sage. "But many more people are messaging than are posting on social media these days."
This is why she thinks bots are the natural successors to apps - the interface is instantly familiar to customers.
Ms Sharma started her coding career at Barclays, where she co-created its Pingit banking app and oversaw its mobile portfolio.
For companies or brands that want meaningful interaction with customers, a conversation mediated by a bot could work well, she believes.
Sage is developing a bot called Pegg that acts as a smart business assistant. It will help small business owners keep track of outgoings and expenses, making tracking cashflow easier.
"Bots don't have to be super-complicated," says Ms Sharma. "But over time they must add a lot more value for a customer."
Bots are more credible because good progress has been made in writing artificially intelligent software, she says. And also because many companies now have huge amounts of data they can use to fine-tune bot responses.
Another advantage bots have over apps is the speed with which they can be developed, deployed and updated, she argues.
'Bots are the new black'
This growing interest is being inflated by work at Facebook, Microsoft and Google, as well as by newer firms such as Slack and HipChat. And start-ups such as Begin, Growbot, Butter, Wisdom and Operator are also helping to take bots mainstream.
One catalyst for the interest was Facebook's announcement earlier this year of a bot framework that streamlines the bot-creation process.
One report suggests that this massive amount of interest has unleashed a $4bn (£3bn) flood of venture capital funding into big and small bot developers.
"Bots are the new black," says Jon Moore, chief product officer at rail ticket booking service, The Trainline.
Although most people now use The Trainline via a smartphone and many regular users have installed its app, the company is keen to investigate what bots can do, Mr Moore says.
For booking train tickets, a website or an app is profoundly better than using a bot, he maintains, but there are times when an app falls short and a conversation handled by a bot may be better.
"We're just at the point of saying it's another interesting piece of technology," he tells the BBC. "We expect that they are going to be useful to us, though it won't work for every context and circumstance."
Tapdaq's Ted Nash warns that though bots might look straightforward, they're not necessarily an easier option.
"A bot is a much more simple technology from a customer perspective, but the AI that powers it is immensely complicated to do," he says.
That difficulty often means that bots are pretty crude.
"A lot of them now have pre-defined inputs and responses," says Mr Nash. "The only way they are going to become truly ubiquitous is when they can respond as a human would."
But even before they do that, says Nick Lane, chief analyst at consultancy MobileSquared, bots are likely to be useful for smoothing out the interactions between customers and companies.
"We'll see bots helping out with customer engagement, queries and product enquiries," he says.
But there is danger in relying too much on a technology still in the early stages of development, he warns.
"Some companies are wondering if they can put their business and reputation in the hands of a computer program."
There is another reason why technology firms are keen to use text-based chatbots, says Mr Lane.
"It could be that they see this interaction as another form of data mining," he says. "People should ask how that information and conversation is going to be used.
"One way or another there is a model evolving around that communication that will see it being monetised."