Start building your own chatbot now

The beginning of apps

I am old enough to have seen apps in their infancy. By now, I am getting so old that I can’t even remember if apps went through the Hype Cycle. But all I can think of when I think of apps in 2009 and 2010 are… farts apps. That’s right.

You press a button and your phone makes the sound of, well, you know what I mean. At this time, other kinds of popular apps were created, like soundboards (press a button and your phone plays back a line from a movie or cartoon) or beer drinking apps (tilt your phone and it looks like you’re drinking a beer). You could even press a button and your phone would make the sound of a razor and you could pretend to shave yourself. So useful.


Time went by and we saw apps like Dropbox, Evernote or Angry Birds grow huge and change the digital landscape. From useful apps to fun games, built by companies we never heard of before. Saving us time, frustration we didn’t even know we had, reinventing the meaning of mobile gaming and communication.

The golden age of apps and the beginning of bots

Flash forward to today: we have apps you can’t even call just apps anymore. Uber, Airbnb and the likes are multi-billion dollar companies.

They exploded in a couple of years, using an underlying technique that hasn’t changed that much. Sure, the Galaxy S7 is way better, faster and slimmer than the Galaxy S1. But the essence is still a touch screen phone with apps. When you look at iOS in 2010, it looks eerily similar to iOS 10 (which will launch this fall).

And here come the bots. Chatbots. As apps have matured to a point where we keep using the same ones (WhatsApp, Messenger, YouTube, Maps) and new killer apps don’t seem to emerge anymore, bots could save us all. At least, that’s the impression I got when I lived in San Francisco earlier this year. Both Facebook and Microsoft announced a bot framework in the beginning of the year, and it was certainly met by a lot of hype. Yes, hype is the important word here.

The Hype Cycle

The first examples of bots on the market seem to underwhelm. So much so that I can’t help but think back to the fart apps: chatbots are currently in that stage. The technique is there, but the ideas aren’t up to the Uber level just yet. Natural Language Processing and Machine Learning have gotten pretty impressive, but remain tricky to adapt to everyday life. Today, you can use bots to order a pizza, but it is sometimes slower than using an app. That means people can give up after getting messages like “Didn’t quite get that, buddy. Please tell me where you are flying from”.


Thinking back to the hype cycle, we have already seen the Technology Trigger and the Peak Of Inflated Expectations for bots. All that is left now is the Trough Of Disillusionment: after the excitement that this new technology created, people realise that it isn’t what they thought it’d be. Luckily, this period is followed by the Slope of Enlightenment: that’s when creativity kicks in. That’s when people think of use cases for bots that are actually useful and will see the Dropboxes and Evernotes of the bot world emerge. And as soon as we hit the Plateau Of Productivity, we will have Uber level bots.

The technique is there. It is up to us to make it happen. 

How will bots reach their own golden age?

As much as I’d like, I can’t predict the future. I don’t know who the winners of the bot race will be and which ideas will prevail. However, I can look into the past. To a time long before even Nokia and BlackBerry ruled the world. To a time when we transferred from one technology to a sort of similar, though much better and more potent technology. The progression from the mechanical age to the digital age. The progression from the typewriter to the PC.

A small company from Albuquerque, New Mexico in the US created the first device worthy of the name personal computer.  The Altair 8800 hit the market in 1974 for $395. It was an instant hit among nerds, geeks and computer enthusiasts. So much so that by 1977, more than 30 companies in the US alone were making personal computers. Among them, they weren’t very consistent. Some could only write in ALL CAPS, some lacked a mouse, and if you printed something on the Apple II, the font on paper would be different from the one you saw on the screen.


The defining moment came when the IBM PC was introduced in 1981. It wasn’t a technical wonder. It wasn’t better or more advanced than the others. But it combined familiar elements that had proven themselves in other sectors: a TV monitor, a standard disk drive, A QWERTY keyboard (or AZERTY or QWERTZ if your German or French), an Intel 8088 chip, open architecture and last but not least Microsoft’s MS DOS as the operating system. Together, these elements came to define the idea of a personal computer. IBM quickly grabbed a market share of more than 30%, thus forcing other manufacturers to design their PCs in the same way, making the IBM PC the dominant design. The dominant design doesn’t meet the needs as much as a customized design, nor is the best technical achievement. But if it doesn’t optimize for the few, it satisfies the many.

The dominant design makes features and accessories essential. Before the IBM PC came along, people didn’t know what to expect from a PC. But after its rise, they knew, and accepted the use cases of a personal computer. They expected their computer to come with a mouse and a QWERTY keyboard. Any other way would have been odd and unsatisfying. IBM managed to dictate the codes of personal computing. And they’re still relevant today! The design IBM created decades ago is still the norm, even if IBM hasn’t been a major actor in the PC market for quite some time.

The dominant design does one thing very well: it dominates. It becomes the de facto design and forces other manufacturers to adjust, when most won’t be able to make the jump from their design to the dominant design.

Today, we’re seeing different approaches for bots, different designs like natural langage based answers or buttons. Users don’t know what to expect from a bot, they don’t know how to talk to one or what answer to expect. Until the IBM PC of bots comes along, consumers will be in a technological blur and bots will remain a high tech innovation without real applications. One reassuring thought is that the inventors, tinkerers and innovators starting in the beginning of the Hype Cycle, before the establishment of a dominant design, are the ones who will survive.

History is being made as we speak. Exciting times.

Lex Oudijk – Founder of

Want to build your own conversational bot? Get started with SAP Conversational AI !

Follow us on
  • sfgower

    I have looked around, and don’t sense that we have reached the trough of disillusionment.

    I recently sat in a big room full of people listening to an abysmal talk from a major vendor on their conversational API. The room was full, and people seemed to very interested. What they didn’t seem to realize is how the vendor’s claims were based on extremely limited approaches. It was a hype pinata.

    And that is why I think why I think we are still many months away from the trough of disappointment.

    Of course progress can be made. Meanwhile, avoid hype pinatas no matter how pretty and colorful they are.

    • Justine Baron

      Hi! That’s a good observation, and I have to agree with you. I believe we’re kind of in a paradoxal period concerning chatbots. The core sector, who’s been experimenting with chatbots for more than 6 months, is facing the disillusion that chatbots can do everything right now. No they cannot, and they’ve experienced it first hand by testing out (and maybe building) many chatbots, many platforms. However, as you explain it, people who are not working in this specific market are still extremely impressed (as they should be) by the possibilities of bots and what they do today. They’re craving for more POCs and want to try everything out. These people are still in the upward slope of discovering the tech, and yes, it’s our job to manage their expectations and provide amazing products.

This site uses Akismet to reduce spam. Learn how your comment data is processed.