Talk at Emerce Tech Live in Amsterdam

Last Tuesday I gave a talk at EMERCE Tech Live on the main stage of the Beurs van Berlage in Amsterdam. It was a lovely event and it was fun being back in Amsterdam however briefly.

It was a business focussed practical riff on my ‘Designing Conversational Interfaces’ talk that may have blown some people’s minds. So it goes!

Tech Live! 22

Tech Live! 24

A post shared by Alper Cugun (@alper) on

Now on stage @alper on conv interfaces. The good and bad. The power of simplicity #tel17

A post shared by Iskander Smit (@iskandr) on

A post shared by Iskander Smit (@iskandr) on

ThingsCon Amsterdam 2016 Talk

Tweet coverage of the 2016 Bot Summit at the V&A in London

I was at the 2016 Bot Summit in London a couple of weeks ago. I did my best to capture salient points from every talk in a tweet. Here are all of them in order.

Notes from Interaction16

I’m still waiting for the pictures of Friday to air on the IxDA Flickr and for the video of my talk to be posted, but here are some notes I found around the web about my talk (for my personal notebook).

Eigenlijk veel interessanter om 14:00 is ‘Conversations are the new interfaces‘, van Alper Çuğun (nog een Nederlander!). Als de mensen van Hubbub iets doen rondom experimentele interfaces, dan let ik op. —

Today Alper Çuğun will lead a session about ‘Conversations are the new interfaces’ and Marcel Schouwenaar will talk about ‘Trade-offs & Sharing’. —Embassy of the Netherlands in Finland

Alper Çuğun (@alper), who used to co-run a gaming company called Hubbub, talked about conversational interfaces. I’ve been really interested in these lately so I found it very helpful. He showed an example of a conversational UI that they developed and gave a rundown of do’s and dont’s when designing for them. One interesting insight was that they found that kids will read a lot of text if they get it in SMS-sized chunks. He smartly called conversational UIs the “UI for AI” and cautioned that they are not perfect for every use case. —Aaron Ganci

Another fantastic talk was from Alper Çuğun. He gave an overview of conversational user interfaces. It was based on his own experience making mobile games that use conversations and messaging as the main mechanic. He also shared an analysis of the tools out there to make conversational UIs and a prediction of where the scene is heading.

In particular, I was struck by how conversations, like the messaging apps we’re all familiar with on our phones, can lower the barrier to people using technology. It now feels very natural to text back and forth with friends. When machines can text with us, that might give us a sense of accessibility and agency in our interaction with them. — Michelle Thorne

Talk at Interaction16

I just got back from Interaction in Helsinki having given my talk about how Conversations are the New Interfaces.

I have been blown away by the response to and kind words about my talk. I think this is a conversation (!) worth continuing. Stay tuned for details on that.

Here now is a collection of photos I found of my talk (for my personal archive):

A not so secret Hitler

Rob Dubbin at the Awl has written a comprehensive account of what is wrong with Secret Hitler. I agree with his critique but I want to highlight one issue in particular.

I tried ignoring Secret Hitler but their design notes kept making their way into my twitter. I skimmed through them and found them to be well put together. The last one about illustration and graphic design however convinced me that the game goes well beyond just bad taste.

The problem is that the identity cards for fascist players in the game (shown below) display them als lizards where the liberal identities are shown as human. Fascists are inhuman, get it?


This is simplistic and immoral. If it’s not obvious why, here are three reasons:

  1. Depicting certain groups of people like vile animals is a way of objectifying them and an excuse to exterminate them. One of the lessons of history is that we don’t produce this kind of propaganda.
  2. Depicting the fascists as animals is not a reversal that makes it all right. The fascist of my fascist is still a fascist.
  3. Depicting fascists as intrinsically different from other people and easily recognizable as such is a deeply wrong and misleading fantasy.

This way of thinking is part of an ongoing trivialization of fascism and spreading it is harmful.

As Rob Dubbin says in his piece:

There should be a high bar for invoking this person, and there should be such a thing as falling well short of it.

The people making Secret Hitler are obviously intelligent, skilled and have vast resources at their disposal. I can only guess why they would make a game about this topic and then do it so poorly.

Insurance in the age of big data and personalized tracking

Last week there was some debate spurred by some of the larger insurers of the Netherlands who want to use tracking data to personalize insurance coverage. A piece in the Reformatorisch Dagblad of all places and Rob Wijnberg talking about it at DWDD.

The problem is that insurance by definition is not personalized and we should be protected from each other’s best interests. I tweetstormed about it and have recorded it below.

This is particularly salient from a design perspective if you see the tweets below. What this comes down to is a policy design problem of a vast scale, a level of abstraction up again from service design. People aren’t well equipped to make these decisions for themselves and they probably shouldn’t have to be. They should be aware of which expertise they are lacking and they should know who they can trust. Creating those two competencies are the two hardest problems of our time.

Chat as an important new platform for user experience

Talk about this is increasing all around us (see this piece by Cennyd) and I think it’s time for me to share some of our recent thinking on the topic as well. We believe that conversational user interfaces will be the way that most people will interact with digital systems from the near future on. That can be chat or voice or something else constrained to offer only specific responses or fully freeform. Natural language processing has improved to a point to make this workable and will continue to improve further.

Chat apps are the sine qua non of mobile devices. They are essential, they are everywhere and many of them are cross-platform. People use chat to connect to people but increasingly chat applications are used to interface with non-humans. Chat apps can offer a flat channel to a digital system or facilitate any and all kinds of persistent bots and application logic to be deployed. A great example is this a16z piece on the wide applications of WeChat in China.

The fact that chat apps are cross-platform creates a new smallest common denominator on which you can build applications that are guaranteed to work on all the devices the app runs on. This is a new OS. That people are used to these interactions and normally use them to connect to other people also creates a convenient habituation.

I argue that the bits of conversational logic deployed through chat can be called applications and do most things that apps do.

Most apps allow you to retrieve information or to perform an action. This is glued together with some chrome filled with awkward ever-changing (hamburger menu in or out?) architectures. They need to conform to stringent visual design guidelines while looking recognizably the same across lots of different devices.

Information retrieval and performing actions can be done via chat as well where an AI/bot counterparty will keep track of your context and give you the right cues at the right time. “Slackbot give me a GIF.” “Domoticz turn off the light.”

I am the purveyor of a small app to find good coffee called Cuppings. There is no reason why that same experience could not be delivered through a chat interface. No reason in fact why it could not be delivered better through a chat interface.

Add to that that making good apps is becoming an increasingly difficult endeavour because of device proliferation (mainly on Android), API bloat (on all platforms) and increasingly high visual and interaction design standards. Increasingly making a pixel perfect app that feels nice and works well is something that only larger companies can afford.

Most of the effort we spend right now into user interfaces could be moot if the experience would be delivered through a chat interface. That every app has a different UI and information architecture and that it has to be learned anew is a huge impediment to its adoption. We have recently built several chat based apps & games inspired partially by Lark. During testing we found that users don’t need to be explained anything because they are so familiar with the paradigm.


Chat is here to stay and I’m incredibly excited to see how far we can push this new medium.