Interaction 17 — A Review

Interaction 17: the happiest time of the year for me as an interaction designer. I call it my very own “Professional Thanksgiving” — because it usually involves traveling to see family. This is what many of the friends I usually see every year at this conference have become.

This time I didn’t have to travel as I usually do. The conference took place right where I live this year: New York City.

This year, I got my my fill in all things chatbots, conversational/natural user interfaces and machine learning (among other things). Design & tech continue to evolve, and Interaction reminds us exactly how each year.

Before the all the talks begin, there is an entire day full of interesting workshops. I attended a great conversational UI workshop at RGA, in which we got to build our own chatbots and Alexa skills. You can read my in-depth review of it on

What follows is a collection of my very own personal highlights. It will give you an insight of how I experienced the conference. However, I encourage you to check out all of the conference videos whenever they come out.

Day One

My highlights for the first day of Interaction start with Gary Hutswit and his talk “Unframed: Crafting Audience Experience in Cinematic VR”. You may know him from the Helvetica and Objectified documentaries. If you know a designer, chances are he/she has mentioned these at least a couple of times.

Gary Hutswit

Gary touched on a few projects he’s been involved with using virtual reality; highlighted some of the pros/cons; reminded us about the importance of storytelling and how it relates when telling stories in VR; recommended to always craft a balance between showing and telling; and even went on to showing how prototyping could be done when creating a VR experience.

With VR being one of the hot topics in technology, I’m sure this talk was more relevant than ever before.

Next, it was Brad Crane & Jon Mann with their talk: “Suspending Disbelief: Immersive Designs for The Future”.

In this talk, they compared designers to magicians: those who rely in making people believe in stories or experiences no one has ever seen before.

Brad Crane & Jon Mann

They went over some rules to have in mind when crafting such “magic”: setting the stage (critique, storyboarding), questioning known tricks (undoing the bias of past solutions); creating the tools (making the right medium for the message) and, being careful — because “magic can be dangerous”.

After some delicious Greek buffet for lunch, I chose to attend all the chat bots/conversational UI talks.

The conversation began with Paul Pangaro and his talk called “Conversation is More Than Interface”.

He touched on various points relating conversation to interaction design itself, but the part I found the most interesting from his talk were the questions he asked for designing conversations (and CUIs): Is the message clear? What does the UI know about the user’s context? What more can it know automatically or from users input? Does the CUI evolve during engagement?

Paul Pangaro

He also gave a few examples of the nature of the structure of a conversation. Paul mentioned the one in which a direct command is involved where A sets the terms for B: “Alexa, news form NPR”; a more guiding approach in which A sets the goal but lets B decide the means to reach it: “Alexa, some news please…”; and finally a more collaborative approach, in which A and B decide on goals together: “Alexa, how about I listen to something?”.

Next up was Elizabeth Allen with her talk “Adventures in Conversational Commerce”. In it, Elizabeth shared a few things she learned about bots from working on them at Shopify: People might expect your bot to suck; bots can reply instantly and this can be seen as “pushy” in the contexts to eCommerce; bot personality is a thing; and putting users in charge of bots is complicated.

Elizabeth Allen

The following talk was by Whitney French, called “Shortcut to Chatbot Emotional Intelligence”. One of the things that most caught my attention was when she pointed out that emotional intelligence becomes a UX problem when it comes to building chatbots. She also pointed to 5 metrics to have in mind when designing bots: Intelligence (the brain of the conversation); Flow & Cadence (conversational language); Helpfulness (anticipation of needs); Personality (the hook of the conversation); and Utility (the heart of the conversation).

Whitney French

After Whitney, came Greg Vassallo from Fidelity Investments and his talk “Prototyping Conversational UI”. He highlighted the importance of conversational UI when it comes to accessing the power of artificial intelligence. He then proceeded to share a number of tools to learn more about (and prototype) conversational user interfaces. Some of them are Botsociety (to export video of a chant interaction); Poncho (weather CUI); Motion AI (for creating modules for CUIs); eCommerce bots (like the Jessie app); among a few others. Lastly, he touched on how bot designers should describe the entities and intents that make up a conversation, and train the AI to handle understanding and flow.

Greg Vassallo

One of the formats of talks I found the most interesting at Interaction this year where the Spark Talks: designed to be short but full of great information. One that particularly caught my attention was Steve Bai’s talk called “Redesigning Urban Interacitons: City as a Service”, in which he highlighted the impact designing for cities can have in enhancing many people’s lives. He mentioned how some basic services like public bathrooms or public transport are not interconnected, but should. He even made a parallel between designing for cities and designing for an operating system. He also reminded us to be aware of who is controlling this particular kind of data and whether we can trust such companies.

The first day ended with Brendan Dawes’ closing keynote called “The Beautiful Inconvenience of Things”. He shared a number of examples he had worked on, in which he highlighted how little inconveniences are able to introduce pleasurable experiences. This similar to the anticipation that can be felt when listening to music on vinyl, which might be regarded as inconvenient, but somehow pleasurable because it gives you a break. One of his project that particularly caught my attention was one in which he used old, unused projection slides with images of various album’s artwork and added an NFC sticker so that when placed on a machine that could read the sticker, would play the corresponding song.

Day 2

The day started with a very inspiring keynoted called “A Journey in Social Entreprenaurship — Pitstops & Lessons” by Juliana Rotich, Executive director of — and incredible solution for creating a “connected classroom” in rural Africa. There is a TED talk by her on this you should definitely check out. The main takeaway I got out of this talk was the importance of designing locally: a product designed in LA or New York is not going to necessarily be the best solution tailored to the population in Nairobi. There are a lot of intricacies involved in designing with a specific population and culture in mind.

Next up was Jon Kolko and his talk “Sh*t Sh*ow: Finding Focus in The Midst of Ambiguity”. This talk was primarily about his recommendations on how to gain creative clarity in organizations that haven’t adapted well to let it thrive, but a lot of it was about design leadership as well. Jon recommended that as design leaders, we should focus on building trust among out peers, as much as we do on building craft — making an artifact with and about the team. He also highlighted the importance of having the freedom to explore in the work place: e.g. hanging things on walls even if “not allowed” (there is a recent Medium post by Co-Founder & CEO of Airbnb Brian Chesky that speaks to something very similar); how rules, unlike constraints, kill creativity; and driving a vision, which doesn’t mean “having the answer”, but more giving your team a reason to go to work by crafting the opportunity space and visualizing the strategy just enough to set the trajectory.

Cindy Chastain was next, with a truly relevant talk to my current work environment called “Design-led Cultural Innvovation”. This was one of my favourites this year.

Unlike many of these talks, she wasn’t presenting a success story — more like work in progress. This was mainly because of the great amount of work it takes to exert change (in terms of design) at a financial institution. I for one can empathize being a designer at Chase: it’s an incredible challenge, but very tough one indeed.

Many industries are struggling to become more digital and think of themselves as tech companies, but they should if they want to survive. Banks are no different.

Cindy called for design to offer a clear value proposition to the company, which leads to better customer experiences and elevates the brand.

Design, according to Cindy, is very important for creating emotional connections. Her first area of focus at MasterCard was strategy and connecting: building repeatable practices where teams can start to bring customer-centric views before they even decide what the product is.

Cindy also highlighted the importance of journey mapping across the organization, which is crucial to make people see the end-to-end journey or the root cause of issues in products and services.

Finally, she mentioned that this culture change was not all about process and design thinking. It is also about socialization. Sharing the wins so people can see the value. Designers need to make the people involved in doing the work feel included: engineers, product managers, legal…etc. They must feel like they in the same journey as us.

The talk ended with a clear remark that for cultural change to happen, the role of design needs to be lifted out of a service function to a platform for driving business value.

Another one of my favourite talks this year was by Tony Chu, principal designer at Noodle.AI. It was called “Designing in a World Where Machines are Learning”.

One of the things I found the most interesting in this talk was how Tony called machine learning “the new ingredient in interaction design”, and as such we need to know what to do with it when designing products and services.

Tony mentioned how as designers, we usually care about friction and constantly find ways to reduce it. He then mentioned data points that might be accessible to us, such as time, energy, attention, streets, location, images…etc; and how machine learning was at the intersection of friction and this type of data.

Tony Chu

Towards the end of the talk, Tony highlighted 3 specific patterns seen in applications using machine learning in design: repeated context-dependent decisions (destination selection in a map depending on the day and the time); inferred intent; and filtering signal from noise, as we can see in somr photo applications today.

Tony finished by introducing in interesting concept he called “Confusion Matrix”, composed of quadrants that he argued should be taken into account when applying machine learning in design. At this point, I stopped taking notes and decided to simply pay attention to what Tony was saying. Unfortunately I don’t have much recollection of said quadrants, so I urge you to see this talk in full once it becomes available online to get the full scope (I sure will).

The last talk from Day 2 I want to highlight is that by Jack Moffett called: “I’ll Gladly Pay You Tuesday for a Hamburger Today — Managing UX Debt”. In it, Jack talked about classifying UX debt into technical, functional and behavioural and visual documentation. He also mentioned ways of avoiding UX debt, namely via usability research & testing; paying attention to details; modularity (something I’m currently working on at Chase) and documentation.

Jack pointed out that there is a comprehensive video of this talk previously done at a similar (also really good) conference called Midwest UX.

Day 3

Among some of the talks I found the most interesting during the last day of the conference was that by zachary jean paradis entitled “Agents vs. Agency: Battle for Control in IoT”.

Zachary reminded us that design has always enabled human agency: the capacity or condition of enabling the ability to feel empowered to do something. Yet, as designers, we are presented with the choice of continuing building agency vs. building more agents: a entity or thing acting in behalf of someone else.

The “agent vs. agency” debate becomes more relevant in the context of the Internet of Things, and this is the way the Zachary wanted to frame his comparison of the two. He also pointed out that because of the Internet of Things and Artificial Intelligence, a lot of different agents were going to inevitably be created — which is why we need to be aware of what they mean for users.

Zachary ended his talk by pointing some of the most important aspects behind agency: access, fitness and continuity; as well as those of agents: anticipation, transparency and control.

Matt Yurdana, from the IoT Experiences Group at Intel followed, with his talk “I’m Safe… Really… Trust Me: Establishing Trust With Fully Autonomous Vehicles”. In this talk, we learned that in order to address the complexity of interactions within a vehicle enabled with voice commands, we need to be able to use natural dialogue, which the system needs to know how to instigate. He also pointed out that creating trust in interactions is about communication timing and responsiveness.

Matt Yurdana

After lunch during the third and final day, I saw one of my other absolute favourite talks of the entire conference. Chelsey Delaney from Planned Parenthood gave a talk entitled “Designing to Combat Misinformation”. I chose this talk mostly out of support for Planned Parenthood, given the current political climate in the US and how affected this organization is likely to be under the current political administration. I thought Chelsey’s talk was going to be somewhat politically charged, but instead it ended up being a talk full of practical applications for designers everywhere.

Chelsey addressed the relevance and impact of misinformation in a general context, as well as in the context of designing for user like those of Planned Parenthood’s birth control app Spot On. She also highlighted the difference between misinformation (neutral intent, condition of being misinformed), disinformation (the action of disseminating deliberately false information) and propaganda (biased information used to promote a political case, which can be true or false).

One of the main points of Chelsey’s talk was for designers to “GOOB” (commonly known as “Get Out Of The Building” — except with a twist: “Get Out Of The BUBBLE”. She also highlighted some of the things that make misinformation happen: exposure (the more you see something, the more familiar it gets) and oversimplification (something that looses its meaning because due to being too simple).

The talk then went on to discuss the case study of how the Planned Parenthood app Spot On dealt with users who were affected by misinformation and miseducation from similar apps. Other apps were oversimplifying birth control and the menstrual cycle, while also providing medically inaccurate information as to when it was safe for women to have sexual intercourse and not get pregnant. Chelsey demonstrated how Spot On solved for this by making the app medically accurate, educational, informative, neutral, inclusive, genuine and fun. They achieved this by committing fully to accuracy and clarity; learning the sources of misinformation; addressing the truths, not the myths; and engaging users in contextual eduction.

The talk followed with another similarly-addressed case study Planned Parenthood was involved with (Global Mobile) about bringing information about safe sexual and reproductive health via a mobile-responsive site to countries like Nigeria, where the origins of misinformation are difficult to track, as it begins offline an spreads rapidly. The talk ended with a tips for debunking myths, such as choosing to show instead of tell; the importance of visualization; getting out of one’s bubble and understanding the end user’s own bubble.

Chelsey Delaney

Debunking a myth leaves a gap. It is important to fill it back up with an alternative narrative (do not confuse with “alternative facts”).

Kevin Gaunt, from Samsung Research gave a talk called “Past and Future Speculations on Smarter Homes”. Kevin shared a few videos of concepts from the 60’s all the way though the 90’s of what smart homes could be. This footage showed the way in which many of the concepts of smart homes we have today have been influenced by these ideas. Kevin then challenged these ideas by calling out that smart home technology needed to continue providing a clear sense of agency the more and more it acts on out behalf. He also highlighted the importance of purposely designing limitations into how smart this kind of technology appears to be, a well as the awareness of the moral boundaries around designing technology that has a ig emotional role in our minds.

The conference talks came to an end with a great keynote by Marc Rettig and Hanna du Pleases called “Design as Participation”, but the last talk I want to highlight was one I saw right before, which turned out to be one of my other favourites of the whole conference: “Ethics in the AI Age” by Cennydd Bowles.

Cennydd called for deeply questioning the ethics behind not only artificial intelligence, but overall technology and the products we design through it. We have grown accustomed to automate many things through technology, but should a military robot be automated to make the decision to kill? Is our technology treating people as end or means? Are we maximizing happiness for the greater number of people?

After posing some intriguing questions about ethics in technology, Cennydd pointed out a number of useful things to account for a better sense of ethics in technology and design, like ever loosing attention to accessibility and being aware of how the culture of a company affects it’s ethical decisions. He also proposed the notion of building an “ethical infrastructure” by appointing a “designated dissenter”; engaging in constructive antagonism; challenging assumptions much like a QA engineer does with software.

Cennydd Bowles

The talk ended by highlighting how diversity serves as an ethical early warning system; the various ethical benefits of user research; how policies and documents do (sometimes) help. He reminded us that cultures have always been shaped by their technology, the making those directly involved with creating with and through technology have significant power in our culture.

These were all the talks and presentations I attended and found the most interesting. The conference closed with the usual soiree of Interaction Awards at a beautiful venue in Manhattan called Gotham Hall, where we saw inspiring designers be awarded by their incredible submissions to the different award categories; ate delicious food; dance and chatted the night away before engaging in one of Interaction’s best and not-so-secret “secret” tradition of Karaoke somewhere in New York’s Koreatown.

Interaction Design Awards at Gotham Hall

It’s already been a whole month since it ended, but I already can’t wait for next year’s conference, which will take place in Lyon.

I’m going. I sure hope to see many of you there. Again.

On y va?