As a Design Anthropologist, design thinking in conjunction with Socio-Cultural Anthropology for UX research is my focus. I design conversational user interfaces (ie. chatbots) that leverage AI and IoT. My approach is to extend Service Design and User Experience Research across all stages of my practice to consistently and reliably foster insights of high value for policy, customer experience, design, services, usability, and tech development.View the profile
Rory is a freelance technology consultant and founder of user research tools startup, inChat. He has worked with public sector organisations, start-ups, and universities, to help them take their early stage concepts to completion. Rory is passionate about sharing and advocating the advantages of digital technology, delivering training on development, civic technology, and Open Data.View the profile
About the talk
In this presentation, we want to open up the possibilities of designing chatbots for research purposes. Our journey in AI at inChat, began as what we see as the opposite end of practice and design for most chatbot developers. In doing so we have created a methodological approach that we call the Ethnobot (ethnographic chatbot). Bots designed in said method can be used for UX, Service Design, and Ethnographies. Equally important we have learned some best practices for embedding research infrastructures into services and products. These yield the benefits of having a UX researcher on hand. By sharing our story through key case studies, we offer insights towards designing chatbots as a viable resource for teams with varying levels of expertise in user experience and service design.
Presented by inChat Co-Founder & Design Anthropologist, Hector Fried and inChat Co-Founder & Technology Developer Rory Gianni at the 2021 Rasa Summit (https://rasa.com/summit/) .
Hector Michael Fried is a Design Anthropologist and Artist. He works between UX research, service design, and ethnography. Hector’s research explores the use of technological probes towards qualitative and ethnographic inquiry, while applying it to industry issues around UX and service design. He has specific interest in IoT with an emphasis on contextualising the growing relationship between things, humans, and AI. At Design Informatics he has worked with Edinburgh City Council towards their Public Art Research Project and conducted research under the Internet of Things Research Hub (PETRAS). He has become a leading expert in Conversational User Interaction Design for research purposes.
Rory Gianni is a creative technologist and trainer. He has over 10 years experience working with public sector organisations, start-ups, and universities helping them take their early stage concepts to production. In his previous role as a research associate at the department of Design Informatics, Rory focused on developing technological probes and tools to conduct design led research. He is now working on the edge of affective technologies to enhance conversational AI experiences.
- Learn more about Rasa: [https://rasa.com](https://www.youtube.com/redirect?even...)
- Rasa documentation: [http://rasa.com/docs](https://www.youtube.com/redirect?even...)
- Join the Rasa Community: [https://forum.rasa.com](https://www.youtube.com/redirect?even...)
- Twitter: [https://twitter.com/Rasa_HQ](https://www.youtube.com/redirect?even...)
- Facebook: [https://www.facebook.com/RasaHQ](https://www.youtube.com/redirect?even...)
- Linkedin: [https://www.linkedin.com/company/rasa](https://www.youtube.com/redirect?even...)
#conversationalai #aichatbots #nlp
Okay, so let's get in you might as well kick off to the next side, right? So in shock as we said we're design consultancy understanding of users customers and audiences. But we actually started as a research project project at the University of Edinburgh recent public spaces after 2 years of qualitative research, we were not any closer to understanding The Innovation space that could exist. We have been looking at the types of stories and values. Festival, audiences might exchange of points of sales like
a Christmas Market or other local cultural festivals. Like the Royal Highland show which is about cattle in farming. We needed a way to see the types of stores that have been. We have been studying to participant, observation, and qualitative research methods from a digital perspective. After a couple months and difficult discussions, we had a breakthrough moment. What if we used a chopped off? Could we take the rolls? I played as a qualitative researcher and supplement a chatbot intervention. The short answer. Yes. And it worked really well we deployed
chatbots to incentivize festival-goers at the Royal Highland show. This is our first experiment where we handed them an iPhone with a chatbot out, that talk to them and ask, what's going on and about their experiences as it happened. We honestly thought we would get short one, word answers, but people surprised us and they were very candid and treat the bought like anyone else that was interested in the conversation as an ethnographer and he's a researcher. I was actually a bit shocked that this all worked out so well we got results that were equally rich and insightful as my previous
research conducted in person. At the same side of the same Festival, just a year prior. It will be up town in research and practice is that chatbots are easily adopted by a wide range of user. Participants after all my grandma send me emojis, we can take qualitative research approaches as well, take them and apply them at scale. Remotely Most Fascinating lead and truly an unexpected results is how users are more likely to be honest with the boss and they are with the human So what do we call this endeavor, we call this a stenography. So,
what is ethnography means, people Grotto's to write literally an easy translation would be to write about people or nation is the main practice of cultural Anthropologist and one widely used by other social sciences and is actually the origins of user research and which many people still practice today and its original intent method. When utilizing chop box for research purposes, though, there is a huge ask to contextualize the relationship. Between what I call an integral triangle between
things humans. Need AI. This is actually what we like to think of as a key paradigm shift and understanding culture and human interactions in general. To this end we are conducting ethnography ux research service design and product design. But we do this by combining multiple approaches and not one novel methodology, mobile technology, fee, diary studies, technological cultural probes ecological momentary assessment, and the practice of field notes, So we believed in strategic Ai and smart design. Basically, we're trying to employees
not to reinvent Alexa from scratch, but to make early design decisions to the build systems are practical and cost-effective for the use case at hand. So what we're really doing here is reimagining classic ethnography to Foster. Empathic understanding is a, I and I do to become more prevalent in our everyday lives. In in a very practical way, we're taking what the famous Anthropologist Clifford geertz used as a methods. Thick description. And this was an integral practice too early, ethnographers in the
practice of actually taking field, research notes. I'm so from a design standpoint and a wider range of social science is concerned with the practice of photography. I hear you have some service design. We offer a very useful approach. One where we can ants remove, what can be an ethnographic or field researchers subject to bias by allowing a i to take that role. So, in our research and work, we have found particular advantages or benefits or insights that are over these common practices that were combining because
easily and widely adopted. We can almost guarantee that there is going to be no, friction of use and because the chapel can be deployed in an intended environment of the problems faced or service. We have an in-the-moment moment context Rich understanding of what is happening by asking right the right questions in the right ways at the right time. We open up the possibilities of the audience in users to express their emotions and problems really. As a field, giving them space to surprise us and ensuring key insights in the conversation. Don't go on Earth. The
most importantly, we are able to take a research approach and apply it at scale. And just as importantly where you able to take those qualitative insights at scale in real time and make key analysis. We're able to do this over a multitude of digital platforms. That's easy to deploy remotely and we still and I'll go I mention this again. I want to drive us home, find it extremely fascinating. This unexpected result, that people will be more honest than they are with the human, in a research practice we have seen it clearly over
and over again at East of the river effects. So we thought I would like to get into our first case study as parts of the iot UK Hub Network in partnership with the Royal Bank of Scotland under the smart transactions in public spaces, research Kraft release. The first prototype of the shop to do graffiti at the Royal Highland show. The initial driver of this was focused on studying visitor transactions in Festival markets so that we can understand their stories and value constellations of various stakeholders. And as such we had value buttons to list of the user
to share stories. More explicitly with the ball. Was a ticket to the event. It's closed for entry without tickets. So, nobody's going once inside the visiting public and enjoy the spectacle, livestock parades and sales, and farm machinery demonstrations. There's a wide range of food, drinks and agricultural. Crop products on sale from a range of vendors. There are also many opportunities to sample where is for free. Another research at the Highland show was conducted in previous up nography by both researchers and students from designer for Maddox and the University of
Edinburgh in the following year, we introduced a chop box and replaced. Every one of the study was too full to understand the implication of using a chop box to gather ethnographic data from participants for outdoors and on the move and second to gather data for the steps project to achieve these objectives to learn about the responsiveness of participants to chop all questions and directions. And how often do they take the opportunity to expand voluntarily on the responses? Secondly, the participants perceptions and experiences of the chop Bok,
how they felt about the experience and how comfortable they were using in The Choppa context and thirdly the success of capturing the information required by examining, what was and what was not actually recorded with The Choppa. Will basically we have to show that we can be successful in capturing a detailed Corpus of data valuable for the steps project and as well understands. Understand the volume understand the experiences in this context, implementation can be rich and informative has data. We have shown also that in the simple Act of
prompting for specific actions such as go to taking We Gather a Richard, you to set the example 85% of all the photos in this trial provided when were provided when participants were specifically asked for them while seemingly against the usually open at a nature of ethnographic inquiry because of the to direct response, preset responses were important to explore because of the Practical benefit, they offer, and speeding up the interaction for Disney park. Participants who in the trial. When negotiating, a very hectic, events are also used as a way to notch, participants to
report on particular activities, but the aim was to produce greater in depth and breadth of response with preset responses, providing us. An open-ended questions afterwards. This however, was also where we found our first big mistake in this approach. We limited and prescribe the user and did not provide an opportunity to break from that. You might notice in this picture and there's the bun. Selections at the bottom, you can see that there's a blank. Blank wasn't intended. It was just not thought of the Gap in the button now. Sure, if we had done it correctly, you should have had
one more that says other user to provide any information. They deem fit of the experience. They were having Okay, I'd like to move on to our next case study. This one is just a study focused on the city council's. Public art project aims to understand that influence, public art practice in the City of Edinburg, including funding legislation policy, and planning, and to determine what needs to be in place to ensure the local settings are beneficial to The Devout and the public are using a chatbot intervention. We collected. In
moments with public Bar for 12 piece. Stakeholders recovered, a representative sample of the division in Edinburg City Council art, gallery's director artist in academics as a method to remove participant surrounded. The City of Edinburg Does annihilated the temporarily of space and place in the complexity of audiences and sending a larger public space in real-time participants that chair made, sure what we're calling curated moments focused around transformative and contentious spaces which allow us to problematize development of
new works of art. This was part of a larger for Stagecoach, design, service solution, towards Edinburgh City. Council of our project, the end result was a setup code design guidelines that were placed to the comedian planning on Ann Arbor city council for the betterment and change of the public art commission process. There is an example of a user Journey from Ann Arbor city council, public art, research project at conversations about the public art in the city moving around three distinct places. The chop up would be asked, to be taken to a transformative work of arts, a contentious
work of arts and then somewhere to place, or replace the work of arts. Each of the three places tied into questions pertaining to five key insights towards understanding public art of the City of Edinburg. They were one quality taste inappropriateness to authority responsibility and ownership 3, expertise, or trust and risk. And five a reactive or confusing lack of process. In essence, we were getting users to be self-reflexive and examine, what they are talking about in context in real time.
also sometimes abuses cut distracted on the journey and it was important to know when and where that happens and the reasons wise because deviating from your journey, can also be a problem and if not, maybe a great insights so what were the real results of this intervention is a really complex Endeavor that involves a lot of stakeholders such as artists Gallery, directors, politicians, and Council employees, The Choppa actually provided a framework, that allows diverse stakeholders to create a moment in time and space which provided a solid foundation for
a thick descriptions of practice and attitudes, Clever to allow this opportunity. We couldn't really accomplish it by the designer of researchers alone. It took a series of interviews and workshops with all the key stakeholders to find the right questions and he incites to open the discussion in the first place. This is lettuce to believe that if you leverage co-design early with your stakeholders to ensure the right types of interventions are built, and as a result, You will get thick and Rich detail now. The data has its own Journey
chat. Entries in our example are saved with timestamps along with media, and geolocation data into an event. Log are thinking of this as reimagining, practice practicing the Arts or science of the old is entitled to think it is entitled to think to your data that you will and want captured and how best to analyze it in the early steps of your design. You make these interpretation readable. What we call the ethno reader app similar to a qualitative PDF? That should make sense to any research or designer or clients that you need to demonstrate this to.
So you said a lot of things. These are two case studies. Obviously we can go in a lot more detail. But here are some key takeaways that I want to go over. That might help you in your Endeavor if you like what we're talkin about here. So basics for Levering shop box for research, purposes, number 1, leverage co-design, you can do this in collaborative setup Theory stages. We have made Co design workshops as a weigh-in. You can bring the right people in at the right times in various ways to develop and the Funky questions and provocations what is it? You need to understand.
What do you want the user to share or divulge? I remove the word Interruption from this list because it's a different mindset. Although key interactions are equally important to remove. Mechanic for a moment and make sure that we are thinking in the research Manor bus interaction will then be defined by or based on the key questions and publication as intrumental steps. In your user Journey. Next remove technological Hang-Ups. This is important for our own bias and but equally important to manage our clients and partners if tech bias filters into the
conversation too early. You all the missed opportunities in the real progress. Next this I stress. This one's very largely and it's a very different chicken and an egg kind of problems, but you want to embed a research infrastructure, you must find appropriate ways to improve your products and service your chop box. But also the service that the chop off might be providing and better. Yet I suggest stop thinking about that box, a simple products and instead as an integral means of service in a service design chain, that should be able to enter 8 in the bug itself from other factors.
Along that service line. Boxer Versace Insight allow for unstructured or non-prescribed responses that other button that we mentioned earlier you need to be you need to give an opportunity or potential for the user to surprise you on Discover something that maybe you would have missed. Annexed make the most of anthropomorphizing ocean over and over again. It just seems that whether you like it or not people, accept job and they will. This idea that the buttons either really intelligence or some or actually really human-like and you can't do anything about it. So design with it instead
of against it. And finally, we suggest to make the analysis, transcribe a bowl. What I mean here is that you must design for your client and end-user appropriately but also for yourself, this is quite hard to think about that once but since the but will be an automated service, you must consider the end interpretation of the uptake of that early on. If you want to have a continued practice instead of a one-off also, you must consider cool and how they won't trip with what you have done. And if you are as I am suggesting cure embedding a research infrastructure, how to iterate that
Buy this talk
Buy this video
With ConferenceCast.tv, you get access to our library of the world's best conference talks.