Future-Proofing Participatory Research Through Human Insight and AI

Ana Couvinhas, Catarina Nunes at UXDX Community: AI Research and Designing for Developers. Video: https://youtu.be/3WowcFoUX5I

Readable transcript: edited from the recording's captions for readability (fillers and false starts removed, punctuation and section headings added). Wording is the speaker's own. Timestamps are positions in the video. Names marked [?] could not be verified against the audio.

Why participatory research needs AI now

[00:00:08] Ana: Hello. Participatory research: we've talked about it before, so why are we bringing it up again? Because our ways of working are changing. AI is transforming everything, including research. Let me take you back for a second. A few years ago, when I did participatory research in African rural communities, everything was manual. Then when I moved to tech, the research became digital. We were still collecting qualitative insights, but the participatory research process was much faster, with digital inputs and some automations. And now we are in the generative AI era, and the way we all do research is changing again.

[00:01:02] The truth is, if we want to keep doing participatory research, we have to do it with AI. Otherwise we simply won't be able to scale it, sustain it, or even convince stakeholders to invest in it. So today we are going to talk about how to make participatory research future-proof without losing its human dimension. Thank you for joining this session. My name is Ana Couvinhas, and I'm specialized in systems thinking and ethnography.

[00:01:44] Catarina: And my name is Catarina Nunes. Hi everyone. I specialize in psychology and social cognition. Let me just acknowledge the other author of this talk, Andrea Mita [?], who is a product experience manager. Currently Ana and I are strategic UX researchers at OutSystems, and OutSystems is an AI-powered low-code software development platform. We are here to talk about how participatory research can humanize software development in this ever-evolving digital age.

[00:02:11] As we just saw in Ana's introduction, participatory research isn't new. It stems from activism, anthropology and community-driven research. These are all fields where the goal wasn't just to study people, but to involve people in shaping solutions. So instead of a researcher coming in, gathering information and leaving, participatory research puts people at the center. People document their own experiences and raise their own concerns, becoming active participants in the research. And what is more interesting is that even though this type of research started in the social sciences, the exact same approach works in tech. Whether you're conducting a scientific study on human psychology, designing for farmers in Africa, or for software developers in Silicon Valley, the core idea is the same. It's the people using the product that should be shaping it.

From diaries and cameras to generative AI

[00:03:05] Ana: So let's go back to that journey that I mentioned earlier: manual process, digital AI-powered process, and now generative-AI-powered research. When I first started doing participatory research, everything was handwritten. People would document their daily experience with diaries, and we provided research participants with cameras to capture everyday life. They captured unconventional scenarios that outsiders, people like me, would never ever be able to witness, from witch doctor rituals to household routines, and traditional dances to farming spirits. The insights were quite rich. They were invaluable. It was amazing. But this manual process was really time-consuming.

[00:04:04] Later, when I moved into tech, I started to apply the same approach, except that the diaries were digital, and instead of photos our research participants captured screenshots. And we used AI-powered tools to accelerate the participatory research, for example to do automatic transcriptions, as you can see here on the screen. So comparing these two realities, in the tech industry participatory research was faster than the research project in rural Africa. But fundamentally, it was still about understanding real people and collaborating with them.

[00:04:49] If you are interested in these real-world examples, and in participatory research strategies both in rural communities and in tech, you can check out these two videos, called "Software Developers Are People Too" and "Using Participatory Research in the Tech Industry". But today we are focusing on what's next. Now we are at another turning point. GenAI is completely transforming the way we work and the way we do research. So how do we evolve participatory research without losing its core human element? Well, to answer that, first we need to understand why participatory research is so powerful. And to understand that, we need to talk about the psychological side of it, and Catarina is the psychologist here.

The psychology behind participatory research

[00:05:51] Catarina: Okay, psychology to the rescue. Let's first try to understand why this is a psychology matter. As we saw, participatory research isn't just about gathering information. It's about the whole experience, the way people feel when they contribute to shaping a product. In participatory research, knowledge isn't simply transmitted, it's co-created. Participants actively engage in discussions, problem solving and reflection, making the research process more insightful and impactful. In fact, participatory research shapes both the mindset and the actions of participants, driving real benefits for the research and even for the business.

[00:06:36] In participatory research, the knowledge isn't simply transmitted; it's co-created between the participants and the researchers. Ana, can you please move to the next slide? This ties to the idea of social constructivism, which tells us that knowledge is built through collaboration, dialogue and shared experiences. By having researchers and participants working together, we create deeper insights that wouldn't emerge in other research methods.

[00:07:08] Another particular aspect of participatory research projects is that taking part in them strengthens the notion of social identity. What does this mean? What is the notion of social identity? People see themselves through the groups they belong to, and they take on the group's identity as their own. So research participants who belong to a group of, say, developers feel more connected to a research initiative where the subject matter is, for instance, the developer experience. Essentially, these developers take greater responsibility for the research outcomes. This leads to more thoughtful contributions and a stronger commitment to, and alignment with, the research goals.

[00:07:57] Participants... another... I'm sorry. Basically, we saw that participants feel ownership over their work, collaborating with researchers and with their peers, and see their knowledge and efforts making a real impact. Participatory research enables all of this. Participants are in control, feeling autonomy; they feel connected, experiencing relatedness; and they contribute to the solution, experiencing feelings of competence. These three dimensions, autonomy, relatedness and competence, are the core drivers of motivation according to self-determination theory, a very well-known theory in psychology. And higher motivation levels mean that people perform better and are more creative and more persistent.

[00:08:49] Enough with the theoretical side. Let's look at an example of this in action. In our research practice, in the OutSystems context, we do plenty of user research. Our users actively collaborate and engage in the research initiatives. They feel ownership of the solution they help create, and feel even more that they belong to the OutSystems developer community. They also get research participation badges, which bring recognition among the members of the OutSystems developer community. So then they proactively participate in the community forums and submit their ideas for improvement on the community ideas web page, which is pretty cool. They feel that their opinion matters and makes a difference to the improvement of the tool that they actually use at work every day.

[00:09:39] So we saw the explanatory theories backing the power of participatory research. Now, what about practical outcomes? In fact, all those psychological and behavioral phenomena we were talking about do contribute to interesting benefits that make this research type so powerful. For participants, the benefits are clear. They experience higher motivation, a greater sense of empowerment, improved performance and overall well-being. They don't just contribute; they feel valued and invested in the process.

[00:10:12] But the impact extends beyond participants. It's also beneficial for the quality of the research insights. Ideas are tested and refined faster, ensuring that the final product truly meets our users' needs. This has a direct impact on product quality and innovation. And that is a good thing, because ultimately this means there may be gains for the business. From a business perspective, participatory research can enhance customer loyalty, and a more engaged customer base leads to better products and stronger business success.

Using AI without dehumanizing the process

[00:10:46] Ana: So participatory research is powerful. It's indeed awesome. But it's also time-consuming, and in fast-moving industries that's a big problem. Convincing stakeholders, like our managers, leadership teams and decision makers, to invest in participatory research isn't always easy. They care about efficiency and speed, and if something takes too long or seems too resource-heavy, it's at risk of being deprioritized. And let's be honest: traditionally, participatory research really does require time and effort. And in an industry where deadlines are very tight, it's easy for teams to say, "We don't have time for this."

[00:11:42] So what happens if we can't convince them? The risk is clear. If participatory research gets deprioritized, we lose the human connection to our users. We start making decisions in isolation. And when that happens, we end up with products and services that don't truly meet user needs. Now, with emerging technologies and tools, we can make participatory research more efficient. In fact, AI is not here to replace participatory research. It's the only way to make it sustainable in today's fast-paced environment.

[00:12:27] But, and this is critical, we have to be careful not to dehumanize the process. AI can support participatory research, but it cannot replace humans, that is, real participants. What does that mean in practice? It means that AI should assist with tasks, but it shouldn't generate artificial users or replace real participant insight. AI should help us optimize the research processes, but not at the cost of genuine human experiences. If we use AI blindly, we risk losing everything that makes participatory research so powerful.

[00:13:19] Here is an example of how we can use AI effectively while maintaining the human-centered focus. Let's start with planning studies. We can use AI to find trends and important research questions that matter to participants. But let's not let AI decide everything. People really need to shape the study, to keep it fair and to keep it relevant. Next, while we are conducting the research, we can use AI to help us with note-taking and automatic translation, so we can focus on real conversations. It is very important that we don't let GenAI replace real human participants.

[00:14:04] When we are analyzing data, let's use AI to quickly sort and find patterns in the data, making it easier to understand. But don't let AI decide what the data really means. Only people can understand context, empathy, nuance. And finally, when we are doing the reports of the research, let's use AI to create charts and summaries that help explain the findings clearly. But let's not use AI to take over the storytelling. Only people can bring this depth, this empathy and the true meaning of their experiences to the narrative.

[00:14:47] So if we use AI wisely, we can keep real users at the center and at the same time make participatory research more scalable, more efficient and, most importantly, more convincing to stakeholders. The key takeaway here is that AI isn't a threat to participatory research. It's the tool that will allow it to survive. Researchers should use AI as a collaborator, not as a replacement for humans.

[00:15:21] In a nutshell, just to wrap up: if instead of humans we perform research with AI-generated participants, the research process will be faster, but it's only a quick win. When we collaborate with humans, the real users of our products or services, these people feel motivated. They feel engaged. And when they feel that they contribute to the improvement of a product or a service that is for them, that affects their daily lives, what happens is that they feel ownership. And because they feel engaged during these research activities, these participants give detailed and contextualized feedback, which leads to richer research insights. They also talk positively about the product and service with other people, which is good for the business: the word of mouth.

[00:16:23] Machines feel nothing, of course. And we won't have all these psychological and social benefits that result from the collaboration between researchers and research participants, because in reality these AI-generated participants are just machines. They are machines. And participatory research methods emphasize even more the feeling of engagement and ownership, because they make participants co-creators rather than just subjects of study. And of course, with these co-creators, the research insights are richer, which leads to research recommendations for the business with a higher confidence level and less risk. And as we mentioned before, it also contributes to more customer loyalty, which is good for the business. So participatory research is a win-win situation for both participants and the business. Thank you very much.

Q&A

[00:17:35] Rory: Excellent, thank you very much. I think you're definitely selling participatory research there: win-win for everybody. Thank you very much for the case; it was really interesting going into all of the different aspects. Again, just to repeat, if anybody has any questions... I did miss a question from the last session that came in just at the end, so please do try to get your questions in early, before we run out of time. But I'll lead off with the first question.

[00:18:06] When you were saying don't use AI interviewees, there is also the point about AI interviewers. There are a couple of tools out there now whose pitch is basically: "We'll talk to real humans. You can only scale to how many people you can talk to, because you're one person, but we can interview a thousand people overnight, and we'll follow whatever script you give us." What are your thoughts on that?

[00:18:39] Ana: It's similar, because all these benefits that come from participatory research actually come from this relationship, this exchange, this communication between humans. If we take the human from one side of the relationship, all this feeling of belonging, relatedness, we lose it. We lose this word of mouth. We lose this human contact that makes people feel that they own it, that they are engaged.

[00:19:18] Of course, we are talking about a type of research, for longitudinal studies, in which we are studying the source of problems, or we are doing narrative research, which means it's not quantitative; it's more qualitative. We are trying to understand the problems. We are trying to understand how people think, how people feel. This is why we need participatory research. There are other types of research that need more quantitative and less qualitative methods, and there most likely AI can accelerate the process more. But with qualitative work, these kinds of things where we are really exploring and trying to understand the depth of humans and their lives, it's important to keep this human connection. Catarina, do you want to add something as a psychologist?

[00:20:20] Catarina: I totally agree with you. It's the social nature, the co-constructivism. If you interview, as you said, Rory, a thousand people a night, you'll still have plenty of data in the morning, and AI can help you process that data. But that won't work in cases of participatory research. Of course, this is not perhaps the best research method for all research problems; that's not what we're advocating. But that huge amount of data will turn into snippets, and there's going to be missing the interaction, the co-creation of knowledge that's based on social constructivism, as we discussed in our thesis here.

[00:21:07] Rory: When you say that kind of knowledge, do you mean it's literally the aha moments that you're receiving as the interviewer as people are talking? Because that really instills it. Whereas if you read a report, it's never as hard-hitting as if somebody says it to you and you can see and feel the emotion they have when they're saying it, for example.

[00:21:30] Ana: I agree, but it's not only the aha moment. It's not only what we collect while we are doing the research. It's because they are co-creating with us, because they participate really during the research process. There is again this exchange. It's a win-win situation. They feel, we feel, they feel engaged, word of mouth, better for the business. It's like a snowball.

[00:22:03] Rory: And just to tie in, because you mentioned cost, and that is a big factor for a lot of companies. I'm actually a big fan of having the full development team in a lot of research interviews, because when you're picking something up, you pick up different things depending on your background. A developer sitting in and listening to an interview would pick up on very different things to a designer or a researcher, just given their personal backgrounds. So what's your opinion on getting more of the internal team involved, versus just...

[00:22:41] Ana: In our opinion it's a very good practice. It's something that we also apply at OutSystems every time we can. Actually, it's often: we bring the product managers, we bring developers, we bring architects, we bring people with us to the interviews and research sessions, participatory research sessions as well. And please be aware that we are talking mainly about while we are performing the research. The moment in which we are gathering is when it's important to have the human with us. But then there are the other stages. We have the preparing stage, then we have the analysis stage and the reporting stage. Those are stages in which AI can definitely help us a lot to accelerate the process: doing transcriptions, helping us analyze, and doing reports. AI can really accelerate that. That's why we need to have AI: to make sure that we accelerate all the parts of the process except the part in which we talk with real humans. That's the key thing here. Let's keep the real humans and use AI for all the rest, of course, as far as it makes sense.

[00:24:02] Rory: And I know you're against them, but there are a couple of tools out there that have AI fake users. Is there any value in it? Can it maybe be a starting point, to get you started?

[00:24:11] Ana: Yes, maybe. Watch out: we are saying that during the research phase, in which we need to collect real information from users. But before the research we have a phase in which we plan the research. While we are planning the research, it can also be handy to have these non-human research participants to help us, for example, to test the script, or to anticipate potential feedback from the user that we were not anticipating, and then we improve the script. So non-human research participants can be useful while planning, for us to do dry runs [?] and so forth, to improve the script and improve the research plan. But not during the real information-gathering phase of the research. Does what I'm saying make sense? While we are conducting the research.

[00:25:13] Rory: Yep. It's a great point, about helping. I love the analogy of fixing the script, because often you'll be in a talk with a real person and you're like, "Oh, I wish I had asked this, or I should have done something else," and it's too late at that stage.

[00:25:26] Ana: Yeah, it actually helps us, like role-playing with a machine, without making other people lose their time. We role-play, we waste the time of the machine, and that's okay. And so we are better prepared for real users while we are conducting the research.

[00:25:47] Rory: And the last question is just around the analysis. How do you mitigate the risk that people get lazy? That's a lot of text to review in a transcript, trying to pull out insights, whereas ChatGPT will do that in 10 seconds. So how do you avoid losing nuance and context in your interviews when you're doing that analysis step and utilizing those AI tools?

[00:26:19] Ana: From my side, it really comes with experience. Where you can use AI, it's very, very valuable to help you find the initial patterns, to cluster the initial things, maybe even to raise some flags or raise some topics that we were not seeing, or interrelations and interconnections. It helps you in the beginning, but then you actually need to come in and validate everything, and really perceive and assess whether it's losing all these nuances, all this empathy, all this nonverbal communication that is missed with AI. Catarina, do you want to add something?

[00:27:10] Catarina: Yes. I was thinking that this actually connects to the question you were asking earlier about AI doing interviews. I prefer to go to all the interviews I can when I am going to analyze the data later on, because I'm already learning while interviewing the users. People here on the team say that I'm always in every user interview, because I do feel it really helps you overcome those challenges. AI and an analyst alone would not have the context that you have. So I think that going to talk to the users is the easiest way to prepare for that data analysis moment.

[00:27:58] Rory: Brilliant. Well, thank you both. I think we're just up at time now, so thank you both for that great talk. It's really interesting. I love hearing how people are using AI, and how they're not using AI, to accelerate, because as you said, it's a fact of life. It's out there. It's something that people are going to use, so let's use it in the best way possible.

[00:28:18] Ana: Yeah, exactly.

[00:28:20] Rory: All right, well, thank you very much.

[00:28:21] Ana: Thank you.