I'm Matt. I am the CEO here at Lyssna. Little bit about me. I'm as I said, I'm based in Melbourne, Australia but Serbian origin I'm a former design practitioner. Although I try to stay out of the way of the design team as much as possible these days to not annoy them.

And started Lyssna back in 2008 actually as as a single methodology called Five Second Test. And it's grown over a significant period of time to what it is today. It's a platform for both conducting research and recruiting participants. I need a glass of water, obviously.

Might hand over to to Nikki.

Yeah, thanks so much Matt. I also want to know how to pronounce your last name that came up in, in the chat.

Oh yeah, it's Milosavljevic, but it's funny and when I was growing up at school assemblies and stuff, whenever, names got called out, everyone would get their last name read out except for me. So it was it was almost like being called Cher or something.

Love that. That's awesome. What a great little story. I'm basically Cher. Yeah I'm Nikki, an American in the UK. And I have been in the user research field for about 12, 13 years now. I just don't count anymore because it just tells me how old I am.

And I specialize in qualitative user research and I do things like coaching and mentoring as well as a lot of consulting for companies. Both doing actual user research as well as consulting for services like in the AI space, which I'm sure we will call. We will call. Nope, we will talk about. So yeah. Super happy to super happy to be.

So my name is Julian, and similar to Nikki, I've been in the UX research space for, yeah, almost a decade. I started actually as a product manager, so I always count, like the whole tech experience. So that's pretty much yeah, 12 years more or less. But I just more than a decade, so I don't feel old or anything like that.

I also specialize in research, but in mostly in research operations for small teams. So throughout my career I was always either the first researcher or one of the first hires. So my focus has been on building practices from scratch, so setting up the infrastructure, helping with embedding research into the culture, into the processes, and the communications and working on all that.

And for a change, last year I joined a company called DuckDuckGo after years in consulting only for companies, and especially in Europe and the Americas. And now I'm, I specialize also in, in qual research at DuckDuckGo. So that's pretty new. And we, we do many things, Nikki. We have podcasts, we have stuff, so we can talk about that later. Yeah.

Fantastic. Thanks so much. Before we dive into the actual conversation, just a, a quick bit of housekeeping. For context we asked Julian and Nikki to share their biggest challenges coming into this year ahead of time. So the conversation today is going to focus on a few key issues that they're actually navigating right now and things that they're thinking about.

These are going to be honest perspectives. They're not necessarily reflective of any one particular company. We're going to be talking about what's working and what isn't working, and these things are fluid. We're working, we're, figuring this stuff out. So it's not set in stone. And as research practitioners yourselves I definitely encourage and invite everyone to share comments, ask questions in the chat.

We'd love to get your perspectives as well. We'll have time at the end to go through some of the questions. We'll try to go through as many as we can. Obviously we'll dedicate about 10 minutes or so towards the end to do that. And if for any reason you do need to drop off at any point, that's totally okay.

We are recording this and everyone that's here today will get a link at the end for the replay. Hopefully that takes care of that. And so with that little bit of housekeeping, I think it's, we may as well just get stuck into the thing that's probably at, the front of everyone's minds, which is AI and AI integrations.

How do we work with it? How do we integrate it into our workflows? And both of you flagged this as a top priority. Nikki, you specifically mentioned being really motivated to do this intentionally and efficiently and as and ethically. And I'd love to hear a little bit about what that means for you in practice.

Yeah, of course. So for me AI, I remember when I first heard about ChatGPT I was, I was actually in the UK visiting friends and somebody turned to me and was like, oh, have you heard of this? And I was like, how stupid. This is never going to be anything. Sure. So don't come to me for any advice on the future.

But I think that like a lot of other people, I was scared of AI and still am a bit scared of AI. And I was mainly scared of it because of the sheer possibility of it. And what it actually meant for all of us. And I've played around with it quite a lot both in, on the dark side of AI.

So getting AI to do things for me that it's not meant to do in order to experiment to understand where my own boundaries are. So when I talk about using AI in an ethical way, I can give some examples from like the, my own personal like space. So for instance, fiction writing is something that I want to become a famous fiction writer.

I want to become an bestselling New York Times bestselling horror author. That is my dream. You can't use AI to create fiction. It's absolute garbage and it's unethical, right? I tried, trust me, I tried. And it is, it's just not great. And it also, what it does is sucks all the creativity, all of the fun and all of the like, exciting parts of the journey and process from things.

And I actually feel that way as well with the research process. I think that AI can be great in streamlining some of the like administrative tasks, but asking it to give me like the answer to my research it, it's not, it for me, it's not really provided anything that I feel like I can use without such heavy editing that I might as well do a lot of this stuff myself.

So when it comes to those like ethical boundaries, it's am I using AI to try and replace myself because that doesn't feel like ethical? And am I or am I using it to try and speed up my work in really like intentional, thoughtful ways? And I'm always asking myself that question and finding out where that boundary lies in like a multitude of things.

Like I've used it to write LinkedIn posts before and I will say that some of my LinkedIn posts that I wrote are disgustingly well performing to the point where it's phenomenally annoying. But some of them are absolutely terrible and a lot of the phenomenal ones are based off of something that I started and then massaged and made right.

So like I think it's such a gamble right now that using it ethically is really thinking about how we use it. Sorry, that was like really long. I hope that made sense.

No, that was great. And I particularly like that you liken so the, the practice of research to a creative process.

Yeah and I mean that, that resonates for me. And there are some, there's a lot of conversations going around about, how AI does impact to the creative process, and do we want to even use it for that purpose. Yeah, I think that's a really useful jumping off point.

Julian I'd love to get your take on this as well. I know that you mentioned one of the things that you're interested in is how to embed AI into research workflows at DuckDuckGo, and then also just more broadly. I'm curious what this looks like for you day to day.

All right, so I have I'll give you a layered answer. So just to build on what Nikki just ended with I'm using AI for writing. I'm using AI to help me write my book, but I really run into limitations right away. So I ended up using it in a very limited way. So at the beginning I have a podcast.

I have 40 something solo episodes. So it's me just like riffing 15 minutes about a topic that's content. And I wanted to just put that into the book. And I also I'm more of a talker than a writer, so I could walk, I walk my dog, or I go for a walk and I talk, I record myself. And I wanted to use those transcripts for the book.

And at the beginning I was, playing around with it. And in the end I gave this bit and the AI, filled in the blanks. A lot of stuff that it's whoa. But this is not what I was saying here. So I, in the end I had to learn how to limit it in many ways. So as Nikki said it doesn't add its own thing, its own, and I don't want to call it hallucinations, but its own set of ideas, so to say.

It's if I'm talking about how you should frame a research project, I want you to use my take. I want you to use the transcript I'm giving you, but not fill in the blanks and maybe quote authors and put this, it's I didn't even say anything of that. So I am facing my many limitations.

In the end, I managed to tame it, if you want to put it that way. And I'm using it like, just like use only the stuff that I'm giving you, just clean up the repeated words, the fluff don't rewrite the question. So I'm using it in a very specific way. This also translates in how I am using it for research.

I'm, I don't like or I experience that le letting AI lose and play around with your stuff and supervised can give you a lot of stuff that you are not even sure that is right, that is based on facts and so on. So in the end, you need to be very mindful of where you use it. So I've been conducting some experiments myself in terms of how to use AI generated users or using AI for synthesis or using it for moderation.

I've been trying with different use cases and some of them have really good points pro, but it also sounds like stuff again. So there are tools, there are new tools and I think they have the place in the workflow. Also depends on the context, right? So you asked me about my context at DuckDuckGo. So we have our own flavor of AI.

It's called duck.ai, which is a fully private AI. So everything is everything you put there is not used to trained models and so on, and has different models inside. You can use GPT, claude, Lama Misra, you have different options there. And that's our go-to tool for AI. So basically everything that you put there is anonymized, it's not used for training and so on.

And it's a controlled environment. And we use that especially for crafting documents or polishing how we, when you have like a big data set and then you want to party some stuff. So we use it that way particularly and it's helping us accelerate, right? So once you know what you want to say, once, like uh.

Once you did the analysis yourself, and you have the materials for, to craft a report, the AI can really help you craft a report really quickly. So that's one of the use cases that for us, has been pretty useful so far. But but yeah, I think that also in different contexts, also, depending on how you treat, like we can talk about ethics and privacy.

There are many things that you, we can double click on. I think that also depends on the particular context. But so far, AI is helping us, but we are trying to figure out what's the best way or the best place to place it and how to best use it without, any ending up with a different information than what you actually need or what you actually put into that.

Yeah, it's interesting. You mentioned you're speaking to it and sharing your thoughts and maybe some of the transcripts of these these kind of walk and talk dog walks and it exerts its own gravity in a way. It tries to be helpful and it suggests various things, maybe fills in some blanks.

And I think, in your instance, you have a clear idea of what you want to achieve and how you want to, articulate it. But I think if you're not feeling super sure, like these kind of helpful suggestions can pull you in a different direction. Not necessarily good or bad, it exerts this influence.

And I think that's very interesting and I think that kind of dovetails into the ethics piece that you were mentioning Nikki as well, because it's hard to even be sure about whether you are being ethical when the lines start to become blurry around what's your take and what's the take of the algorithm?

Yeah. I really quickly I just want to say on that I feel very fortunate that I am an, I will call myself this, which I will rarely do, but am like an expert in what I do and that I became an expert in what I did before this came along because I know what good looks like and I have a lot of fully formed thoughts on the craft of user research.

And I have a lot of data of like real life stuff in which has formed those opinions and thoughts and those frameworks like for me I don't need AI to write a screener. I don't need AI to write a research plan. Like I can do those literally in my sleep. It's actually I'm at the point where it's probably more efficient for me to just do it anyway than to wrangle AI to do it for me.

Because I can do it, but when it comes to things that I have less authority on or less knowledge of, that's when I feel worried about AI and like how we're using it. Because yeah, it can pull you into these different things and like you're like, it gaslights you so much, like the number of times that AI has changed its opinion on what I'm saying, because I'm like, it's like this is the best idea.

A great idea, Nikki. And I'm like isn't it dumb? And then it's yeah, you're right. You're so dumb. Like, how did you ever think about it? So like it has no actual like. It is just it doesn't actually give you anything, and so it's about you utilizing your brain and AI together to like maybe make sense of something and show different perspectives.

But I just feel like it's almost like I feel very lucky that I didn't, I grew up without a iPhone, right? It's that same vibe, I grew up without social media. And it's, I think that we need to remind ourselves that there are multiple ways of learning and that AI isn't a learning modality.

Yeah. I think that's really well put. You mentioned, you know what good looks like and I'm curious for both of you are there areas where you feel that AI does generate or does help? And you can. There are, like, are there any domains in which you feel like you can utilize it?

And it does look good, it is actually genuinely helpful. Are there any examples Julian you mentioned, you've been running lots of experiments. Has anything really stood out to you as a particularly interesting use case?

Yeah, many. I think there, what Nikki said is exactly like that.

You need to be able to tell good from bad because it's like, if you never tried, this goes with food. If you never tried food, I don't know, Italian food, and then you go to an Italian restaurant in different parts of the world, they would trick you and they would say this is a typical dish.

And you say, oh, this is great, but then you go there in Italians, never heard of that. That's the common thing. And the same comes here. So you don't know what, if you don't know what good research or what good output looks like you're easily tricked. It's because AI will always give you an answer.

So if you try to ask it something, it will always say something back. So that's something to be careful of. I think there, it also depends on how you prompt and I think there's a lot there, there's a lot of magic in the prompt. So two people trying to achieve the same results. If they prompt differently, they might get completely different answers.

Especially using the same stuff. So there's a lot of, you need to be very precise with the prompting. But I'm going to go one, one take this one notch further. And I think that it also depends, because here we're talking about out of the box AI, right? It's if you go to chatgpt.com, Claude or whatever, it's like you just go and you use chat bot experience.

If you go one step further, and if I know there are any prompt engineers in the call or people trying to get into that. There's if you lift up the hood and you do a little bit of fine tuning and you have data sets where you have you, you have the sense making of the LLM of the language model, but then you're using it with a very specific data set and you control that and you train it on your data with no access to the internet because if you add access to the internet, that's, that becomes a bit anthropic, right?

It's it's there's a lot of the anthropic non-A by the way, andro, just to make sure I'm not an native speaker. So it could sound Yeah. Here. It becomes, it starts adding stuff from anywhere, but if you have it in your own little sandbox with all, with your particular data, and that's how I'm using it right now, especially what all the book thing that I was saying, I'm just very, in a very controlled sandbox.

That's when you can get. The best results. For example, some of the experiments I run was with training with question banks or with documents. So it's this is our research plan looks like category 1, 2, 3, 4, 5. If they learn from that from your controls dataset, then whenever you ask it, it will spit back what it used to learn, right?

So it won't give you like something completely out of the blue. What makes stuff up, right? So if you go to that level of precision and that level of engineer, I think AI has a lot of of potential there. But it's not easy. And I understand that people say what is this guy even talking about?

How do I even get to fine tune? It's I have no time to learn that I don't have, I just want to use that. That's when it gets tricky. One of the things that I really liked about AI is that if you train it with the right thing, it can be. Especially for those environments where you have other roles conducting research it can be really helpful.

So when crafting documents with what I just said now for if you load question banks or the structure of artifacts like research plans or templates and so on, whenever someone else needs to craft a type of study, they will get good like vetted data. So they will say, Hey, here's a research plan.

And that looks like a research plan because it was vetted by a researcher with right questions. It was like, these are good questions to ask. No double barrel questions, no leading questions, no would you buy this questions and all that. That's one thing. And then another thing, and this is with AI generated users also called synthetic users.

I'm not a fan of using that like for real research, so I'm going to make that clear. I don't think that's well apart from ethical, we can, this is a rabbit hole that we can go is there are many reasons why I don't think that's good. But you can use them as a, as a gym or like a training ground for people who do research who maybe don't know how to conduct an interview and they don't know which questions to ask.

You can say, yeah, you can access, I dunno, you can access any time of the day at a really low cost, your research gym and have a sparring partner and it say, oh, this is how it looks like. Great. And then I will go and test with real users or interview real users. So for that kind of training, I think there's a lot of potential in AI.

I don't see that much potential in the actual research then, but I do see in this education and training and getting, accelerating that bit.

Yeah, that makes a lot of sense. And for what it's worth I have similar sentiments in relation to synthetic users op, open-minded, but still a little bit cautious and skeptical at this stage. But I think that use case that you described sounds like a really interesting one.

And yeah, definitely sounds useful. Perhaps I can shift us a little bit towards something that both of you have touched on here. And it pertains to something that's been talked about a lot in UX research for quite a few years now, democratization of research and kind of bringing more folks into the fold and getting them to be maybe not fully fledged practitioners, but certainly engaging in the practice of research. Nikki, you said something interesting, which is that, you feel fortunate that you grew up without social media and perhaps without, having, everyone has a mobile phone. And the same goes for doing research prior to AI.

And then Julian, you mentioned people are feeling more capable because they have tooling and they, you can give them templates and AI can make suggestions. There's an interesting interaction here that's emerging, right? So it's, so people are coming into the fold, engaging in research practices more and more.

We want to, in invite them to do that and encourage them. But we've also just said that, AI can be helpful, but it can also be overly helpful at times and suggest things. And unless, unless you can tell what's, good or bad I think it is possible to lead people astray.

Nikki, I'd love to get your perspective on how do you see that playing out? Where do you see that, working well? Where do you see potential issues?

Yeah. So the democratization issue has been one for, since the beginning of time. It feels, and it's this, so it's the same thing over and over again.

AI is just a different mechanism of searching these things. Before AI, my colleagues would just go on Google and find a podcast by somebody who said, this is how you ask people how much they're going to pay for something. And then they would've just gone and asked that question. It's not I don't I the market difference to me isn't really there.

AI just maybe makes people think. A little bit more like they know things because it's like AI and AI is being a lot kinder to you than it needs to be. Or it should be, but like from my standpoint I consult and I set up like a lot of these research practices at organizations. So I go into organizations and help them like set up from zero to one because that's like pretty much what I did for most of my career as like a solo similar to Julian as like a solo and first researcher.

And so I have to, I usually have to democratize. But what I try and do is I've worked at organizations where we try to democratize, which is like, everybody should do this, everybody can do this. Everybody gets trained. Everybody's sitting in a three hour workshop with me wanting to like not be there and literally do anything else. Versus now what I do is much more like select champion based. I'm actually interested in this democratization. So the way that I try and enable research across organizations is by people who actually like, are interested in learning the things. And I think that AI, similar to what Julian was saying, AI can help if you give people like single purpose. So like I use ChatGPT more than any of the other models. So if you give somebody an agent with a single purpose and without access to the internet, that's how I have used AI. And it's trained on knowledge documents that I've created with custom instructions that I have created.

That's how I've started like using it to amplify democratization and training across an organization. And I think for that it's really great, but just like with everything with AI, there needs to still be that human like kind of touch to it or that human oversight to it. For instance, I was working at an organization recently where I did a training on how to ask open-ended questions, and I gave I, I used the Ted w methodology for asking open-ended questions.

That's something that I have always used and I trained an agent on that. And so people were only able to like, it only spit out open-ended, unbiased, neutral, non-leading questions. So it would turn their questions into that and like that single purpose GPT agent allowed them to play around with things.

And I think that's what AI can help with when it comes to democratization and education is stakeholders can play around with these in low stakes environments. We just need to make sure that those environments are super, like guard railed by creating something like, as I said, like single purpose agents that are highly trained on custom instructions, really great knowledge documents, and don't have access to the web where it's like giving a child a phone without access to the internet.

Totally. I yeah, so that makes a lot of sense. And I, and obviously like tooling often when you have AI embedded into tooling, it provides some of those guardrails because it's not, just like open season chatbots. It's, constrained to the particular context that you're working in.

But I'm also curious in addition to just the guardrails around like the capabilities of the model, I'm curious about guardrails or perhaps just guidance around like what context is it appropriate to leverage AI and what context is it not? And I'm anticipating this is, falls on a spectrum somewhere.

But, let's say that I am not a a seasoned researcher and I'm interested in conducting some research, from your perspectives. Where do you feel comfortable letting me just run with things, perhaps with the use of with the benefit of some AI assistance. And where do you feel like, hey Matt, I'd really to just make sure you're on the right track.

What, how do you see those different use cases?

I would not let you run with anything.

Yeah. Yeah. Yeah that's obviously the right answer.

No I, it really depends on like your so let's assume that you're not, you've not done research before, really.

Like you have an idea of what research is and what I care more about is your willingness to take feedback and your willingness to change and listen to somebody who is. Who knows what good looks like. And so what I would be okay with you using AI for is as a thought partner, brainstorming like, what do I what is this research answering?

What do I really want to achieve with this research? Who are maybe some people who I might want to talk to, like the planning and scoping of it, but I would still want to see that because I'm not wasting anybody's time, yours or the participants or the organizations. So I think like those and for me, if we do that due diligence early on in a project, that's when research can have a really good outcome.

If it sucks from the beginning garbage in, garbage out. Yeah. So I would gauge your like willingness to take the feedback and listen to somebody who knows what good looks like and using it as more of a thought partner or within some sort of parameters that we sp set up.

Yeah, that makes a lot of sense. Julian, I'm curious whether, you've had any sort of experience in this and whether it's good or cautionary.

I'm going to change the angle completely here. Sure. If I think AI is just the latest representation of the latest phase of an underlying problem of people trying to bypass research or skip it and whatever, and let's ship fast and do, and break things on whatever that mantra came from.

That's, this is the latest, it's like a video games, like you just one, you kill one, and then comes the next one. It's it's a different, it is a different thing. Sorry. I love video games, by the way, so I might use a lot of analogies here, but totally fine. That's pretty much how I see it, right?

It's like the people who are willing to go and replace research with answers from ChatGPT or Claude or whatever is, are the same people who be forced, is we don't have time for research. Why do we need research? I know what my customers want and all that. It is just another, it's a 2026 addition of let's not spend time in research.

Why? Because AI will always give you answers. It can also confirm your con, your preconceptions, the sort of confirmation bias. Great idea, Bob. Great. Fantastic. Go ahead. You're the best. And then they say no, I'm not the best. No, you're not the best, Bob, you're the worst. It's that's how AI sometimes work works.

So I think that this is the same danger. You get re like answers. W very, I mean with very low confidence, how to say, but you get answers right away. It feels like you're making progress. It feels like you're moving fast. And if you move fast, it feels like you're achieving things. And this is the same thing.

So in the end, it's more of a culture problem of how to make product. And I'm not saying particularly about product manager, it's also like we, I mean I don't, sometimes I don't even think of research design. It's like we are product people. In the end, what we want is that we want to build good products that, are useful for people or that people see value in.

And you can also build a business around. So that's our goal, right? So we tackle it from the research angle, but we are part of a larger team and we all need to go that way, right? So when I think of this is like if I had to, I mean you can do a lot of stuff right now and just to avoid this it's research always to blame.

Like now you can use like Cursor, Lovable and you can like design or you can vibe like vibe code. But as a researcher, do I really feel confident that I would, I know ship in this code that I maybe did with AI and like vibe code is I dunno, that will break and that will waste an engineering, like engineering time fixing all this patchwork code that I put together.

It is it good for getting a first approach? Like trying out the water. It's oh, this is good. This might work. You can probably like, iterate on that. That can work. But I don't know how far this will go without the expert in the area. I can play around with designs and I can test and iterate and you can use Lyssna and test and, and or any tool, you can use it and you can get good results.

But at some point it say I'm not a designer so I just, I can't really, take that far, without them. And this goes is connected to the research. At some point it's you don't skip design, you don't skip engineer, you don't skip research at some point's like how good the product culture is and how do you understand that You need to understand what you're building, who you're building it for and how to build it.

And then this is just another, manifestation of the same issue. It's like Henry Ford quotes and Steve Job references, faster horses and whatever, all those excuses that we always hear just set by a an LLM now.

I, this kind of maybe again, shifts us a little bit, but you, you mentioned it's a cultural issue.

I think it's, I'm a little bit heartened and I appreciate you're both saying that there's this is symptomatic of we're wanting to take a shortcut here and and find a way to do the research quickly, but not wanting to actually invest.

On one hand I, I agree. I think this is, it can be a little bit dangerous and can lead people to think that they're actually doing research when it's questionable. Maybe sometimes they aren't. And it's, I think it's not, I think feeling like you've done something but you haven't is probably worse than knowing that you haven't done it.

That said, I think at least it points to some sort of there is a, there is some motivation to, to actually do research. And Julian you mentioned that this is the cultural issue. I agree. In terms of getting more buy-in, and this is probably another thing that kind of gets talked about in the UX research space a lot.

I wish it would, I wish we would finally solve this, but given that we're still here, when it comes to actually proving the value of research, proving that it's worth doing and demonstrating the impact that it has, how do you think about that? Where do you think that the challenges still exist?

What does it look like for you to get people to have buy-in to see that there's value in doing it? How do you get them to consume outputs of research and actually take, make decisions and take actions on the basis of research, whether it's AI assist or not?

Yeah, so from my side, this is something that I care quite a lot about.

But, the whole thing that I have learned in the past 12 or 13 years is that, like when you're, I used to just present research. I used to create a deck painstakingly because I suck at design. And that deck had a whole lot of stuff in it. It was probably like 97 slides. I could have, presented a thesis on the topics that I researched.

But nobody cares. That's the thing is nobody cares about the methodology. Nobody really cares about, they care, but they care about the worst things, which is oh, you only spoke to in certain number here of a low sample size, and so what I realized is that my research is only impactful if I talk to what people actually care about.

You're go only going to impact somebody if you know what that person cares about. So the people that I am presenting research to now, I talk to them about what they care about, which is normally revenue or some sort of like company based or team-based metric, which is wholly unsatisfying sometimes because I joined user research to help aid in creating products that move the world forward and make people happy.

That was like literally why I joined research. I was like, cool, like this is very interesting stuff and I want to change the world in a positive way. And now I'm talking about conversion rate all the time because that is one of the most important metrics in the world. And so I think that. What, for me at least is that I try and do when it comes to getting buy-in, is trying to balance those two things now.

So like talking about how changing something could not only help the business move forward, but also help positively impact the people who are using our product. And that is a very difficult balance to find when you're talking to a C-suite who have five minutes who only care about who I'm assuming are mostly caring about revenue.

Sorry Matt, I know that you are.

No, i'm not feeling, I'm not feeling seen by this. So we're all good.

I think that it's about unfor. So I think unfortunately when you are an information giver, which is usually what like researchers are, we give information people to people so that they are making decisions. We are usually not the decision makers. When you are an information giver, you need to give the information that people care about and need in order to make decisions.

If you don't know what that is, you are going to go into a presentation saying, what am I supposed to present? Or, I tried to put together this presentation and I had no idea what to put in it. I don't know what to say. I don't know what to do. That was the symptom of me not knowing the decisions that people needed to make and the information that they needed.

So you are there to provide that information so that they can make that decision. And if you don't know that you are going to have a really hard time getting buy-in and building a research culture that is like where research is positively viewed. And if you are there, just trying to say oh, research is so important.

You're not going to get buy-in. I tried that. I tried to tell people it's really important. And you know what? People ask me, why? People five why'd me? And that's really annoying.

The, they pull the reverse uno card.

Yeah. I was like, this is the one time that you asked why you did all those research interviews, and you never ask why.

That's great.

So you have to be able to five why yourself of why the research is important. Your organization. If you can't do that, you're not there yet and you are not going to convince anybody else. Yeah. The truth.

Yeah. Julian, what about yourself?

Yeah, I agree with everything that Nikki just said. Just to add on top of other couple of things. One is that decisions are never fully rational. So there are different components to a decision. So these are the facts. Now you're convinced, please act on the facts.

It doesn't work like that. So you need to cater to all the emotional side and try to understand the psychology behind what other roles might be experiencing. That's part of how you should present stuff. Just like focus on the facts, but also focus on the emotions and try to get what could deactivate those emotions.

Because in the end it's all about safety and, making the right bets and making the right decisions. So that's on one side. And something that I always like using is the risking kind of speech, which is if we did, if we like made the wrong decision, how much would that cost us? So it's if we had to, if we selected this feature and this other feature, and then we went and developed those features and then it turns out we didn't do it or we didn't design them the right way, or how much would it cost us in terms of redeveloping or fixing plus customers loss? Because if you experience but features, I'm not going back to this thing anymore, it's like maybe it's churn.

So there are like many things that can come from rushing. So the investment is if we invest two weeks doing some research on this and we can de-risk your decision a little bit, isn't that worth, not spending the rest, like not if redeveloping costs you can put a number to it.

If you had the number that would be even great 100 K. It's if we do this wrong, it costs 100 K. If we take two weeks more, the cost would be this, but the risk changes. So that's one of the usual arguments that I use. I think it's, this also caters to what I said before about the emotional part.

Then another thing, and this is also an extra layer to this and I'm also aware that, I don't know, we haven't asked, we ask about where people are based or nationalities and so on, but we haven't asked. You, you are you are, for example you are also a designer you said like you used to.

So I know that a lot of people who are conducting research are not necessarily researchers. So that's also part of what I've seen in my work. That's also, I'm, and this is not a shameless blog. I swear the book I'm writing is not aimed at researchers. It's just like how to build a research practice. If you're a researcher, you can also use the book, but if you are a product manager, if you're a designer, if you're a founder, how can you build that practice?

Because you need to build the muscle. And sometimes the researcher comes later, research happens and you have other folks doing it, and then. The conversation changes because it's like, how do you justify if you're a designer, how do you justify spending two weeks doing research? If you can then, and then design, it's like how do you justify the investment and the change in your workload?

And this is also part of what we need to think, it's not just researcher justifying why should be part of an organization and why your work is valuable, which of course is if you're a researcher, but also think that way. It's if you are spending that time, how do you justify spending that time, that effort and that money in learning more about the users, the problem space and so on.

And that's usually, it's what Nikki said is what I just said before. De-risking, seeing, trying to put what's important there and try to bring up to the surface all those things that probably can derail how you are building your product altogether.

This reminds me of something that you flagged as a challenge, Nikki.

We're talking about how do you get people to buy into the experimentation, but you mentioned something that I thought was really interesting, which is how do you experiment and fail without kind of getting into trouble? Because it's not just, yes. We can run with the assumption that all the experiments will yield something interesting, but it's also totally reasonable for an experiment and not go according to plan.

And you want to be able to be able to do that and then be able to do that again and not have stakeholders say Hey we gave you that one chance and you messed it up, and it's now it's all over. How do you navigate that kind of thing?

It's so frustrating because I remember, so we have this like experiment and fail and that's how you learn, right?

As like a, as a product thing. And it's so annoying because when I first started doing user research, I'd be screwed if I couldn't fail these days. But I feel like that experimentation and failure isn't there like it used to be. Like I used to run like more like creative research sessions that I don't have the ability to run now because if it fails or if it doesn't get like the exact results or if it takes a little longer or I get in trouble.

And I'm not delivering fast enough on my, like ROI as a researcher. Because everybody wants research done yesterday. So what I recommend doing is because again, this goes back to I feel like we've lost a lot of creativity and fun that used to be in user research like paper prototypes.

Like those were phenomenally fun. Like I remember when that was a blast. I also used to do design workshops with participants and I used to we used to collage, I used to collage with five participants as a methodology for developing mental models and understanding mental models and customer journey maps.

And so what I would recommend doing is if you are looking to experiment and fail, is doing like micro experiments with things in on, like low stakes projects. So let's take like a usability test, for example. A lot of people don't have the ability to do generative research at their organization.

So I always say book an extra 15 minutes for every usability test and for that 15 minutes ask a generative research question or ask a question in a different way that you would never actually ask it. Or, with a team that you trust, run a workshop that you've never run before, but do it on a, like a very sm low stakes project.

So that's what I try and do is utilize low stakes environments and low stakes situations in order to do that experimenting or try to partner it with something that I know will work.

Do you have a a take on what's worked well for you in the past Julian?

Yeah, in terms of sorry, I wasn't sure if we had time for q and a or if I'll try to wrap it up quickly so we don't, because we have like just some minutes left. I think the, when the conversation is about, like you get one shot, you need to then think of what a, like a minimum viable project and try to really be sharp.

It's okay, you give me me this time. Then I would ask, I would answer two questions for you and reduce the scope and make it really specific. So that way you ensure that even if you have limited resources, limited time and limited everything, you can still get some answers. And it's with what I had, this is what I could deliver.

If you try to, boil the ocean in in because you got this is my research chance and like now I'm trying to do everything. That's probably not going to work. So be my, if you have, if you want to reduce the risk, like just go minimal and try to get like a good answer to one question, and then that will lead to a second question, and then that will lead to another one.

It's usually like that, it's like a little flywheel, it's one question will fit the other, and at some point you will get the ball rolling.

Awesome. And thank you for keeping me honest. Yeah we don't have I promised 10 minutes. We've got about five, four minutes. So let's try to get through as as many of the questions as possible. So I might start with Gaia here. And she asks if we can tell us if we can share a little bit more about the challenges and steps that we've taken to build research cultures around places that we've worked in, what was successful and what has failed.

Yeah I've tried to make big sound. It's big one. Yeah, it's a big one. I can give a really long word. I actually, it's three chapters of my book on this, so probably I'll try to keep it to one minute. First I, one of the key things that you need to get right is you need to map out and understand the dynamics.

So in terms of who, how, what kind of experiences people had with research, how do they make decisions? What's important to them, how are they evaluated or how are, is their product measured? You need to understand stuff from their perspective and then see how you can plug research into that. Also it will depend on how much.

Appetite for research they got, if you need to help them develop that appetite, what can you do for that? So basically there's a lot of mapping to be done at the beginning. Mapping, what kind of research is happening, mapping who's who, mapping stakeholders, understanding how decisions are made, understanding dynamics, understanding how people might think or how might, how people might go about because the company dynamics and so on.

And I think that from there, and this is I will not go forward because there 15 different answers is 15 different ways this can go. But like from there you can start finding your way, if, this company has, like, all these people have like less appetite and research, you can start with building allies on this part and talking to these roles and, presenting stuff this way.

And then from there you work your way up well, your way up your way, like horizontally. You just like your way around. True.

No notes.

Maybe under one minute. Yeah.

Yeah.

One other interesting question here Lucia has asked which cross-functional relationships are most important to build to ensure research has impact across the organization? And I suppose the typical one we hear of is, product. But yeah, I'm curious to, it'd be interesting to get your perspectives on it.

I would say I actually think that building with marketing, with sales, with customer support, customer solutions account management, like literally every single department has something that. Either would benefit you. So account management, for instance, with recruitment or customer support with like data triangulation, like every single department has something that like, where there's a symbiotic relationship with research.

So I would say that there aren't like certain cross-functional relationships. It's who in your organization is and can be the one who's using your research the most, who then can promote it to other people. So you want that domino effect. So looking towards people in these different departments who might be struggling with a problem that you could solve with your research.

I think those are the most important relationships, agnostic of departments. But look throughout all departments. Find the pain points that you can solve with your research and then that will amplify research across an organization in so many different ways. And that again is agnostic of like role department hierarchy.

Just look to solve people's problems because your user research is a product that solves people's pain points at your organization. And your stakeholders are your users. So maybe thinking about it more like that, find the stakeholders who you can solve pain points for and start with them and keep going and building it. 'Cause that's how it gets amplified.

Yeah. I would say everybody, thanks for coming to my TED talk. Agree. Thank you. Yeah. As much.

I like the I totally agree, but I do the the suggestion of kind of building close relationships with sales and marketing, just because we you both mentioned drawing that line to ROI and I think like when you can point to an impact to like a dollar number.

I appreciate that's not always the most sort of gratifying thing to be pointing to, but but yeah, it can be quite persuasive.

Some of my best products or some of my best products, some of my best projects were outside of product.

Yeah, makes sense. We are at time and we do have a few more questions, but I, we might just make an effort to to share some answers asynchronously don't want to encroach on, on people's time and I want to be respectful. So thank you, Nikki. Thank you, Julian. It's been a real pleasure speaking with both of you. I'm just going to put up some, a slide here just so that people if anyone's interested in getting in touch or connecting with you there's a slide here with your social links.

I'll leave that up. Thank you everyone for coming. Obviously we could spend a lot longer talking about these things. There's, it's quite nuanced. So I feel like it's fair to say that we've only really just scratched the surface. It sounds like there's a book that we're all going to be looking forward to that kind of touches on a lot of these.

So I'm excited to to get my copy. And yeah, it's been a real pleasure. And I'm keen to have more of these conversations. Definitely looking forward to opportunities to, to chat with you both again.

Yeah. Thank you so much. It's been real fun to talk about all this stuff. So glad to be here.

Thank you. Thanks for having me. Also, pleasure to be here, so feel free to get in touch also for any questions, any comments? We are around.

Awesome. Thank you to to everyone who joined us today, and we'll be sending out a link to the replay video and yeah, be seeing you all again soon.