00;00;06;23 - 00;00;21;03
Speaker 1
Hi. Welcome to Dig In the podcast brought to you by Dig Insights. Every week we interview founders, marketers and researchers, from innovative brands to learn how they're approaching their role and their category in a clever way.

00;00;23;22 - 00;00;49;29
Speaker 1
Hello. Welcome back to Dig In. This week we have Kathy Cheng from Nexxt Intelligence. She's the founder and CEO and we are excited that today we're also going to be talking about AI again, specifically conversational AI and how that's affecting market research. And I'm sort of adding to the amazing research that some of our clients are able to do.

00;00;49;29 - 00;00;52;17
Speaker 1
So Kathy. Thank you so much for joining me today.

00;00;53;10 - 00;00;54;28
Speaker 2
Thank you so much for having me, Meagan.

00;00;55;16 - 00;01;13;11
Speaker 1
Yeah, of course. We're both based for both based in Toronto, is that right? Yeah. Yes. It's nice to talk to it these days. Yeah, it's nice to talk to a fellow Torontonian. I'm hoping, seeing that correctly. Why don't you tell me a little bit about your background?

00;01;14;18 - 00;01;42;18
Speaker 2
Sure. So I've been the industry for a while. I actually started to be involved in the industry as a simultaneous interpreter for focus groups when I was a student in university. So, yeah, I wasn't intense, but fun job for a student. I did that for a while. Upon graduation, I got a job offer from an agency and my mentor, I still believe she is one of the best qualitative moderators.

00;01;43;18 - 00;02;04;29
Speaker 2
I told her that I was going to go to take on that job, and she said, No, no, no, please. I think you should be a moderator. So yeah. So then I just I thought I decided I'll just be a moderator. Fast forward 20 plus years. I'm very, very grateful that she pointed me to this direction. I really do think this is a job that I enjoy.

00;02;05;18 - 00;02;18;27
Speaker 2
So, yeah, that's a that's the bulk of my career for a while. I came to Canada about 20 years ago. After I came to Canada, I started to do more punked. So yeah, so I'm more like a hybrid kind of researcher. I do love both.

00;02;19;23 - 00;02;22;15
Speaker 1
Very cool. And when did you set up next Intelligence?

00;02;23;12 - 00;02;47;17
Speaker 2
That was seven years ago. Yeah, seven years ago. There was a very intriguing project. And also just at that time, almost like everything came together, I had the feeling that time and I wasn't very happy about the kind of quantitative research I was doing at that time because some ideas all came together. Yes, that's the beginning of mixing photons.

00;02;48;15 - 00;02;55;14
Speaker 1
Very cool. And just for anyone who's not aware of next intelligence, how would you describe it? Like, how would you pitch it to me?

00;02;56;23 - 00;03;21;19
Speaker 2
A We are a technology company. We build conversational AI We we say we build conversational AI solutions because we have a platform that is powered by conversational AI, but it's not just a platform. We also have a conversation as a service, like a microservices that can be used as a plugin. I guess at API to power other service platforms.

00;03;21;28 - 00;03;33;18
Speaker 2
That is actually the latest new development. So yeah, it's a multiple microservices but under the umbrella of conversational AI and that's, that's what we do.

00;03;34;20 - 00;03;41;03
Speaker 1
Okay. And to take this one step further, what is conversational AI? What does that mean to you?

00;03;42;18 - 00;04;19;06
Speaker 2
Conversational AI is the AI that mimics human conversations, actually, especially in the context of market research. I think I use for conversation. That's probably not a new concept because we've had chat bots for customer service for a very long time now for market research. Our purpose is slightly different, almost the opposite. Chatbots services, they're built to answer questions, while for us our goal is just like moderators.

00;04;19;26 - 00;04;53;21
Speaker 2
Our goal is to be there to to listen, to understand, to acknowledge, to build our report, and to ask good questions so that the ability to of asking good questions serves the purpose of getting report because the other side of the conversation, the participant feels like someone is actively listening. Is there understood. So that's engagement. At the same time, for researchers, obviously we always want a deep, deep insight so that that ability helps us and gain deeper insight.

00;04;54;05 - 00;04;58;19
Speaker 2
Now, in the context of the AI, obviously that deeper insight is made possible.

00;04;58;22 - 00;05;20;24
Speaker 1
Still very cool. I mean, that deeper insight piece makes me think of what you said at the beginning, which is you're so happy that you went down this road of qualitative research. What is is it the fact that you can get such deep insights through qual research? What is it about qualitative research that you really love?

00;05;22;07 - 00;05;29;25
Speaker 2
Yeah, I think initially I, I felt it was more about the thrill and the thrill.

00;05;29;25 - 00;05;30;17
Speaker 1
I love it.

00;05;31;16 - 00;05;58;08
Speaker 2
I think I probably spent too much time of my university years sitting in the back room of our focus groups. But as you sat there doing the interpretation, a lot of times you just got really fascinated by some really good questions that the moderator was asking. We're just really sometimes surprised. How could how could the direct people to that direction to get that really interesting insight.

00;05;58;15 - 00;06;17;12
Speaker 2
So so that's that's one side. But there are there were other times that you were sitting there just thinking, I wish I was the person sitting in the front room. I would ask questions differently. Right. So, yeah, well, that's interesting. Initially, I think it was that. So by the time I got into the job myself, I thought it it's actually much more challenging than you then just as an observer.

00;06;17;17 - 00;06;40;15
Speaker 2
But it was very useful to have that vantage point. But after I did it for a while, I started to realize what's really interesting about this job is maybe because I've done this voice for so long, I am convinced that in this role there's probably no such thing as true or false or like what's really right or wrong.

00;06;40;24 - 00;07;06;24
Speaker 2
It's really just perspectives. So so almost like I feel like I get paid to hear different perspectives and to learn from people that I can relate to because they are these are just normal, regular people, just I honestly, I feel like I've learned so much more from my research participants than maybe biographies from celebrities, from like wassup to use to to life philosophies, for example.

00;07;06;28 - 00;07;17;06
Speaker 2
Yeah. So I think that's that's the that's the part that's really, really useful and enlightening.

00;07;17;06 - 00;07;49;03
Speaker 1
That sounds like a job I would want to do. It's one of the reasons I love doing a podcast is because I get to chat to really interesting people on a regular basis, and I'm wondering this is kind of taking us off piste a little bit, but is there any advice you can give people who maybe don't have a strong background in qualitative research or in moderation about how to ask the right questions to get interesting or relevant feedback?

00;07;49;04 - 00;07;57;26
Speaker 1
I mean, where do you start when you're trying to to create the right types of questions with the right type of environment for people, to be honest?

00;07;58;24 - 00;08;21;15
Speaker 2
So that's actually a really good question because we have to ask that question when we train our bot to ask questions, right? So we'd be thinking about that a lot. That's definitely not an easy job. It's not like our customer service for you to have a position for you basically, you know what people will write like there are just this many information leads that people make.

00;08;21;15 - 00;08;26;09
Speaker 1
There's like a specific number of jobs to be done. Is that one that's customer service? Yeah.

00;08;26;20 - 00;08;48;07
Speaker 2
Yeah. But when it comes to qualitative research, the the fact that people are doing well is because they don't know it's exploratory. So it's really up to the asking part to do a good job, to do to to dig out the really useful information. So the guiding principles we've been using when we train are both really there are two things.

00;08;48;17 - 00;09;04;21
Speaker 2
One is, first of all, it needs to be on topic in these to be specific to the research objectives. These days, chatbots can do a lot, I mean, especially in the context of the past couple of weeks. The chat up really.

00;09;04;21 - 00;09;07;00
Speaker 1
Oh my gosh, it's taken the world by storm.

00;09;07;12 - 00;09;29;01
Speaker 2
I know. Yeah, it's exciting. But at the same time we'd been exploring LGBT for a very long time. It's the same exact exact challenge. It's it can do so much, but how can we make sure it does what exactly you want it to do? It's the thing we can use. Use the same analogy as a moderator. I mean, there are good moderators.

00;09;29;01 - 00;09;59;03
Speaker 2
There are not as good moderators. Sometimes difference is really how they can ask the questions specific to the research objective in that moment. That is that is critical. And I think the second thing is this is this is the first the guiding principle. The second guiding principle is empathy. As a moderator, I think that building that rapport to make sure that people want to talk to you are able to talk to you, that is very important.

00;09;59;11 - 00;10;31;07
Speaker 2
Sometimes people may may feel certain things may try to articulate, contribute, but it's just very difficult to articulate some of the deepest the feelings, some of the what we call the drivers people may not be able to articulate. So it comes down to tools that we can use to make it possible to make it enabling. So so yeah, so these two things, I think probably they apply to human moderators as well.

00;10;31;20 - 00;10;36;17
Speaker 2
Essentially, if we could do these two things, most of the time we should probably be okay.

00;10;38;02 - 00;11;01;04
Speaker 1
And I feel like I've kind of buried the lead here. But when you say it probably applies to human moderators as well with the conversational AI tool that you're building or that you've built, is it designed to replace human moderation? How do you see it working with other elements of research?

00;11;01;04 - 00;11;27;25
Speaker 2
I personally don't think qualitative research can ever be replaced. Maybe because I'm just such a happy, proud, funny that there are elements. So I think what there is a spectrum so that the type of cool that I'm very, very proud of, that I do not believe can be replaced by is a deep, deep flaw involves not just a conversation.

00;11;27;25 - 00;11;57;29
Speaker 2
And sometimes observation. Sometimes it's not just the conversation between you and me. Sometimes it requires stimulation, whether there's focus groups, it requires people talking to each other to constantly come up with new ideas. It requires possibly to even live their lives, demography. So it's a lot more than just conversation. The people that on the other end of the spectrum, what we call a light point, I think will also serves the purpose of just listening.

00;11;57;29 - 00;12;31;03
Speaker 2
Sometimes it doesn't even have to be too deep, but it provides an opportunity for participants to voice their own opinions in their own words. This versus want. The difference is want. We still call it voice of customer, but the way we collect the voice of customer is by providing some opinions to them and then basically they check right on while for the light side of the pool, it provides the opportunity to for them to articulate their own voices.

00;12;32;04 - 00;12;40;18
Speaker 2
This type of flaw, it's already been replaced by technology. There are clipboards write right on all those new communities.

00;12;40;19 - 00;12;40;26
Speaker 1
Yeah.

00;12;41;18 - 00;13;10;22
Speaker 2
Yeah. So so that's it's obvious that the skill is is is good is helpful and then that's very indepth kind of cool techniques may not always be necessary. So so this is the type of tool I think could benefit a lot from conversational AI because conversational AI can actually bring some additional benefits to just listening, because the essence of quiet is supposed to be two ways.

00;13;10;29 - 00;13;38;15
Speaker 2
It's a conversation, it's reciprocation. While the latest development of those listening type of call enabled by technology is almost like a lot of times going back to the wild way, it's almost like a longer form of the survey. People can do their own work at their own pace. With conversational AI, we can actually enhance that experience to make that type of call even more interesting.

00;13;40;09 - 00;14;11;29
Speaker 1
I think the question so that totally makes sense. It actually it makes me think about how things have moved and changed within the quant world. So that moved to sort of digitizing quite happened maybe earlier on than in call because of what's been going on with AI. And you know, just it was a little bit easier to replicate a lot of the sort of in-person quant and move it more mobile friendly, move it more online.

00;14;12;28 - 00;14;30;07
Speaker 1
But I'm wondering how do you see Quant and Klawe working together in a platform like yours? Do you see them working together? Do you think that quantum qual should be merged for them, like lighter qual aspect? What does that look like from your perspective?

00;14;31;06 - 00;15;09;02
Speaker 2
So this I think I've heard a qualitative depth at quantitative skills for a very long time now. So yeah, so it's clear that there probably is a place for this kind of hybrid approach. I've heard skepticism as well at the same time. So I think the fact that we've been talking about it possibly because two reasons. One is the industry has set up these two types, the two methodologies as two separate disciplines, possibly because of the availability of technologies.

00;15;09;25 - 00;15;35;09
Speaker 2
There's also this pond. But at the same time, we've been struggling for the longest time. That plumbed is wonderful, but it doesn't have enough depth or is wonderful, but it's sometimes not rigorous enough. So it's almost like we hear both sides of the voices all the time. But then the skepticism. I can understand that to be as again, as a cool person, I really, really don't like to say our platform, for example.

00;15;35;09 - 00;15;56;03
Speaker 2
I really don't like people to think about us as a qualitative and quantitative scale. Okay? I think it's the same kind of qualitative insight. It is. It is more open ended, it's more listening, it is more in-depth, but it's like if we just lightly use the qualitative insight as if it's the same as the deep well, it could confuse people.

00;15;56;10 - 00;16;18;15
Speaker 2
That's probably what the skepticism comes from. So I think as an industry, we should do what we can to provide that benefit of the hybrid approach to more and more clients while managing that expectation to make it very clear that when we talk about that hybrid approach, what exactly is the type of lapse in gaps inside that we can get at scale?

00;16;19;07 - 00;16;43;13
Speaker 2
How is it different from real clock? And what then? Is it hybrid? Is it a methodology that sits in between, or is that in our view? Actually, we think we still believe there is one. It's just more like what inspired or enhanced the plot. And then there are sequels. It's possibly like font enhanced the plot, but there's kind of still a little different because the process can be different.

00;16;43;29 - 00;17;00;11
Speaker 2
How you recruit people can be different. So so I think, yeah, this is a very intriguing topic. If we manage all the expectation properly, I believe more and more clients will benefit from a hybrid approach, either enhance the point what you have to for.

00;17;02;13 - 00;17;14;00
Speaker 1
The where does the conversation or the late qual or the conversational AI, where does that fit within the model you just describe? Not the model, but the the two sort of sections. Where does that fit?

00;17;15;18 - 00;17;54;06
Speaker 2
We from our point of view, currently we focus very much on what inspired want. It's more a more in-depth understanding at scale. However, we also do use the same technology methodology for Knightfall as well, when it's just more dependent on the use case. Sometimes if the universe is just so small, it's not simply not possible to have that scale or not beneficial, then we can easily pivot to a different process to make it more quality, but with some of the quantum benefits added to it.

00;17;54;27 - 00;18;24;21
Speaker 2
So it's more fluid. But at the moment I think the most benefits we have seen is on the con side because there are the benefits are pretty obvious. One, it feels that engagement because we know that in full quantitative survey we've been talking about the quality of our responses for a while now. The industry has been working very hard identifying that response to kick out the bad respondents or robot respondents, for example.

00;18;25;06 - 00;18;55;02
Speaker 2
Yeah, we believe those are all necessary that possibly sometimes we should probably should think maybe we should also try to provide a better experience for our participants so that it's not that respondents being bad, it's us whether we give them the best experience. Yeah. So. So with conversational AI, it's two way, it's conversational, it's more human that can help.

00;18;55;06 - 00;19;35;19
Speaker 2
And we've seen we see enough evidence that helps a lot. And second, we can get the depth of information and then we can get the speed. I should have said conversational. It does apply 2 to 2 quizzes. One is data collection, the whole conversation right? We also apply conversational AI on the back end as well. When we get so much text data back, we need to find a way to to use it properly quickly so that conversational AI does apply in text analytics to to kind of quickly summarize the key themes so the researchers don't have to go through.

00;19;36;14 - 00;20;00;02
Speaker 2
I'll use transcripts, don't have to go through lecture transcripts like ours in our space and sometimes days and weeks. Really. So, so so currently we see the most obvious use case application of benefits. They seem to lie more on the front side, the plus side. It could be beneficial for the life.

00;20;01;05 - 00;20;28;25
Speaker 1
Right? I know that that's some of our analytics team at Digg. I know that they do quite a bit of work with that. I to speed up some of the the quantum qualitative analysis on the back end. So no, I definitely understand where you're coming from there. I mean, this whole debate around qual and quant and you know where the industry is going and how we leverage AI within both of those types of research.

00;20;28;25 - 00;21;01;11
Speaker 1
It makes me think about, you know, I guess it was maybe it was more than ten years ago, but this idea of different So on the creative agency side, you had these specialized sort of like branding agencies, advertising agencies, UX agencies, campaign management agencies, and they all started to form sort of one mega agency, if you will. So they started to move together and say like, you know, we're specialists in all of these things.

00;21;01;11 - 00;21;31;02
Speaker 1
And it seems like from a research agency or research specialism standpoint, it's the the waters are getting more muddied. That sounds like it's negative, but it's not. It's they're sort of coming together to provide. So, for instance, with Digg, we started out purely quant and now we have one of the biggest qual, you know, agencies in qual sections or agencies in Canada, which is super exciting.

00;21;31;02 - 00;21;55;25
Speaker 1
But I think it's because people are really looking for that. They're looking for that like dual qual and quant, like, as you said, they want the depth that you would get through qual for certain parts of their research needs, but they also want that scale that you would that you would get through quant. So yeah, as you were talking I was just thinking, man, I wonder where the agency sorry where the agency landscape is is sort of moving.

00;21;55;25 - 00;22;03;15
Speaker 1
Do you have an opinion on that? Like as you think about building out next intelligence, do you have an opinion on where you think things are going to move?

00;22;04;20 - 00;22;36;27
Speaker 2
And I think that is very exciting. And your observation, I totally agree with you. I think clients, their needs are multidimensional as an agency to best meet the client's needs, We gradually become multidimensional. So that's that's natural and also that's concentric. So from our perspective, I can see that maybe I'll just use multi modality from the technology perspective, this is happening as well.

00;22;38;11 - 00;23;23;23
Speaker 2
We see even though we focus very much on conversational AI with the explosion of AI explosions, maybe I'll like this aspect. Yeah, we probably can expect that multi modality to happen very, very soon. Right. Basically, yes. As we talk about like an expert in terms of expertise, we as a one agency, we have both that quote unquote in the future in terms of technology, possibly within one solution or one agency, there will be many, many microservices for example, there could be VR.

00;23;24;01 - 00;23;51;18
Speaker 2
This is what I've been educated by our CTO. Josh There could be VR, for example. Then there is conversational services and you really don't need a human to be there to kind of interpret chat with the person experiencing that. And then there there's also services that can really read the content in the brain at the same time. So you get you get the information, not just the articulated opinions perspectives, but also the subconscious.

00;23;51;18 - 00;24;10;09
Speaker 2
How exciting is that? It's like it's possible now. It's just it's just happening so quickly. So I'm just trying to yeah, I'm just thinking responding to the analogy of the multidisciplinary services that we could provide. I think in terms of technology, this is probably going to be the future as well.

00;24;10;26 - 00;24;32;01
Speaker 1
That's a very good point. I didn't even think about it in that, in that in that way, but it totally makes sense. You know, we're exploring A.I. is all over the place. Everyone is talking about it. We're exploring how that can be applied. But I'm sure that there are so many other modalities that can be applied to the client needs that that were meeting.

00;24;32;01 - 00;24;51;18
Speaker 1
So very interesting. I'm very conscious of time and I want to ask you a couple of questions that we ask at the end of every one of our podcasts. The first being if you gained double your budget tomorrow as a business, what would you spend it on?

00;24;53;12 - 00;25;20;22
Speaker 2
As of today, I would say probably marketing. Okay, Now we appeal. This is happening so quickly. We have really good things to offer to help clients and it's more about how to get more clients, get to know us to benefits from new technologies. So yeah, that's probably the first thing we'll think about.

00;25;21;20 - 00;25;26;18
Speaker 1
Very cool. And then on the inverse, if you lost half of your budget, what would have to go?

00;25;27;29 - 00;25;49;03
Speaker 2
I certainly hope that's not going to happen. That's that I would say we have such a huge long list of features to be added. I think Roadmap, it's very, very, very exciting. So I would think that if we had to cut, we would have to be more disciplined to what to focus on. I mean, this has happened sometimes it's not a bad thing.

00;25;49;03 - 00;26;00;07
Speaker 2
It just makes you become more disciplined to really understand where the market is going and then to really focus on what you have to focus on. So yeah, that probably would be what we have to do.

00;26;01;20 - 00;26;10;08
Speaker 1
And finally, to our listeners who are typically people who work in research, would you leave them with any advice?

00;26;11;10 - 00;26;58;09
Speaker 2
I think maybe human centricity is probably even more important than ever. It's interesting. I was just thinking as you were talking, Meghan, I was thinking we are in the business of understanding. What we do really is to provide insight for clients to develop human centered products, for example, to consumers. Research probably could use a lot of that human centricity as well to develop that research process to experiences the best it possibly can for humans and including participants, including research creators users.

00;26;59;27 - 00;27;22;05
Speaker 2
In the context of we're talking about it today and everyone is talking about yeah these days. So I think it's yeah is really, really exciting. No doubt that I will become more and more available and accessible very soon. I think very soon that we'll see. It's not about what to develop, but rather how to package, how to use it.

00;27;22;14 - 00;27;58;21
Speaker 2
It's almost like it's actually submitting a paper to echo one of the top tier conferences with our academic advisor at University of London on the topic of cost benefit tradeoff. So it's it's a matter of how to it's almost like conjoint analysis like we do for like just the products. When air becomes a commodity, it becomes how can we packaged a product that is really useful, beneficial to our users.

00;27;59;21 - 00;28;15;22
Speaker 2
Then it's, it's human. It's about we should probably switch our focus more on the human needs. The AI will be there will be there will be a lot of attention. But what really will continue to make a difference is that human centricity.

00;28;16;17 - 00;28;38;17
Speaker 1
Yeah. So essentially what you're saying is you need the air is so exciting and people don't really know what to do with it to some extent, like if you are thinking of like the average person, it's like, Oh wow, you know, for instance, I just keep seeing people creating AI versions of themselves on Instagram right now. There's like, you're able to do that.

00;28;38;17 - 00;29;14;18
Speaker 1
And then I was playing around with the open API chat bot asking questions. Right now it's like for the for the layperson, it's it's almost like a little gimmicky in the sense that you're thinking of like, Oh, what are some like funny, cool, hilarious things I can do with this? But when we think about it in market research, there's going to need to be a lot of sort of care taken to what actual plea is this useful for and how do we package it in a way that demonstrates how useful it can be to to the research community, which totally makes sense.

00;29;14;18 - 00;29;30;29
Speaker 1
I mean, yeah, it's almost like you need a ton of product marketing, a ton of product marketing to, to figure out exactly where it gets used and how and create some boundaries around that to make sure that it's high quality inputs and outputs.

00;29;31;26 - 00;29;55;07
Speaker 2
Yeah, the purpose, right? I mean, we can all have fun with some AI development. It is really fun. But then when it comes to commercialization, it's the purpose is how we can do things that can actually serve the purpose. And then that's where inside, that's where that deeper understanding really play a role.

00;29;55;15 - 00;30;11;25
Speaker 1
That was great advice. So people need to think about human centricity. Like the air is very exciting, but you always need to bring it back to the human and how we can actually apply it in a way that's human centric.

00;30;11;25 - 00;30;39;04
Speaker 2
Yeah, yeah, yeah, I would. Yeah. I think that's not what you said. It's exactly that's the, that's the essence of it I think. Yeah, essentially. Yeah. AI is here too. I don't know. I might be naive. It probably isn't going to be the case, but I still believe that is an assistant. And at the end of the day it's the human creativity that make the best use of AI.

00;30;39;04 - 00;31;06;08
Speaker 2
It's the understanding, human understanding of the purpose that so. So if I, if I could leave any final thoughts to the listeners, I would think our experience has been that human centricity is so key to everything we do. And then that's almost like the guiding principle every time. If we're too fascinated by certain technologies, we come back to think, Is this what we really need?

00;31;06;08 - 00;31;08;28
Speaker 2
Is this Does this serve any specific purpose?

00;31;10;19 - 00;31;38;27
Speaker 1
Thank you so much, Cathy. That was very eloquently said, I am so happy I got to chat to you today. We are doing a ton of work internally on A.I. and market research as well. We've just we've got a bunch of articles out on our blog, so please go check out our resources section. And yeah, I'm hopeful that we can chat again in the coming months to talk about where things have progressed to.

00;31;40;03 - 00;31;44;13
Speaker 2
Yeah, I look forward to chatting more. This is such an exciting topic. Thank you so much for having me.

00;31;44;28 - 00;31;57;28
Speaker 1
Of course. See you next week guys. Thanks for tuning in this week. Find us on LinkedIn at Digg Insights and don't forget to hit subscribe for a weekly dose of fresh content.