Ep 21: Augmented Intelligence for Market Research with Victoria Sakal and Ainesh Ravi

[00:00:00] [SPEAKER_00]: Hey, you, with the podcast, you're right.

[00:00:03] [SPEAKER_00]: You've already activated your mobile app, then you'll have to wait for a few days for a long time to start the game.

[00:00:10] [SPEAKER_00]: It's great to have you and all of these mobile functions are now available for the home of the telecom.

[00:00:16] [SPEAKER_00]: Now you can see one hour.

[00:00:18] [SPEAKER_00]: The mobile app is in the mobile app of the Magenta app and it's already up to you.

[00:00:22] [SPEAKER_00]: In the 5th of the day, you won't be able to activate it.

[00:00:26] [SPEAKER_00]: in their mind magenta app de telecom.

[00:00:36] [SPEAKER_04]: For one spot, pull over.

[00:00:42] [SPEAKER_04]: In this episode I'm joined by two great guests,

[00:00:44] [SPEAKER_04]: I Nish Ravi and Victoria Secau from Wonder,

[00:00:47] [SPEAKER_04]: a startup that combines human expertise

[00:00:49] [SPEAKER_04]: with AI to provide market research solutions.

[00:00:53] [SPEAKER_04]: Personally, I was a market researcher when chat

[00:00:55] [SPEAKER_04]: chief of TV was publicly released

[00:00:56] [SPEAKER_04]: and I worked at IBM's Market Intelligence group many years ago.

[00:00:59] [SPEAKER_04]: So I'm fascinated by this topic.

[00:01:02] [SPEAKER_04]: Victoria, I Nish and I dive into the evolution of Wonder,

[00:01:05] [SPEAKER_04]: the challenges and benefits of incorporating AI

[00:01:06] [SPEAKER_04]: into their workflow and the importance of human expertise

[00:01:09] [SPEAKER_04]: in the research process.

[00:01:11] [SPEAKER_04]: Our conversation explores a role of AI in the future of work

[00:01:14] [SPEAKER_04]: is discussing adaptability, productivity enhancement,

[00:01:17] [SPEAKER_04]: and how to navigate changing landscape of AI and market research.

[00:01:21] [SPEAKER_04]: Hope you find this discussion is interesting

[00:01:22] [SPEAKER_04]: and it's like full as I did.

[00:01:24] [SPEAKER_04]: Thank you again for tuning in.

[00:01:26] [SPEAKER_04]: Hi everyone, welcome to another episode of LVature AIQ.

[00:01:30] [SPEAKER_04]: I'm your host Bob Pover with me today

[00:01:32] [SPEAKER_04]: or two guests from an interesting startup called Wonder.

[00:01:37] [SPEAKER_04]: I Nish Ravi and Victoria Secau.

[00:01:41] [SPEAKER_04]: Thank you guys for being here.

[00:01:43] [SPEAKER_03]: Excited to be here and it's an interesting conversation

[00:01:45] [SPEAKER_03]: saying excited to get into it.

[00:01:47] [SPEAKER_04]: Yeah, absolutely.

[00:01:48] [SPEAKER_04]: So just to kick us off,

[00:01:51] [SPEAKER_04]: we're not each of you just give a little bit about

[00:01:53] [SPEAKER_04]: your background and how you got to wonder.

[00:01:59] [SPEAKER_04]: Yeah, I'm just curious about the paths that each of you took.

[00:02:01] [SPEAKER_04]: So Victoria, why don't you go first?

[00:02:02] [SPEAKER_03]: Sure.

[00:02:03] [SPEAKER_03]: I'm Victoria Secau running sort of growth,

[00:02:07] [SPEAKER_03]: go to market effort specifically.

[00:02:09] [SPEAKER_03]: How do we get more people to know about Wonder?

[00:02:11] [SPEAKER_03]: And now with that company has really seen a lot of great success

[00:02:15] [SPEAKER_03]: with different types of clients across the board.

[00:02:17] [SPEAKER_03]: My journey to Wonder was squarely through I Nish.

[00:02:21] [SPEAKER_03]: We went to school together actually, so a fun little product.

[00:02:25] [SPEAKER_03]: I don't know how to grow a partnership here,

[00:02:27] [SPEAKER_03]: but my journey was through research through the primary side for

[00:02:31] [SPEAKER_03]: the most part.

[00:02:31] [SPEAKER_03]: I did consulting, worked with a few different companies,

[00:02:35] [SPEAKER_03]: had a few different types of clients over the years

[00:02:38] [SPEAKER_03]: working with worker 500 and then B2B SaaS companies.

[00:02:42] [SPEAKER_03]: But always found myself doing a ton of desk research and googling

[00:02:45] [SPEAKER_03]: and getting smart on my client or my industry

[00:02:48] [SPEAKER_03]: and so having saved a touch with my nature of the years

[00:02:51] [SPEAKER_03]: and hearing how Wonder evolved and developed

[00:02:53] [SPEAKER_03]: well shocked that there's no other solution out there

[00:02:55] [SPEAKER_03]: to make this whole space much easier around a glass started

[00:02:59] [SPEAKER_03]: in April 2023 when AI was really hitting the scene

[00:03:03] [SPEAKER_03]: in terms of, as been around,

[00:03:05] [SPEAKER_03]: which healer and it has been around,

[00:03:06] [SPEAKER_03]: but in terms of incorporating this into your business model.

[00:03:10] [SPEAKER_03]: So I missed almost a decade of Wonder's existence

[00:03:12] [SPEAKER_03]: but a new version of Wonder was born around the time that I joined.

[00:03:16] [SPEAKER_03]: So I'll pause there because I nationwide

[00:03:18] [SPEAKER_03]: that transition but also has been through all the evolution before

[00:03:21] [SPEAKER_03]: and I'm sure we'll get into it,

[00:03:23] [SPEAKER_03]: but that is how I learned it.

[00:03:24] [SPEAKER_03]: I learned it.

[00:03:25] [SPEAKER_04]: Thanks, my niche.

[00:03:27] [SPEAKER_02]: Awesome.

[00:03:28] [SPEAKER_02]: Yeah, I know.

[00:03:29] [SPEAKER_02]: I can't believe it's been about like 15 years or so of a tourist

[00:03:32] [SPEAKER_02]: since we've never had any other development.

[00:03:34] [SPEAKER_02]: It's been a long time.

[00:03:35] [SPEAKER_02]: I'm glad that it's led to,

[00:03:38] [SPEAKER_02]: what do we are today?

[00:03:39] [SPEAKER_02]: Yeah, life of a tourist set of wonders been a wild ride.

[00:03:43] [SPEAKER_02]: Right?

[00:03:43] [SPEAKER_02]: There's been kind of grew up in sort of the marketplace

[00:03:46] [SPEAKER_02]: time for startups whenever everyone was sort of the Uber for X sort

[00:03:50] [SPEAKER_02]: of company, whatever they're just trying to build sort of these large

[00:03:54] [SPEAKER_02]: marketplaces and I think we always knew that technology and AI

[00:03:58] [SPEAKER_02]: was going to be a big part of that piece.

[00:04:02] [SPEAKER_02]: The question was much more when.

[00:04:04] [SPEAKER_02]: So much that we even invested on our own side on the AI front

[00:04:08] [SPEAKER_02]: until we were actually, wait, this is a losing battle.

[00:04:11] [SPEAKER_02]: That's going to require a ton more capital to actually make this

[00:04:15] [SPEAKER_02]: make this a reality and hey, if other companies are going to

[00:04:17] [SPEAKER_02]: pour billions of dollars into it, we will certainly on the

[00:04:20] [SPEAKER_02]: application side of the door.

[00:04:22] [SPEAKER_02]: So the moment those models came out,

[00:04:24] [SPEAKER_02]: we were on it like white on rice.

[00:04:26] [SPEAKER_02]: And that's what sort of led to definitely a lot of the journey

[00:04:31] [SPEAKER_02]: like Victoria mentioned.

[00:04:32] [SPEAKER_02]: This is sort of the new version of Wonder which is heavily

[00:04:35] [SPEAKER_02]: technologized and it's been a journey, right?

[00:04:37] [SPEAKER_02]: To take a company that was very,

[00:04:40] [SPEAKER_02]: much more so sure we had that technology focus,

[00:04:44] [SPEAKER_02]: but at the end of the day we're still incredibly human-driven

[00:04:46] [SPEAKER_02]: and then taking that experience that we had from how do we actually

[00:04:50] [SPEAKER_02]: do research?

[00:04:51] [SPEAKER_02]: Well, and then translating that into AI.

[00:04:53] [SPEAKER_02]: And I think that there's a lot of really cool stuff, especially now

[00:04:56] [SPEAKER_02]: with chains and agents, especially that allow that institutional

[00:05:00] [SPEAKER_02]: knowledge to be capitalized upon versus just making one, you know,

[00:05:04] [SPEAKER_02]: open AI call and expecting it to come back with something that is

[00:05:08] [SPEAKER_02]: right.

[00:05:09] [SPEAKER_02]: There's a lot more complexity that you can build into the system based

[00:05:12] [SPEAKER_02]: on the institutional knowledge that I think we have,

[00:05:14] [SPEAKER_02]: just excited to keep building here.

[00:05:16] [SPEAKER_04]: That's awesome.

[00:05:17] [SPEAKER_04]: I feel like certainly I could have used more of these tools

[00:05:20] [SPEAKER_04]: thinking back to my market intelligence phase at IBM,

[00:05:24] [SPEAKER_04]: where we were sort of scrambling.

[00:05:26] [SPEAKER_04]: I mean, it was around when people started talking about

[00:05:31] [SPEAKER_04]: big data and do we have the horsepower to analyze the

[00:05:35] [SPEAKER_04]: data, do we have the right data, you know, relationships through

[00:05:38] [SPEAKER_04]: data aggregators and things like that to get it,

[00:05:41] [SPEAKER_04]: you know, trustworthy information.

[00:05:43] [SPEAKER_04]: And certainly we were starting to play around with some of the

[00:05:47] [SPEAKER_04]: IBM Watson powered AI stuff into me advanced analytics,

[00:05:51] [SPEAKER_04]: but one of the things I find interesting about what you guys

[00:05:54] [SPEAKER_04]: are doing first of all because our future is, you know,

[00:05:58] [SPEAKER_04]: humans plus AI and you guys are making both and basically

[00:06:03] [SPEAKER_04]: take the best of both in a sense and putting those together to come up

[00:06:07] [SPEAKER_04]: with this optimized sort of output for different clients.

[00:06:12] [SPEAKER_04]: And so were you helping them figure out, okay, now that we know

[00:06:18] [SPEAKER_04]: what large language models are, let's take a very close look at

[00:06:23] [SPEAKER_04]: where their limitations are so that we know,

[00:06:26] [SPEAKER_04]: as I imagine like what tasks or what portion

[00:06:30] [SPEAKER_04]: is done by a human versus what was done by, you know,

[00:06:34] [SPEAKER_04]: even advanced analytics before there was gender to the AI.

[00:06:37] [SPEAKER_04]: You know, that balance may not always be consistent, right?

[00:06:41] [SPEAKER_04]: I mean, it's going to take on more tasks and I guess

[00:06:45] [SPEAKER_04]: I'm curious if any of your experiences led you to some epiphany

[00:06:50] [SPEAKER_04]: that said, you know, this is the right approach.

[00:06:54] [SPEAKER_02]: Yeah, I'd say you hit the nail in the head, right?

[00:06:57] [SPEAKER_02]: It isn't moving target.

[00:06:58] [SPEAKER_02]: And I think the way that we've tacked it is more of a behavioral change

[00:07:02] [SPEAKER_02]: in terms of how our company thinks about this versus like a,

[00:07:06] [SPEAKER_02]: this is like a prediction that we have.

[00:07:09] [SPEAKER_02]: And for us, when I say behavioral change,

[00:07:11] [SPEAKER_02]: I think it like actually goes as deep as the organization structure

[00:07:15] [SPEAKER_02]: of the teams such that we actually now brought on a lot of our

[00:07:19] [SPEAKER_02]: individual really amazing researchers onto what we call our innovation team

[00:07:23] [SPEAKER_02]: to actually understand what the limitations are because they have

[00:07:26] [SPEAKER_02]: to deep understanding of like what amazing research looks like.

[00:07:30] [SPEAKER_02]: But they also now have played around with this technology and

[00:07:33] [SPEAKER_02]: playing around with the technology, day in and day out,

[00:07:35] [SPEAKER_02]: 40 hours plus a week where that sort of allows us to kind of bridge

[00:07:41] [SPEAKER_02]: that gap a little bit and help get over some of those components

[00:07:45] [SPEAKER_02]: of fear but also be extremely realistic with like,

[00:07:49] [SPEAKER_02]: hey, it's just not there yet with these things and we're not afraid of that,

[00:07:53] [SPEAKER_02]: right, that's okay if AI isn't great at these things.

[00:07:57] [SPEAKER_02]: But at least we know that there's a path forward for the AI to be really

[00:08:00] [SPEAKER_02]: great at these things or on the human side,

[00:08:03] [SPEAKER_02]: we build incredible processes such that our humans can be excellent in these things.

[00:08:07] [SPEAKER_02]: So I think that's sort of how we've tacked it much more like organizationally

[00:08:10] [SPEAKER_02]: how do we shift what the team structures look like and how we actually

[00:08:13] [SPEAKER_02]: look at really the evolving opportunity of AI,

[00:08:17] [SPEAKER_02]: not AI and it sticks a point in time.

[00:08:19] [SPEAKER_03]: The other interesting thing that the team thought through and I say the team

[00:08:24] [SPEAKER_03]: because I joined and I'm like, what, how is this going to change?

[00:08:26] [SPEAKER_03]: Like we were totally human-powered and now we're in this

[00:08:30] [SPEAKER_03]: inflection point of evolving the offer just till it's

[00:08:33] [SPEAKER_03]: in tangibility because the mindset piece is obviously critical for

[00:08:38] [SPEAKER_03]: bringing the talent along who made the business successful

[00:08:41] [SPEAKER_03]: and who are still going to be a very big part of this human-the-loop solution.

[00:08:46] [SPEAKER_03]: But then there's the process piece and so I guess even using the consulting

[00:08:50] [SPEAKER_03]: framework of people and processes and tools we need to think about

[00:08:54] [SPEAKER_03]: the actual process that we follow to take a client request all the way

[00:08:59] [SPEAKER_03]: through to request the liver or results and research delivered.

[00:09:03] [SPEAKER_03]: And so when we thought about the knowledge assembly line and the steps that we took

[00:09:07] [SPEAKER_03]: as humans, that was how very tangibly you break down all the micro steps

[00:09:12] [SPEAKER_03]: of first you have to take in the request as their way so we can

[00:09:16] [SPEAKER_03]: audit that and accelerate that. Even in that step, we have built a different approach

[00:09:20] [SPEAKER_03]: then a lot of what you experience on the LLNs where we're clarifying questions

[00:09:24] [SPEAKER_03]: or having a dialogue with this submitter instead of just, you know,

[00:09:28] [SPEAKER_03]: I'm interested in a competitive landscape and GPT's off to the races

[00:09:31] [SPEAKER_03]: and you're like, wait, I didn't mean global, I mean, you know, emerging players

[00:09:35] [SPEAKER_03]: whatever we have a dialogue the way a human historically have that conversation.

[00:09:39] [SPEAKER_03]: You sort of take each other steps of the process and you break it down into

[00:09:43] [SPEAKER_03]: what a success look like where there's an opportunity to automate

[00:09:46] [SPEAKER_03]: to make things more efficient, an opportunity to augment and you can do more now

[00:09:50] [SPEAKER_03]: because you're starting 50% of the way there because all the force

[00:09:53] [SPEAKER_03]: power of AI got you there in seconds and now you can take things much further

[00:09:58] [SPEAKER_03]: but just to make it really tangible for any listeners who are thinking through,

[00:10:02] [SPEAKER_03]: yes, you have to bring your people along and make sure there is, you know,

[00:10:04] [SPEAKER_03]: understanding or support of where they're almost remote as a human

[00:10:08] [SPEAKER_03]: contributor is but also how their actual workflow is going to evolve when they think about,

[00:10:14] [SPEAKER_03]: you know, almost the widgets going to this assembly line and what they're,

[00:10:17] [SPEAKER_03]: you know, taking his inputs and eventually creating his outputs.

[00:10:20] [SPEAKER_04]: I'm curious if you guys publish your own research.

[00:10:23] [SPEAKER_04]: I say that because I feel like so many people trying to figure out their own

[00:10:28] [SPEAKER_04]: workflows and trying to figure out their own, in part, their own workforce planning,

[00:10:35] [SPEAKER_04]: how are we evolving together?

[00:10:38] [SPEAKER_04]: How are we moving people to higher level tasks?

[00:10:42] [SPEAKER_04]: How are we only giving AI things that we know it can confidently do,

[00:10:48] [SPEAKER_04]: not just, you know, hack at in a way.

[00:10:51] [SPEAKER_04]: But yeah, I'm just curious if you guys have come up with,

[00:10:56] [SPEAKER_04]: you know, some novel approaches that would be relevant and beneficial to all my listeners

[00:11:02] [SPEAKER_04]: and way beyond.

[00:11:04] [SPEAKER_03]: The quick comment I can make is that we did have a, you know,

[00:11:07] [SPEAKER_03]: five weekly events series last year.

[00:11:09] [SPEAKER_03]: We sort of took attendees under the hood and showed how we are building things.

[00:11:14] [SPEAKER_03]: We can certainly link in the comments a couple of the resources that we shared for listeners.

[00:11:18] [SPEAKER_03]: I think in terms of groundbreaking and unique approach,

[00:11:22] [SPEAKER_03]: we, you know, I say the team because I had no one to know credit for it,

[00:11:26] [SPEAKER_03]: but you know, thought through how do we approach moving the business into this new world

[00:11:30] [SPEAKER_03]: and had this structure?

[00:11:32] [SPEAKER_03]: I don't know if it's materially different or, you know,

[00:11:35] [SPEAKER_03]: shocking or insightful, compared to other companies,

[00:11:37] [SPEAKER_03]: our world just kind of doing the same thing out as we go.

[00:11:40] [SPEAKER_03]: And it worked for us.

[00:11:42] [SPEAKER_03]: And so what we, you know, did spend a good amount of time and continued to do is talking through

[00:11:45] [SPEAKER_03]: here's a process we followed, here's how we broke down our steps and then you just need to translate that to your own workflow.

[00:11:51] [SPEAKER_03]: And that can mean as simple as like me when I go to create a marketing campaign,

[00:11:55] [SPEAKER_03]: there are 15 steps that I as a human take and all their steps within that very specific way.

[00:12:00] [SPEAKER_03]: The way that we do this is to be able to get the best of the task that can get automated,

[00:12:02] [SPEAKER_03]: and augmented or whatever.

[00:12:04] [SPEAKER_03]: Or, you know, as a full, you know, research function at a Fortune 500,

[00:12:08] [SPEAKER_03]: there's a million activities and you can map that.

[00:12:11] [SPEAKER_03]: But also say we can certainly link and help make it tangible.

[00:12:13] [SPEAKER_03]: I don't know if we've published any super-super-undid ahead semantics,

[00:12:17] [SPEAKER_03]: but I need to feel free to keep it on.

[00:12:19] [SPEAKER_02]: Now, I think, you know, as always a delicate balance of like what's the secret sauce?

[00:12:23] [SPEAKER_02]: Plus what also is why have the relevant?

[00:12:25] [SPEAKER_02]: I think, you know, maybe one of the things that we definitely am sure talked about in a few of the sessions that will link is this idea of prompt engineering.

[00:12:36] [SPEAKER_02]: And right how do we actually look at each step within the workflow and understand how successful we are at that step?

[00:12:44] [SPEAKER_02]: And not just how successful we are that in terms of that prompt but how successful it different model is.

[00:12:49] [SPEAKER_02]: Right, like our ability to choose which model determines how high quality that is,

[00:12:55] [SPEAKER_02]: how functional that is, right what the speeding costs, sort of considerations are for that step of the process.

[00:13:03] [SPEAKER_02]: And for us we've seen a lot of success on the prompt engineering side to just do a bunch of testing.

[00:13:08] [SPEAKER_02]: Thunder sent, okay, where's high queue amazing?

[00:13:11] [SPEAKER_02]: Where's Clouds on an amazing?

[00:13:12] [SPEAKER_02]: What's the value of GPT 40 many?

[00:13:15] [SPEAKER_02]: And I think that there's a lot of things in there that we probably could do even more, you know, publishing around that space because I think that's probably where we have a bit more experience.

[00:13:25] [SPEAKER_01]: Before we move on, I need to let you know about my friend Mark Feffer and his show people to.

[00:13:32] [SPEAKER_01]: If you're looking for the latest on product development, marketing funding, big deals happening in talent acquisition, HR, HCM,

[00:13:41] [SPEAKER_01]: that's the show you need to listen to.

[00:13:44] [SPEAKER_01]: Go to the work to find network, search out people tech, Mark Feffer, you can find them anywhere.

[00:13:52] [SPEAKER_04]: Right, okay, that makes sense.

[00:13:54] [SPEAKER_04]: Victoria, you said something at the beginning that I wanted to follow up on a lot of startups.

[00:14:01] [SPEAKER_04]: They have kind of jumped into the general AI with both feet and because of what.

[00:14:10] [SPEAKER_04]: And that's what we're talking about like the speed at which some of the capabilities are improving.

[00:14:17] [SPEAKER_04]: You've got this existential risk in a way if you can't keep up because if you're just tapping in to some of those for, you know, specific.

[00:14:28] [SPEAKER_04]: You know, tasks or you know, steps in your workflow or what have you but they're releasing it whatever frequency, you know, all these new capabilities.

[00:14:38] [SPEAKER_04]: I mean, I've seen some young granted their younger than than wonder, but I've seen some startup get just dropped it like everything they've done for the last six months was almost.

[00:14:48] [SPEAKER_04]: You know, the new technology, just the new release just basically made it all available and threw it out there like what's unique and what's differentiating about what you're building.

[00:15:01] [SPEAKER_04]: And at this point, I mean, you've got to have that right out of the out of the gate.

[00:15:05] [SPEAKER_04]: I mean, if you're looking for funding, they're going to ask you that right.

[00:15:08] [SPEAKER_03]: It's a great point. I mean, they're stressed out as a lot is a flick of the rest, right? You, the smaller companies can be.

[00:15:15] [SPEAKER_03]: Blood sweat and tears into solving a problem and open a eye throw is a cool million forget about billions at something and overnight they have a solution that competes.

[00:15:24] [SPEAKER_03]: The couple of things that we we sort of thought about and we push the most piece pretty heavily here actually think it does go back to the people process tool piece.

[00:15:33] [SPEAKER_03]: Obviously, we're betting on people right who've not only been researchers but we've trained them very strategically and very extensively to be executing a certain type of research.

[00:15:42] [SPEAKER_03]: Very well these are questions, you know, we just talked about the risk of hallucination mark their research is a challenge because quality is really important even if you're just getting directional insight on the ballpark size of a new market to maybe build something in you need to be like directionally right more than directionally wrong.

[00:16:01] [SPEAKER_03]: So we've got the human component. We just talked about AI being a part of the tools and the technology, the whole other bucket for us is IP and the process and having studied the space knowing what you know good competitive intelligence looks like.

[00:16:17] [SPEAKER_03]: You know, you might go to an LLM and ask for a competitive landscape first of all they're not going to ask you helpful questions to help really refine your ask are they going to go and know the right sources for you and the financial services and you know where you're emerging

[00:16:31] [SPEAKER_03]: players are going to exist there's so many layers to this around what good looks like in terms of not just the questions you asked not just what goes into a good competitive landscape not just the data sources.

[00:16:41] [SPEAKER_03]: Not just have access data sources that maybe you are eyes and average Googler aren't going to be able to access but that we're going to bring access to our clients.

[00:16:50] [SPEAKER_03]: I think more and more about this I guess the two pieces here are the combination of strategic value and you know having an intersection is where most tend to come from because someone can pick one of those things that if they don't have both are all.

[00:17:04] [SPEAKER_03]: Then you're selling advantage and I think there's a bigger conversation here around there's kind of operational risk if you're choosing to buy a vendor who is great at whatever AI solution they're offering and then they either go out of business or they release something and you built your entire staff on integrating.

[00:17:25] [SPEAKER_03]: So I think there's a lot of risk right now because things are changing so quickly and so that's again where for us.

[00:17:30] [SPEAKER_03]: Our humans and our experts are sort of the shop of the servers where something might change but worst case.

[00:17:35] [SPEAKER_03]: Humans are doing this research for a decade and we can still get to answer pretty damn quickly.

[00:17:40] [SPEAKER_03]: So there's a little it's tricky for people and I say this is someone who's shopping for different tools do I trust a lot of these AI tools and that they're even going to be there in three months isn't worth it to try them out now if I get used to them and they're gone so I think I can play for you around this.

[00:17:54] [SPEAKER_04]: Yeah, I want to just kind of take a step back and just make sure people understand like how wonder works in terms of like the workflow and the human in the loop around you know this type of.

[00:18:09] [SPEAKER_04]: Market research so.

[00:18:12] [SPEAKER_04]: But my mind just kind of walking through what happened you get a request from a client that you know this is something that I want to do a bit of research on I'm not I don't want to pay.

[00:18:24] [SPEAKER_04]: You know some big research from to spend you know months and you know ten people you know putting this together and the information that I think is in like all these different places.

[00:18:35] [SPEAKER_04]: And then you guys come in and say we can do that.

[00:18:39] [SPEAKER_02]: Yeah, yeah definitely can walk through the these experience which our focus is to figure out how to make it as simple as possible to execute on that curiosity.

[00:18:48] [SPEAKER_02]: Because I think whenever people have that curiosity whenever it's super complex to do it tends not to just get asked.

[00:18:55] [SPEAKER_02]: Or you do it in a very like hot hazard way so what we wanted to do and when we thought about okay in this new world of AI how do we make this experience as simple as possible.

[00:19:05] [SPEAKER_02]: And we also landed like many others didn't we have an quite early on this is like in a chat base interface where we actually looked at all of the different.

[00:19:12] [SPEAKER_02]: And we've actually had in chat form and also in person form over the past almost ten years we've been in business.

[00:19:21] [SPEAKER_02]: Tenderion okay what does like that perfect clarification experience look like this is over 180 plus thousand projects so we've done.

[00:19:29] [SPEAKER_02]: And we understood that actually wasn't just the straight clarification process it involved this way of thinking divergingly before we actually converge on like a set of specific things that we want to talk.

[00:19:42] [SPEAKER_02]: That gives the person this piece of mind that oh wait they covered everything that I wanted to cover and thought about things that I didn't even think of which breeds this like for much trust that you then have without a product that you're using.

[00:19:55] [SPEAKER_02]: Regardless of whether that's human or AI and then from there once you understand hey this is now the specific things we want to tackle we leverage it to actually go out.

[00:20:04] [SPEAKER_02]: Scour the public domain look at specific data sources so whether that's SEC fileings whether that's the different crunch base whether that specific.

[00:20:13] [SPEAKER_02]: You know links that are like that outline the particular companies products and services or you name we have one of your so different agents that run based on the type of project.

[00:20:25] [SPEAKER_02]: That we have and then usually that takes anywhere between seconds to minutes then from there we have an analyst go through and actually orchestrate between the tools based on what exists what we actually were able to find from the ASN and then we put that now the tools into the hands of the humans.

[00:20:42] [SPEAKER_02]: And then from there within 24 hours you actually have that four research report for that that customer so ultimately want to make it super simple to go from that question to then what that final answer looks like.

[00:20:55] [SPEAKER_02]: And much faster cheaper than they could go anywhere.

[00:20:58] [SPEAKER_04]: My last full-time role was with a small you know but to market research from specifically in the talent space and I was there when type to be was released and.

[00:21:10] [SPEAKER_04]: We just started playing around with it even in that form which was about 3.5 I think of ChetchipiT and you could tell immediately that.

[00:21:19] [SPEAKER_04]: I don't know I guess all the other market research analysts kind of looked around like.

[00:21:24] [SPEAKER_04]: I'm back in our resumes together because when I have to do not going to need to grow the organization by adding people but they could certainly probably take on more with less as everyone always wants to do but now this technology.

[00:21:38] [SPEAKER_04]: It seems to really be able to do it so I guess I'm curious like I know you have a lot of clients but like our a lot of them like bigger companies that don't have the.

[00:21:49] [SPEAKER_04]: In depth investment and their own market research or is it more you know start up trying to enter a new market as it's just you know is it.

[00:22:08] [SPEAKER_02]: From the solo printers to like the to the SMB market overall though why we could build up that expertise.

[00:22:17] [SPEAKER_02]: It takes a lot to both build up the human expertise to then translate into the tech to actually see you need this combination of both in order to make something like this happen.

[00:22:28] [SPEAKER_02]: In a way that then is economic feasible to actually go and invest in a built solution versus a buy solution however they come to us.

[00:22:36] [SPEAKER_02]: Anific to it as any.

[00:22:37] [SPEAKER_02]: If you want to add to that.

[00:22:39] [SPEAKER_03]: Well, I mean going back to your point as well Bob when you think about the potential replaceability of humans here.

[00:22:45] [SPEAKER_03]: I think in general any of our clients any size and a sage they're coming to us for pretty much one of three reasons they don't do research or they don't ask questions so our nation's point before it's too hard to get the answer.

[00:22:58] [SPEAKER_03]: There's don't ask that I go off and got or my CEO told me to do whatever so I do it.

[00:23:02] [SPEAKER_03]: So you're kind of going zero to one there that we we know we need to add some rigor and we need to be better there is one to you know whatever call it a hundred where we're doing it.

[00:23:12] [SPEAKER_03]: But in our research with our own audience show this as well we're not doing it efficiently or not doing it a cost effectively it's too much time having to verify things you sit down and I'm curious for your perspective Bob right you pull out GPT.

[00:23:24] [SPEAKER_03]: I have GPT caught and Gemini open and I have to go across all of them and my answer is still not completed like I still have to verify so none of that there's things that were missed and you know that's not like we're not there yet not to say that in and you're we won't be but still there's this element of like I'm doing it but I want to do more of it.

[00:23:42] [SPEAKER_03]: I also want to do it higher fidelity and higher quality and then there's the element of great now I've got the thing on my desk.

[00:23:48] [SPEAKER_03]: So what now what then what right like taking it not just what is the answer to the question what do I make a decision based off of what do I recommend to my stakeholders and then how do we actually start thinking.

[00:23:58] [SPEAKER_03]: Ten steps ahead of what the impact of this decision what kind of competitive moves do we make then so we stop playing checkers and that's the world where.

[00:24:07] [SPEAKER_03]: You know it's almost a stop a couple of research away I don't know what strategic insights is strategy go away definitely not so we might have to like.

[00:24:16] [SPEAKER_03]: Brush art farmers skillsets but our resume and getting laid off it's a whole different piece right such as the role that we play in helping action on and bring insights into action and revenue and all that good stuff kind of seems like there's plenty of use cases for having a single.

[00:24:32] [SPEAKER_04]: You know interface have a conversation with one you know super agent or however you want to define it and then they act as sort of a.

[00:24:42] [SPEAKER_04]: The general contractor to say okay now I know what you want and clarify that if necessary but now I'm going to go and go find myself contractors to execute specific.

[00:24:52] [SPEAKER_04]: You know sort of specialized tasks that part of what you guys are going.

[00:24:58] [SPEAKER_02]: 100% right I think that you know it's hard to say who's going to win the UI of it right like we have a very light UI that works to get products in which I think will you know be good enough for the you know the foreseeable future until there's like.

[00:25:13] [SPEAKER_02]: future player whenever Siri dukes it out with Google, dukes it out with Amazon all of those big those big guys will really focus on like what does distribution look like.

[00:25:23] [SPEAKER_02]: And then from there I think it really comes down to the quality of the agents themselves I imagine what's likely going to happen with all these big players is that they're going to now say okay great we need agents that do this thing really well.

[00:25:34] [SPEAKER_02]: Well who's the best company that does this thing really well like who are the best agents at the web and I think like even as we think about it in the micro case for us like we're making partnerships with our our enterprise partners and previously sold a research but now we're looking at okay if they're building their own internal LL solutions.

[00:25:53] [SPEAKER_02]: How do you actually just make our agents available of unAPI call right and I think that we'll start to see more and more of exactly that where you have the big the big players really do get out on the UI mass distribution site.

[00:26:06] [SPEAKER_02]: And then you'll have the applications that are really focused on just extreme quality and their niche get really good at that thing in focus.

[00:26:18] [SPEAKER_03]: Right okay.

[00:26:19] [SPEAKER_03]: And then I really like we think of it as sort of like a chef right where you've got all these spices on the spice rack or all these different elements and ingredients and it goes back to your point about mode as well.

[00:26:29] [SPEAKER_03]: The point solution where you can go to our nation's point to a certain company because they are the best at a certain thing and you use their agent for a certain thing.

[00:26:40] [SPEAKER_03]: It'll be interesting to see how things fall out our hypothesis that at least where we're spending our time.

[00:26:45] [SPEAKER_03]: We have world-class agents that certain things that will go in mind this source or you know pull this type of information and you put 10 event together.

[00:26:53] [SPEAKER_03]: And no other company is offering that combination of 10 you're either going to have to go pay 10 different companies and add it all up or you're going to have to supplement that with your own bandwidth and energy and cost whatever your own internal talent.

[00:27:06] [SPEAKER_03]: So there's an interesting kind of unknown to us we see things fall there but we see that in every industry always right centralization and decentralization but when you think about the formidable nature of like what our clients are trying to do is.

[00:27:19] [SPEAKER_03]: Are they trying to just do you know a quick scan of a competitor now they're doing that for a reason and the more that we can help them get to wherever they're trying to go quicker through a really strong.

[00:27:30] [SPEAKER_03]: Stack of you know contractors or spices on the spice rack or whatever more valuable we are to them because them more effective they can be in their role.

[00:27:38] [SPEAKER_04]: I do like the cooking analogies right because there's a bunch of different ways to think about it as your prompts might be you know recipes and you've got to.

[00:27:49] [SPEAKER_04]: You know not everyone who uses the same ingredients is it's gun day outputs not going to be the same we've seen that in countless you know cooking shows where people just.

[00:27:56] [SPEAKER_04]: You know it was disastrous and yet they had the same exact kitchen and the same.

[00:28:01] [SPEAKER_04]: Of and then same ingredients and everything it didn't turn out quite the same I guess the other thing that the name you think about is like this whole concept of part of future work is around thinking about how jobs sort of more fun change and.

[00:28:14] [SPEAKER_04]: You know if if AI takes more tasks or more tasks get automated at some point you reach a threshold where it says well this person's job is now fundamentally.

[00:28:24] [SPEAKER_04]: Change so let's argument sick let's just say it's you know 50% will now you have to redefine this scope of these different roles and I just again think back to some of the market research that we did at IBM I had it I had a whole team.

[00:28:42] [SPEAKER_04]: Of people who did the actual research and then brought me back to results and we would sort of iterate back and forth until it got to.

[00:28:53] [SPEAKER_04]: Something actually insightful that I could go take to my executive stakeholder and say these are some insights that we have and here some potential you know actions that you could take us as a result tweak this you know.

[00:29:07] [SPEAKER_04]: The shift to shift this here shift this budget here you know things like that I mean there's just recommendations and have to.

[00:29:13] [SPEAKER_04]: Take the advice but just the whole back and forth between me and another you know a whole team of people in another.

[00:29:20] [SPEAKER_04]: You know time zone and in another continent and that back and forth you know you've you basically applied AI to do a lot of that work but.

[00:29:30] [SPEAKER_04]: I will say those were very bright people capable of doing a lot more so I do think some of these things lead to real conversations about what do we do but these people and how can we it's either how can we.

[00:29:47] [SPEAKER_04]: Reinvest in them and and find you know new roles for them as the company grows or do we take more of a sort of myopic you know take the win kind of approach and let some good people go.

[00:30:03] [SPEAKER_04]: Unfortunately I know that's not an easy.

[00:30:06] [SPEAKER_04]: Business decision that some would not take very lightly but still doesn't make it any easier.

[00:30:12] [SPEAKER_02]: That's something that we've molded over for you know and gone in a few different you know directions over the past couple years.

[00:30:19] [SPEAKER_02]: I think maybe the most salient one that that comes like you'll never regret investing more and people and then people will show whether they're there for it.

[00:30:30] [SPEAKER_02]: But the people that then show that they are there for it they more than offset those that show that they aren't and that you invested in someone that probably didn't work out and then you have to have the tough conversations after.

[00:30:44] [SPEAKER_02]: Because of people that you did invest in their super motivated and they're super clear to have this historical context that they're extremely adaptable for this now new world where adaptability is actually the key thing to select for.

[00:31:01] [SPEAKER_02]: I've witnessed first hand that like so many of our analysts have just really taken that next step or maybe a few steps up and it's just beautiful to see it's beautiful to see it.

[00:31:11] [SPEAKER_02]: That makes me very excited whenever new updates come out not for oh my god now we have to understand how do we bring this workforce along.

[00:31:20] [SPEAKER_02]: I get to know say oh my god I get to let my workforce tell me how do we get brought along how do we actually now leverage this thing which is just incredibly valuable.

[00:31:31] [SPEAKER_04]: Yeah, I love to hear that just to jump ahead a little bit I mean when I think about a IQ and AI literacy and skills and readiness mindset is critical right you really need to constantly look for opportunities for growth right and so.

[00:31:50] [SPEAKER_04]: That may not be a linear path I mean there's plenty of folks including myself who have had a non traditional you know career path there's plenty of rewarding things that you can do and there's plenty of transferable skills that you may not even realize.

[00:32:02] [SPEAKER_04]: That you have that you could take on right so you don't have to be replaced you can choose not to be replaced by technology.

[00:32:13] [SPEAKER_04]: But you've got to put in the effort and you've got to have the right mindset that says I can learn I can grow and to your point I can adapt.

[00:32:23] [SPEAKER_03]: You can have you know an open mind and a mindset and a curiosity I think it's worth calling out there's a very real.

[00:32:31] [SPEAKER_03]: There's a very fair year with the complexity even what we were talking about before that companies are either making an operational risk or investment to say we're going to build on this model.

[00:32:41] [SPEAKER_03]: And that means we're writing on the hotels of this thing like if it changes we got to figure out how to stay up to date and you know that's the strategy we're betting on now.

[00:32:49] [SPEAKER_03]: I think a lot of leaders that I've spoken with and we try and understand how you know where they stand on all of these AI components and you know to the care how they are.

[00:32:57] [SPEAKER_03]: I'm just telling you that there's a real barrier of like I'm interested I know it's relevant.

[00:33:02] [SPEAKER_03]: I know I need to learn how to use these things but where do I start or okay I've not my workflow now there's 17 different free AI tools that help me write content better.

[00:33:12] [SPEAKER_03]: I'm going to now spend a week investigate all these tools learn one maybe is dead in a week there's there's I just want to acknowledge anyone listening that there's there's a weird limbo we're in right now where.

[00:33:23] [SPEAKER_03]: You kind of just got to throw yourself into it and LLMs at least are the simplest barrier to entry to you know cross over but there isn't element of getting yourself I talk a lot about even for myself right I'm in a company we're doing this and I know there's a lot more efficiency I can bring to my own workflow but to me there's a perceived barrier and there's behavior.

[00:33:41] [SPEAKER_03]: The mindset of their behavior is not very at the experience of components because maybe I'm just a CB with like wanting to pick the right tool but it's like there's so many options that's not to navigate and there's so many free things and then there's GPT's and it's like how do I where do I begin it's a very real thing and if you're listening to their feeling feeling it.

[00:33:59] [SPEAKER_03]: At least they're not in denial about non-meaning this skill said to change but there is just a challenge is actually acting on it to empathize.

[00:34:08] [SPEAKER_04]: You know if you're at a company that's starting to adopt it obviously you should probably be a good.

[00:34:14] [SPEAKER_04]: Good for racism and try to if they try the tools that they have you know sanctioned to give you but I know people are looking you know maybe on that they've already been playing around with things and their personal lives and things like that but no I definitely appreciate that situation Victoria I mean I.

[00:34:32] [SPEAKER_04]: Try to play around with a lot of tools I mean I would be a hit a great if I didn't play around with quite a few but it is a lot right because all the tools that you've got you were already using now say oh it's you know now with AI right like bolted on and you're like well.

[00:34:49] [SPEAKER_04]: How do I how much do I trust this I mean yeah I've been using this tool and I'd like to continue using it but you know I am going to kick the tires and I am going to see if these new features are reliable but we can't.

[00:35:03] [SPEAKER_04]: I just have to apply our own you know critical you know thinking and not you know it's not a calculator at least for most of these general.

[00:35:11] [SPEAKER_04]: You know LLM power tools but so you yeah you got to figure out what's going to work in your workflow but I think you know to some of the points raised earlier sometimes they have to work together in order to really make.

[00:35:24] [SPEAKER_04]: The whole process you know stream line right because if you have an address the workflow and the end to end.

[00:35:33] [SPEAKER_04]: Process then it's like individual productivity when you're working as part of a team right like if you if you have one cog spinning at 10x and the others are at their original speed something's going to break right so.

[00:35:47] [SPEAKER_04]: I do think organizations need to take you know a careful look at and certainly move beyond productivity individual productivity.

[00:35:56] [SPEAKER_04]: To start thinking about other metrics and start thinking about where the real value is generated right like when you if you do save.

[00:36:04] [SPEAKER_04]: You know, an ax amount of man hours or you know person hours in the week how might you be reinvesting.

[00:36:12] [SPEAKER_04]: You know some of that it goes back to what we talked about before about potential you know job displacement is right like that you have a choice you can reinvest some of the cost of avoidance cost savings.

[00:36:24] [SPEAKER_04]: In the people that got you to where you are and the people they're going to help you execute your strategy going forward.

[00:36:31] [SPEAKER_04]: And as well as in more technology and find more use cases around the company or you can just take.

[00:36:38] [SPEAKER_04]: You know take the win and and just say what we don't need those people anymore hopefully a lot more companies choose the former not the latter but you know we know not everyone's behavior is.

[00:36:50] [SPEAKER_04]: Outchewistic and human centric unfortunately. So yeah, so that's a big piece of it but one of the other thoughts I had in Victoria was you're talking about all these people just trying to figure out.

[00:37:03] [SPEAKER_04]: Where to go and they're going off on these goose chases looking you know trying different tools and more traditional research methods whatever.

[00:37:10] [SPEAKER_04]: You know a bigger company is you guys may have experienced this already but when you go through your annual planning exercise.

[00:37:18] [SPEAKER_04]: You know you're trying to encompass a lot across people process technology data and you broke up into all these sub teams and then everyone went off and did a little piece of the research.

[00:37:31] [SPEAKER_04]: And then it only came together back together in like a you know at the final stages to piece it all together.

[00:37:39] [SPEAKER_04]: Which I thought that's a lot of man hours that went in to putting this together and that could be streamlined by going to see you guys.

[00:37:49] [SPEAKER_04]: But also to not quit those insights together until the end your insights are going to be limited because you didn't quit more.

[00:38:02] [SPEAKER_04]: Potentially interconnected sort of pieces of the broader puzzle together sooner and say how how do these things impact each other.

[00:38:10] [SPEAKER_04]: So I think that is a miss taking these traditional approaches and it seems like in theory you know wonder would mitigate that risk.

[00:38:22] [SPEAKER_03]: Yeah definitely does there's a few pieces that I think clients have found to be especially valuable here.

[00:38:27] [SPEAKER_03]: There's generally most companies have an appreciation that starting with like what exists in the first place which is like set zero before what you just described Bob like what do we what if we just conducted research on it someone just do a transfer to their.

[00:38:41] [SPEAKER_03]: Data that we can mind from our product or whatever or is there stuff publicly available that even a competitor published alone even.

[00:38:49] [SPEAKER_03]: Gartner or BCG or whoever there's stuff out there that we can at least start somewhere and not at zero and then like find the same things out.

[00:38:59] [SPEAKER_03]: And we you know have what you said time and effort and money just to get to like basically what was already out there and then kind of within that.

[00:39:07] [SPEAKER_03]: There's the efficiency piece so that's where you know AI becomes really powerful because if you can get all of that together.

[00:39:13] [SPEAKER_03]: Use a language collecting the dots so that you can more quickly get to the connecting the dot stage and that's either.

[00:39:20] [SPEAKER_03]: You know by the imagine there was value in the different perspectives the different teams brought to whatever they were doing.

[00:39:27] [SPEAKER_03]: Was it necessary that they each had to roll up their sleeves and apply their manpower.

[00:39:31] [SPEAKER_03]: No, but they did what we did want all those perspectives to come together and create a stronger fiber for wherever the strategy went or the vision went.

[00:39:40] [SPEAKER_03]: But then you know you have to just kind of think in general with like where the insights that you're trying to accumulate is it makes sense to.

[00:39:48] [SPEAKER_03]: Have one of these big one off efforts we have seen clients be very quick to adopt you know the my food we advocate for is there's big projects sure you're going to have this planning effort.

[00:39:59] [SPEAKER_03]: There's ongoing monitoring that keeps you smart over time and then there's little questions that'll pop up over time and if all you ever do is sit down once a year or once quarter.

[00:40:07] [SPEAKER_03]: Really stick your hands and learn all the things that you can starting from zero instead of halfway there using all your manual labor instead of AI.

[00:40:17] [SPEAKER_03]: You're so much farther behind your competitors who are again using AI are staying always on their monitoring things.

[00:40:22] [SPEAKER_03]: They've got you know all these different caveats and insights coming to them and then they control half of it out in the trash that's fine, but at least they're getting exposure to them.

[00:40:30] [SPEAKER_03]: And you know able to click or pluck from a better crop of options or insights or data or signals instead of we just had this one week we all went heads down.

[00:40:41] [SPEAKER_03]: They came up with 50% of the season we came up with you know there's some I'm being dramatic obviously because a lot of room for efficiency but just again the curiosity of culture that culture curiosity becomes ways you're to act on when some of the tools get out of the way but also make it way easier for you to.

[00:40:57] [SPEAKER_03]: I'm not it.

[00:40:58] [SPEAKER_04]: Yeah, now culture of curiosity I love it when I think about the way that you're actually leveraging the humans in the process one of the things that came to mind was around.

[00:41:11] [SPEAKER_04]: It's not just that you need human you know intuition and someone to make sure the context is right and make sure you know it's tomorrow you know output especially when it's combined from different you know AI models but I think about cognitive diversity and making sure that you've also looked at potential.

[00:41:36] [SPEAKER_04]: You know bias in the output or even as you're putting a final report together just making sure that you've incorporated different perspectives whether that's you know academic background cultural background geographic.

[00:41:50] [SPEAKER_04]: Background things like that and which sometimes at least to date I don't know that.

[00:41:55] [SPEAKER_04]: Yeah, I'll put necessarily has those sensitivities.

[00:41:59] [SPEAKER_04]: So I just curious if any thoughts on that angle.

[00:42:02] [SPEAKER_02]: Yeah, I think when we this kind of goes back to the history of wonder we have straight away over you know over time from being that like subjective like here's the recommendation years all of these other things.

[00:42:17] [SPEAKER_02]: That tend to lead to the bias of the person actually doing the work and I've always lean towards here is objectively what we found at this source right if you believe this source then you also believe the data point that comes from source.

[00:42:31] [SPEAKER_02]: We have our own sort of master resource list that effectively says hey here great sources here bad sources here is sort of the scoring between them.

[00:42:39] [SPEAKER_02]: And we've taken a lot of those learnings and put that into our AI solution as well to figure out okay well now how do we effectively score sources in a way that that allows us to prioritize the the most trustworthy most unbiased.

[00:42:56] [SPEAKER_02]: Now is there bias regardless they're always right I think that it's very tough to eliminate every source of bias especially whenever any human touches the work.

[00:43:07] [SPEAKER_02]: We try to create the right framework to remove that such that it is as objective as possible and that whenever there's like you know sort of another way to think about is like what tools do you even use.

[00:43:20] [SPEAKER_02]: Right like when we think about cognitive cognitive diversity like the AI chooses its own tools based on you know sort of how however the AI decides to use the tools that it uses.

[00:43:30] [SPEAKER_02]: But the human may choose very different tools and a different human may choose very different so like how do we think about cognitive diversity in actual tool selection.

[00:43:39] [SPEAKER_02]: It's actually one of the topics that me and the team have talked about actually earlier this week was well the AI will choose those tools that doesn't mean that we need to choose those exact same tools on the human side it's actually better if we don't and we think about things.

[00:43:52] [SPEAKER_02]: Separately as so even more tackling this problem leveraging the AI worse.

[00:43:56] [SPEAKER_04]: Nice okay.

[00:43:59] [SPEAKER_04]: So we all are playing around with tools that work at home I'm curious if you guys have any favorite tools that you use outside of outside of work that you have found particularly.

[00:44:13] [SPEAKER_04]: Cool or insightful you know that actually afternoon the specific you know model or program just cares like how like use use case wise of how you're using it.

[00:44:26] [SPEAKER_03]: I mean I'll say that my use of the tools is as much what I said before knowing I need to learn how to use them as very academic right I'm like monitoring all of the different types of tools that could be used for desk research and using them to see where they're flawed relative to wonder.

[00:44:40] [SPEAKER_03]: So a lot of it is like sometimes it's research or if I'm investigating again different tools we might want to use for XYZ or audience insight tools so you know my marketing and content strategy or whatever.

[00:44:53] [SPEAKER_03]: I'm working across multiple at a time usually just to see how they compare and sort of the front end version of what I need to describe it where I started in, you know this model of that versus this model that I'm thinking of like the UI here feels better and still it's here feel better.

[00:45:10] [SPEAKER_03]: So yeah that's pretty much what it's been for me other than that a lot of what I've seen with the tools that we already use is they will slap on an AI you know looks like on a pig element where like do I need notion to summarize the thing that I purposely just pull it out of a point to make it not a water down sentence no but I'm pretty sure.

[00:45:33] [SPEAKER_02]: That's what I started. Yeah cool and they should like yeah I think on my and one of the pet projects that that I'm working on is a bedtime podcast in Spanish so like one that we can actually use AI to say hey this is like this like story and specials at learning Spanish right now and I want to see if there's any way that I can incorporate that into my routine.

[00:45:59] [SPEAKER_02]: So just partner chat Tpt to create different sections in chapter speaking imagine right I can't create like a whole non right because the token context is just way too short for something back complex.

[00:46:11] [SPEAKER_02]: But especially with the newest models right the ability to go a little bit more multi-modal I think that there's a lot of fun opportunities there on the personal side.

[00:46:19] [SPEAKER_02]: The professional side of each chat GPT multiple times a week not just for research I love having chat GPT there is just a sounding board.

[00:46:28] [SPEAKER_02]: This ways to sort of just tighten up any copy that if I'm just like I either in lazy to fix this or I'm like I'd like be my head against the wall and like figuring out how do I like make this as concise as possible then then chat GPT sort of sort of like go to there.

[00:46:43] [SPEAKER_02]: And then on the image side right I think that there's a lot of really cool stuff that Dali does is also a lot of frustrating stuff honestly that Dali does like the whole editing process with Dali.

[00:46:54] [SPEAKER_02]: You've gotten me like really close to it but because you haven't gotten all the way there I actually have to just review this whole thing but it's fun it's really cool to see what it would it can do.

[00:47:05] [SPEAKER_04]: Yeah I had the same frustrating experience well it all really stemmed from trying to get it to include words in the image so I suppose it would get better and might have to take a different approach to get better but yeah it was so close.

[00:47:21] [SPEAKER_04]: It was like 98% and it's like nope can't just misspelled birds right and I don't know how many times I can try to explain it just put these letters one right after the other.

[00:47:34] [SPEAKER_04]: So yeah no it's um they will get better I mean I'm not creative so I definitely went in there to look at you know podcast logos and company logos and stuff like that to see what it could do but yeah I got a little over complex and then.

[00:47:48] [SPEAKER_04]: And then everyone else was doing it so I was like okay maybe I don't want to just hack something together it looks like everyone else is cool colorful images so yeah it'll you know definitely get better but.

[00:48:01] [SPEAKER_04]: I do have a appreciation for people that are really creative and designers or whatever because they do great work and it's worth it.

[00:48:12] [SPEAKER_04]: I know we kind of hit on this a little bit before but I just thought you know in closing if you had any other thoughts around how people can get started if they haven't gotten started to just be comfortable and.

[00:48:25] [SPEAKER_04]: I'm going to let her it when it comes to AI like what do you well advice would you give for someone that wanted to elevate their AI to.

[00:48:33] [SPEAKER_03]: I own experience on channel saying the two pieces are first they the lowest lowest barrier to entry tool to try and just experiment using is your LLM so any of them all of them I mean I've used it for ideas for cooking.

[00:48:47] [SPEAKER_03]: Just whatever you know you have a question instead of going to Google just like get familiar and then you also learn.

[00:48:53] [SPEAKER_03]: Right this is challenge for a lot of people you learn where it's not helpful or you learn what you should have asked or you should have provided this context that's okay but at least you're like also thinking about thinking and there's value in that.

[00:49:04] [SPEAKER_03]: And then the other second pieces if there's some like very micro piece of your day that you feel is very monotonous or you you know think that there must be a tool for it just take on it and like you know figure out how you can.

[00:49:16] [SPEAKER_03]: Test and again there's an element of like being a servant of where it all sure and where it is or is it helping but I think there's just by it all little piece and that's the best the best first step versus feeling like I have to redo my whole workflow and my resume while I'm at it.

[00:49:30] [SPEAKER_02]: 100% of all that I think on the professional side is still treated like an intern you know you might have to redo a lot of that work but it's an intern that could grow into some into something that is actually a really well contributing knowledge worker on your team and I think going into it with that level of like openness and kindness to it also like then lets you not have as you know high of high expectations.

[00:49:59] [SPEAKER_02]: For but also it allows you just to play like you can give an intern to grow sort of that small task and they probably will be you know solid at it and then on the personal side I'd probably say yeah just anytime you have a question anytime you like want to.

[00:50:14] [SPEAKER_02]: I think that's a really good thing to talk to someone about this new thing that you're doing like especially having the new stuff it's like really good at.

[00:50:20] [SPEAKER_02]: I think those are really helpful times just like pop open chat GPT and just ask it something on the engineering side there's a lot of time at there's a ton of use cases.

[00:50:31] [SPEAKER_02]: I think we're seeing a lot of great use cases from Claude our engineers almost exclusively now use Claude instead of chat GPT it's helping I would make a lot of work to grow the engineers listening.

[00:50:43] [SPEAKER_02]: Hi, I recommend checking that out excellent.

[00:50:45] [SPEAKER_04]: Alright, well this was a great great insightful conversation thank you guys so much for being here Victoria I nish.

[00:50:54] [SPEAKER_04]: This was great I will make sure get links to you include those in the show notes for this episode and anyway yeah thank you again for being here.

[00:51:04] [SPEAKER_03]: I'm going to be here thanks again for the time don't read a retouch to us and think we're both on LinkedIn so if anybody has questions and want to chat through how they were thinking about things we're always happy to engage.

[00:51:14] [SPEAKER_04]: Perfect.

[00:51:15] [SPEAKER_04]: Alright guys thank you again thank you everyone for listening and we'll see you next time.