[Automated Transcript]

Deep: Hi there. I'm Deep Dhillon. Welcome to your AI Injection, the podcast where we discuss state of the art techniques and artificial intelligence with a focus on how these capabilities are used to transform organizations, making them more efficient, impactful, and successful.

So, hey there, everybody. I'm Deep, your host. And today we'll be talking about leveraging AI to build authentic online business communities and our guest today is Jason Tan. Jason founded Engage AI, a technology that automates interactions by remembering past engagements across various channels like LinkedIn with the goal of enhancing conversations in both virtual and real life settings. So Jason, thanks so much for coming on the show today.

Jason: Well, thanks for having me Deep. I'm so excited to share with you. I understand that you probably also use the large language model for some of the business, uh, that you're consulting, but really looking forward to share what we have got.

But more importantly, I actually think that this is a good, uh, note exchanging, uh, session as well. 

Deep: Yeah, yeah, no, it should be a really good conversation. So I guess the first question I would start with, um, is why do we want to build authentic business communities online? And, you know, like, why does that matter for the world?

And maybe if you can talk about why does something like leveraging AI to make insightful comments on relevant posts and stuff, like, how does that help in that pursuit? 

Jason: Absolutely. Great question, Deep. I actually have two angle and those are the equally the two reason that, uh, that, uh, that inspired me to do engage AI in the first place.

It probably a little bit too long. So if you bear with me, feel free to stop me if I go too much of the detail but Really, the two reasons that, that inspire me to build Engage AI is that really the, the first one is that your question of like, why does the business should be the authentic connection, right?

So similar like you, I am in the B2B uh, business. So I provide in before I do engage AI and part of my core business is providing data, AI, and consulting solution to the enterprise. And as you can imagine B2B sales take a really, really long time. And as we are not the Lloyd Accenture, PwC or Microsoft, it can be quite challenging because we don't have the brand positioning.

We don't have the hundreds of thousands of the BDM out there doing all the prospecting for us. So 

Deep: you focus on smaller businesses, if I understand you. 

Jason: I actually focus on the large enterprise. So, uh, which mean to some extent I am competing with a big brand like the Essential or Deloitte. And to get the attention of the senior VP or, uh, of the enterprise who would use data and AI to do various things to optimize their business and customer experience.

So, Equally, that means that we need to build relationships, but before building a relationship, though, one step, a pretty good step of building a relationship is breaking the ice and get to know them. Breaking the eyes and let them to get to know me. And that is not easy. It's also easier to say than done, right?

Yeah, particularly 

Deep: in the online world, you know, where in a conference, it's a little different. You might be eating breakfast or lunch together. Uh, you know, you get a chance to kind of meet each other and, uh, I mean, it's not a fully natural context, but it's a more natural context in some ways than online.

And so, you know, and in line, everybody's. Most people have their defenses up. You know, they they're a little bit more guarded and like so many folks are trying to sell so much stuff to so many people and all of us are getting bombarded 

Jason: with it. It's spot on, especially on LinkedIn, you probably have get a lot of these and I know that people who are listening to this would have equally the same experience that you have these people connecting to us and then.

start drip messaging us to say, hey, do you want to read my ebook? It will make you rich. Hey, do you want to book a time on my calendar that my solution is going to solve your problem? I think one thing that people are missing that is it is not a natural setting. And also The very fact that we are in the B2B, in the consulting world, the, the, the price that we sell, often if not a couple of thousand to tens of thousands or hundreds of thousands or even millions of dollars, nobody, nobody is going to get connected.

See the drip messaging and then say, dude, take my 30, 000. 

Deep: So you, you're, you don't think that traditional, at least the last few years, kind of drip messaging where you're getting small incremental messages and maybe you're getting people to like read some content and consume some content and maybe watch some videos.

You don't think that is sufficient, maybe or, 

Jason: or I, I would say. Consuming content would still work, but not in the way that I just described. If they are posting content, and if they find a way to get the content to the eyeball, uh, to get the eyeball of their target audience, that, that will work, okay? But I don't believe that it will work in the manner that I described where I just connected with you before I even talk with you and I just say deep read my case study.

Deep: You mean like the kind of like that abrupt cold outreach that that that won't work. Yeah, I think we're exactly because because you haven't yet earned the. You know the right to the other person's attention. Yeah, there's a bit of a bit of a process there like humans don't want to just hand out attention and we can't in the electronic world.

There's just too many people trying to get our attention. So there's a bit more effort required up front to earn the right to have that 10 seconds or even a minute or whatever. 

Jason: Exactly, exactly spot on. So, and if we, I'm glad that you used the example like how, for example, if we, if we would have met in the networking or if we have met those networking event or conference, where the setting is slightly more natural that we would have a small talk.

And exactly spot on on that one right now. Imagine Deep, you are the SVP of this insurance company. And you have just given a presentation talking about how you guys are using data and AI in solving some of the claim processing problem to speed up the process. Now if I just come back to you and say, Deep, that was a great presentation.

I really like what you were saying that how to use the LLM. where you can process the description provided by the customer in doing that. Um, and then I just follow up and say, how is the result so far? Naturally, I think that it, you would be happy that, uh, number one, not only I pay attention to what you say, but I also ask a solid question, a solid question in a way that not, is not too complex, but equally it's so that I listen to you.

Deep: Yeah, and in that context, it's sort of natural, right? Like someone is You know, at a conference presenting and everybody is sort of trying to learn what they're presenting and, you know, people queue up behind a microphone and they ask their question and, you know, you get your answer. So are you saying with engage AI, you're trying to sort of bring that concept into the online space.

And that is 

Jason: correct. That is 100 percent spot on. So the equivalent of that giving presentation is that you have a lot of these. Uh, really smart people who are writing content and sharing their knowledge on LinkedIn. The idea of that is the same thing that I will be queuing up and say, ask, ask that question.

Now here, here is the question for you. Imagine there are three people queuing up. The very first person come out and say deep to you, um, thank you for sharing. And the second person come up to give you a fire emoji. Uh huh. And a third person come in, which is me, and ask you that question, I said, Dave, I really love what you were sharing in this, using the LLM, uh, to speed up the claim processing.

How, how, how's the result? Among two of us, which one do you think that you will remember the most? 

Deep: I, I get the, the concept there, but wouldn't they have been talking about the results in the first place? Like, that's the part that I find Awkward in in, like, LinkedIn conversations is, you know, people talk and and they present stuff or they might post an article or whatever.

It takes a lot of time and energy to, like, consume that article, figure out what they're actually trying to say and come up with an insightful question. So I imagine that's the core of what you're up to here is, is trying to. Facilitate that with some, some generative, um, 

Jason: technology. So what engage AI does is effectively take the context of the presentation, take the context of the pose, processing it, and then also it come up a small comment.

In according to the tone that you have chose. Um, so in Gage AI we have a five default tone, where you can also create a custom tone, etc. All of those things. That exactly is the whole idea. And also exactly you just point out the very important point, right, is it takes a lot of time to read all of those things just to come up with a question.

And especially in the early, early, early stage of the prospecting, where I am just trying to break the ice. And if reading a post manually and to come up with something more than, uh, something unique that is more than to say thank you for sharing, it really takes somewhere around 6 to 10 minutes per touchpoint, right?

We know that on average, it takes more than 10 touchpoints. just to break the ice and just to get the attention. So 10 multiplied by 10 in in that breaking the ice would easily take a hundred minutes. 

Deep: But if I don't read the article and the machine reads the article and even though the machine, let's say the machine comes up with the perfect thing to say, perfect question, don't I still feel awkward because it's, I don't even know that it's a relevant question unless I at least read a summary of the article or is that part of it too, or you're building in a summary as well.

Jason: Well, the way that we certainly do great question, by the way, the way that we do is that we don't fully automated the entire thing, including posting the comment before they read it, nor do we encourage the user to post a comment before they read. But the whole idea of that is that you can at least quickly.

Uh, scan the article to get some idea about what that is. And then when the comment come out, you can add your personal touch. You can make sure that it is correct, uh, or it is wrong that you want to read, you want it to redraft again. But more importantly, you always have the opportunity to add your personal touch, whether you already know something about that industry in and out, or maybe you know about that person that you want to add that in.

So that is the whole idea. At least doing that, it will at least cut down your time from 10 minutes to easily, uh, just under a minute. If we 

Deep: block upstream a little bit, you know, I'll take myself an example because it helps me think through what you what you're offering is, you know, we have we have our target customers.

We know who we're trying to get a hold of. I imagine your system to be effective has to also identify conversations that are appropriate for somebody to jump into. And in addition to generating the response. Is that correct? 

Jason: Well, we are not at that level yet. I would say at this stage, really, you would be telling me as you are, scroll through the, the, uh, newsfeed of LinkedIn or maybe if you, especially this is a really good one, especially if you have like 30 or 50 people, uh, that you are targeting and you can actually save them into our web application and we will actually monitor and see.

Where whenever they post, we will notify you. But more importantly, right now, in terms of the context itself, we only do the context of the current post. We don't necessarily go and, uh, analyze the entire personality of those things, but we analyze the entire context of the current post. 

Deep: So the entry point is a list of users.

So I, I put in a list of users that I want, whose conversations or whose posts I want to engage with. Is that correct? 

Jason: That is correct. That is more of the advanced feature where I will also equally monitor whenever they post and notify you. And the reason why I come on that is, let's just say if you have 50 prospects and when you go on to LinkedIn, LinkedIn doesn't show you those 50 people.

All the time, or doesn't only show you those 50 people. They always show you whatever people that they want to show you, right? And that is a challenge. You, if you just go into LinkedIn and you navigate to each of those profiles and just to double check whether they post or not, 50 people will easily consume 20 minutes of you.

So what I am saying is deep, forget about spending, wasting all of those 20 minutes to check people are posting or not. Tell me exactly those 50 people are, I will notify you whenever they post and all you have to do is click the link, go to their post, and then use EngageAI to draft the comment for you and then engage with them.

So that is more of the advanced feature, but for some of the users, they are literally just scrolling the news feed and then they're commenting on the people that they are finding interesting. Got 

Deep: it. And then do you track the engagement stats, you know, post, uh, comment? So I, 

Jason: I assume you're I am. I'm glad you asked that question.

Um, so this is a pro feature that we, for, for pro members, uh, for the process description that we are testing. So we are actually testing in terms of like we are tracking the common. And also the post that they commented on. So, say for example, Deep, I am following you, I'm trying to prospect you, you have made about 10 posts for the last two weeks.

Whenever I comment on some of those posts. So let's say I comment on three of those 10. Now, Engage AI will start remembering The content of those three particular posts that I commented on, and I also the comment I make on those three particular posts, we are doing that we are testing that on the for the for the pro member, but I want to go a little bit deeper on that why we are doing that.

If you think back about the way that we communicate with each other you communicate with your family, or your longtime colleague, when you meet with them. You don't talk just about what is happening as of today. Often you would ask question to, to understand what is happening. So for example, let's say deep, if I meet you again, uh, three months later, I would ask you how is the podcast going and how is the business, the onyx.

Yeah. 

Deep: So you want to, you want some longterm memory in your system. 

Jason: So exactly, exactly. It's that sort of long term memory is what we use between human to deepen relationship where it help us to, to, to remember, to, to make the other party care. 

Deep: Have data? Have a hypothesis on some high value insights that if extracted automatically could transform your business? Not sure how to : proceed? Bounce your ideas off one of our data scientists with a free consult. Reach out at xyonix. com. You'll talk to an expert, not a salesperson.

Deep: So one question. So maybe, you know, a lot of our listeners like to know, like, how, how does it work a little bit behind the scenes? So let me see if I can, if I can guess, but before we kind of dig in, is it a plugin, like a browser plugin? Like what's the mechanics of the system? Is it an app that's using the LinkedIn API?

Jason: We technically, I, I would say we almost, uh, we are building an ecosystem. So there are three ways that, well, there are technically four ways that it will work. The first one is, uh, is a browser add on or browser plugin. So, and that is the most common way that, uh, 30, 000 users that are using us. Uh, around the world at the moment.

So once you install the Chrome extension or the Edge, uh, Microsoft Edge add on, as you go on to LinkedIn, as you are browsing the news feed, you go to the, uh, the post, you will see there is a little icon, uh, be injected in the comment field. And all you do is that you mouse over that particular Uh, icon, and then you will be able to choose the term.

So that is, before going into the exact process of how you work, which is your next question, I'll, I'll, I'll skip on that one. So that is the first, first one. And then the second one, in terms of our ecosystem is the, the web application, which is the one that I shared earlier, where if you have the 50 or 100 or 1000 people that you specifically want to target and focus at any given time, What you do is you tell me in terms of those particular, uh, profile and I will do the monitoring for you.

So that is the second part of it. And the third part of it is still related to the second part is a lot of time people would have already have a CRM. Whether it's HubSpot, PubDrive, uh, Zoho, and that is where we again come into the picture to do the integration. So if you are already tracking your prospect or your client in those area, um, once you install the, uh, Engage AI, you also do the integration.

situation to those CRM or maybe Zapier to other CRM. And what happened is that whenever you are adding the new prospect, and also if you take the checkbox to say, I want to monitor this guy. So the CRM would notify, engage AI to say, hey, here are the prospects that did want to monitor. So I would take in a list of those prospects, and I'll start monitoring.

And as soon as they start posting, I would equally send that sort of information back to, to the CRM. So what that equally means is if CRM is your way of working, then you can stay on the CRM and you can see equally all of those information being exchanged and you can get those information with the CRM.

So that is the third way that we are working in terms of our ecosystem. So you 

Deep: have a profile that you're building of So you've got the target that the user who's using the system. You've got their prospects for each of the prospects. You have a profile. The profile gets data from both linked in past conversation history and external CRM repos and then that.

All of that information now can be used, um, to contextualize, I'm presuming an LLM generated response in response to, uh, you know, your like a particular post that one that that prospect just made. And it can also be tailored. So you can also, so you can do something like prompt to chat GPT for Hey. I'm trying to get into a conversation by this person with these characteristics.

They just posted on this, uh, topic. Um, this is my background. This is what I want to do with that person. Um, can you generate a perfect comment here? And then maybe, maybe you're adding some training data behind the scenes. Is that about right? Yeah, that's 

Jason: about right. And you can actually do that in our custom prom.

And our custom pro will allow you to write all of those thing, uh, to, to to do exactly what you just described, where you provide more detail. A lot of the user probably use the default tones or the default prom that we have already, uh, written in that. However, they always, always can update the default tone, uh, the default pro as well.

I was just gonna come back to what I was trying to say earlier. The, my. Final final bit that I was going to say in our ecosystem is we actually also have a mobile phone app. So if you are iOS or Android user equally you can Use the mobile phone app to do to get those things working as well. It's not as elegant as the browser add on But at least it, it, it helps, especially if you're on the run.

So tell 

Deep: me a little bit about how well it's, um, working for customers and like what stage you're at. I mean, are like, what kind of engagement do you have for a given user? How often are they commenting in a per day? What were they commenting before? What are they commenting now? And then do you have some way of actually tracking whether they're overall, like your Holy Grail conversion, which I assume would be, you know, you got, I don't know what a 30 minute call or something with the prospect.

Tell me a bit about all of those stuff. How's it working? How do you quantify that? It's working. 

Jason: Absolutely. Absolutely. So, uh, we built and released in the early January and so far. Uh, organic growth itself. We have got about 30 plus thousand around the world. Uh, in that I have also numbers and hundreds and hundreds of the paying subscriber and also thousands of people who use the API option.

And that is the thing that I actually have been doing with the customer. I have been interviewing the customer and to really understand what exactly and also the value that the hypothesis that we have got, right? Is this thing working? Is this thing worthwhile for the money, subscription money that we are charging to people?

The answer that often I got from those people, including the, the customer who have cancelled because of the budgeting issue, is that It is certainly working because they can really tell the differences in terms of before and after using the Engage AI or even after they stop using the Engage AI, what are the numbers or what are the metrics that they are using to measure the effectiveness of Engage AI then?

Number one, how many conversations can they actually start? Okay, number two, how many profile visit. So profile visit is another way that people would use to decide when they want to start a conversation. So let's just say for example, if I come to comment on you for a number of time. It's very, very natural.

And that's the whole psychology of the Engage AI as well. If I come and comment on your post over and over and over and over again with some reasonable unique of the comment, it's very only naturally, unless you have your profile managed by someone else, but it's only naturally for you to look at. To look at that, right?

Who is this season 10 that often come and support me and engage with me? 

Deep: So you're looking at the, the curiosity pings on their profile. So. 

Jason: Exactly. Exactly. So you would respond to it. That's how we build rapport, but for some people who don't necessarily respond, they would actually come and look at your profile to say, who is this person who always engage with me?

Right, is a combination of those things that how people are using the site. Okay, here's the right time that we should start a conversation. So the effectiveness in terms of the metric that people are using is the profile we see. The rapport that they are building, the conversation that they are having, and more, more importantly, is how much and how many meeting and how quickly they can have a Zoom meeting with the prospect.

Deep: Do you have that built into your, your plugin somehow where you can where you can actually set up the meeting and know that you that you got the real lead there. 

Jason: Not at this stage, not at this stage. I think, ideally, maybe we would want to build all those things, uh, one day, but at the same time, I recognize my role and my role is to break the ice and help people to, to create a meeting, to get a meeting.

I think my role stopped there at this stage, which, and, and I would love, I would. Rather than let CRM take over, I am super focused. I am super focused in terms of our positioning, which is I help you to break the ice and my job stops there as of 

Deep: now. Yeah, I mean, I think the reason you might want to go there is because the selling argument is much stronger with the end user, because now you, you're not just saying like, hey, I can get you to talk to people, but you're saying, hey, I can get you 30 minute meetings with your qualified prospects, which is a much higher value.

So 

Jason: I'm 100 percent 100%. So I think naturally for us to get to the next level that up though, is that we, we, we would be looking to expand ourselves into the LinkedIn inbox. That is where we have them to start conversation, especially with our second brain. Feature because second grade will start allowing them to remember all the conversation where they can find some of those historical information to help we get a conversation.

I, I really think that we would be helping them to converse better and to deepen that relationship. I think about the way that we use the communication, especially in these B2B sales relationship is that. Often you start with the commenting for some people, then you start in the inbox. And as soon as you determine that, okay, I really want to talk, I do want to continue talking with this person, very often they will migrate themselves out of the LinkedIn inbox to the email.

Because LinkedIn inbox is a mess. So. Once you go into the inbox, how do you bring a lot of those memory to the email inbox as you are trying to write the next email to get the meeting. So that is how we about the, the flow of this whole conversation, but the highlight that is that our focus. Exactly as you say is how do we help them to get to that 30 minutes meeting, but true to conversation.

I don't necessarily want to replace the CRM. I think we would be a good compliment to the CRM. Our focus is still to get people 30 minutes, but we don't necessarily take over the role of the CRM. So tell me, 

Deep: like, I understand how, if I'm trying to get ahold of somebody, how this could help me by, um, You know, helping me engage with those folks, but how does it contribute to an overall healthier online business community world?

How does the other party benefited? So, you know, the person who, you know, just posted or, you know, and is suddenly getting comments that are bot generated, how are they, how are they benefiting from 

Jason: that? Absolutely. That is a great question. I think that is a question that people were raising as well. Right.

So, in terms of the user behavior, certainly, I have no doubt that we are seeing that some of the user, they literally just use engage AI. To draft a comment and then before reading it, uh, or before adding any personal touch, they just post the button. I think that behavior equally is very much the same thing, like how people are using the generative AI to write article and then blindly just plot that as a blog article and then hope that the traffic will increase.

Now, the truth is, there will always, always be a group and a percentage of those people. And yet, uh, business owner or the prospect who read those sort of comments, they don't necessarily like it, right? They would not like it. And they are not going to have any relationship. They are not going to be bothered to, to pay any attention to those 

Deep: people.

I think you can summarize the macro argument a little bit like this. Um, like imagine a real life conference, like two real conferences, one real life conference. Um, people present the audience. Um, You know, let's say has one minute or something like a brief amount of time to ask questions. And then you move on to the next presenter and the next presenter and the next presenter.

That's like Scenario A. Scenario B is they present, it opens up to the audience. Maybe it's even unconference like where people can break out and like really get into deeper discussions. And there's a lot of back and forth between the original presenter, those listening, and those listening and each other.

And I think almost most of us have a strong intuition that the latter case is more valuable that just that there's just a lot more engagement. It's healthier engagement. So then. If you go back to the bot, I think your, your points well taken that, you know, there's always going to be people who just blindly take what the bot says and say it.

But to the extent that it's like a high quality assistance tool to like encourage that, get over that blank slate problem. So people know what to say, what to ask. Then all of a sudden, you know, in theory, we're getting closer to that second conference scenario. Like scenario where people are highly engaged, they can break out and self organize and they can go back and forth with richer conversations that are real and not just you trying to prospect and the person trying to avoid being prospected to, 

Jason: I think that is 100 percent correct.

You summarize and you describe it so, so much better than I do. Exactly spot on. And we always, always encourage to have that human in the loop where people should be adding that personal touch. That is a hundred percent smart

indeed. 

Deep: Yeah. I mean, I think that that's a, that analogy of the real life conference. Cause we've all like my favorite conferences have always been the unconference. Are you familiar with that format? No, 

Jason: I'm 

Deep: not. Can you tell me? Yeah, so the unconference format is a bunch of people show up, there's a general theme, and then there's just a room with giant pieces of paper.

And everybody who attends just walks in onto the paper and writes whatever it is they want to talk about. Or whatever, whatever they're interested in, this is in the old physical world. And then there's like a step where, I can't remember the details, but there's a step where, where people who want to talk about X kind of, because people tend to cluster, you know, like people who want to talk about, you know, topic A, topic B, topic C, then they tend to like get around each other.

And then there's like another step where they have specific questions that they ask underneath that topic. And then they basically just kind of all interact. And it's like a conference without. Featured speakers. It's, um, but where everyone gets to talk about what they want to talk about and it's, uh, sort of self organizing.

So it shares like many of the properties of how conversations emerge on the web or in LinkedIn world. Um, but at the same time, you know, it's because it's in real life. Because you're at a conference, then, you know, the ice breakers are there. The relationship building happens, you know, all that stuff is more natural than it is typically on the internet.

Jason: Exactly. Now, if we just zoom in into those, when you were talking to each other, do you notice, do you remember that there will always be one or two people who are often louder than most of the people? Do you also notice that there are people who are more quiet than the others? But does that equally mean that the one who is louder knows the best or knows the most?

Does it mean that the people who were quiet does not know anything? 

Deep: Yeah, I think the, the more extreme introverts in the online case, you know, may or may not share overtly, but, but they're there, and they can self organize out into a different group, or I think what you're speaking to is some tooling that like brings those quieter voices out.

Is that right? And if so, what kind of tooling are you thinking of? 

Jason: Exactly. And that is the whole idea of the Engage AI. That's the second angle of the Engage AI as well. So the first one that where I talk about Engage AI could be useful for the SMB owner to really save time in prospecting. But equally, Engage AI could be really useful for the technical.

manager for the technical founder for the technical business owner who often come from the technical background. Now these people, they are really, really good in terms of what they do technically, whether it is, uh, graphic design, data science, AI, uh, code or whatever, uh, marketing that is. But often these technical people, they are not as good in terms of the marketing themselves or selling like those who are naturally born good in marketing, in selling, or, or speaking.

That is exactly where I come from, where I, this is the people that I think engage AI could be really useful. How we can provide them a tool, uh, AI assistant to help them. To communicate better. I'm not saying that they can't communicate. What I'm saying, though, is that it is a natural personality for them is that either they prefer to listen more or they sometimes they just need a bit of assistance or they give them the tooling.

To help them to market themselves, to sell better, to communicate better, like those marketing or the salespeople who are naturally good at those things. 

Deep: So beyond assistance in, in message generation and maybe identification of, you know, who to talk to or prospect, uh, assistance, like, what do you think the role is of AI in helping build these online communities in the big 

Jason: sense?

I think it would help people to write that and also convey their point. better. I, I probably would use one example, um, for maybe some, someone like myself where English is second language. It took me years to practice, to learn, and also it took me years to, just to build up that sort of a confidence to feel like I'm not making fool of myself.

But I tell you what, 20 years ago, you would not, you would imagine that I can't even bring myself to make a comment on social world in English. If I were to have to write in a second language, and sometimes we just think too much where we feel like the grammar, the vocab, the way that we write, the style, they are not naturally similar to the English speaker, and...

Deep: Yeah, yeah. I mean, I have the same problem when I'm communicating in other languages. I mean, that makes me think of, um, you know, we're seeing... Increasingly powerful uses of LLMs for applying like rubrics, you know, so to like, think of it as a high school English lit teacher, like providing feedback on an essay, but you're getting that feedback down to the comment level.

So it's not necessarily only like writing a replacement version of for you, but helping you think through the kinds of things that might make this a better comment, like, Hey, you're being abrasive. This could be perceived as being abrasive. Maybe if you start with a comment that shows that you're listening to the other person, maybe some active listening, grading, that kind of thing, is that kind of where you're going?

Jason: Exactly, exactly. That is where I was going as well. But also sometimes it would help people like myself to be more elaborative. So, a lot of the technical people, they tend to be one liner. Rather than being descriptive as well, right? And LLM for the online community, for the people like myself, would actually help me to be a little bit more elaborative rather than just one liner.

Rather than just go to the point and, and say that out and thought that, that, that is all I have to say. We embrace degenerative AI in a healthy way. Seeing it as an assistant rather than a replacement, it would really help to build that positive and healthy online. So that's, that's 

Deep: an interesting idea. I want to take it for a little bit away from the more controlled landscape of the LinkedIn community.

It's like LinkedIn or, you know, business folks, it's a lot more respectful in its tone and dialogue. But as you know, places like Twitter, you know, and, you know, and the wilder parts of the internet. They're a complete mess aggressive and people are like intentionally just hostile to each other and it produces a lot higher engagement as a result, which is sad, but true.

Do you think there is a lesson here for for those communities to increase the quality of the engagement? And specifically, is there a business model where it's in Twitter's interest to not have everyone just throwing tomatoes at each other and, you know, flame throwing? Is there some world where we can all just basically leverage AI to be more civil?

The way that I think the communities generally are on LinkedIn. LinkedIn's a very civil 

Jason: place. I. I think that is a really, really deep question and it, I don't know if I can ans I don't know if I have the answer for that. My initial thought of what is just describing though, I suspect that maybe to some extent the L R M can be helpful.

For the people who are less descriptive, who are less communicative, or who are less, who are not as good as, uh, who are not good with the word to, to be better communicate themselves and help them to convey their point and try to bring a bit more civil into those communities. But equally, on the other hand, LLM would amplify those noises.

LLM will also amplify a lot of those Louder noise to make it even louder. So I suppose, I suspect that the, I suspect that each of the community would have their own personality, would have their own type of Of online behavior and LLM probably will amplify that more than balancing. It is my gut feeling.

And that's why I think a lot of people are concerned at how people are going to use for the misinformation. I think that is probably a bigger question that we all have to come to address it as a society. But I suspect that that is what it will continue to happen. 

Deep: Yeah, I don't know if I agree with that.

Like, I think, um, When I think about what's happened in the in our information consumption arena the last 25 years, you know, with the rise of filter bubbles, people tend to see and hear and self select into the communities. Exactly. Well, that part I agree with that. They scream at you all the way that they if they want to be screamers, they go to screamer communities.

That is correct. But I think with AI, we have with these machine learning systems, we have, um, what I think is Kind of interest like possibly good, possibly terrible, but you have the ability to like if I'm a jerk to have this thing automatically tone it down when I post on LinkedIn, and then maybe even amplify it in like, you know, some some rabbit online forum, and then you have the opposite case if I'm really, you know, nice and friendly.

Um, you know, where it can spice it up a little bit to go into a,

if I like people screaming and yelling at me all the time, then screamy, yelly entertainment world. I mean, it feels like it has a very strong chance of just becoming dystopian, um, because everyone can like. Like what does it mean to live in a world where, you know, somebody's actually a complete asshole, but everyone else thinks they're nice because they're because all of their interactions are translated into a nice bubble.

And, you know, like, what does that even mean? Like they never get a chance to improve themselves to become a better person because they just scream and yell. And everybody, you know, like everybody knows that person that's been in an, you know, um, like a social or family structure where everyone's like, Oh, he's just like that.

Leave him alone. Let him just be his jerk self. That's not really better for that person or those immediate people who have to listen to it all the time. I don't know. Sometimes I wonder what happens when we get so good at tailoring to the individual that we tailor these tone, tonal tendencies as well.

Jason: Yeah, I think that is the really the world that I think that is the outcome that I'm looking forward to see as well. I don't know whether it will turn out good or bad, though, that the people that I am serving, the people that I'm helping. Where, who are not naturally a good salesperson. I do feel that how this tool or the vision that I have for them is to help them to communicate better because there are shit lot of those 

Deep: people.

Jason: Yeah, yeah, yeah. Shit. Lot of those people like myself. So this is probably the message for those. 

Deep: Well, I mean, I think that's a good point, right? Like there's plenty of. You know, founders or entrepreneurs or just people running, you know, smaller, medium sized businesses, maybe that have amazing ideas, amazing products, but for whatever reason, they can't sell at the level of somebody who's louder, more charismatic, all that kind of stuff.

It's sort of like a equalizer in some ways, at least your vision. I want to end with like a question, um, where let's say that everything you imagine happens, your assistance tool is amazing. Everybody uses it. Maybe it even gets better and you automatically identify conversations across like a ton of different platforms to jump into you automatically jump in fast forward five years for me.

How's the world different? How's it better paint for me? The vision of how I helped create a better world or a worse world. Whatever 

Jason: you think. Well, I hope there are two outcomes. Certainly for that. The first one is these very people that I serve, the SMB owner, the engineer like myself, they are going to have a chance to sell, sell, sell their product, or sell their service.

They is an equalizer. They can sell as good as those naturally good salesperson. That is really my number one outcome. Number two is, I hope people can have a better relationship because of the idea of the second brain being there with them for 24 7 and helping them, uh, to remember all the conversation and then provide them those data, provide them those information so that they can use, or they choose not to use, as they are communicating.

I always come back to this scene that I, I remember watching, uh, the Devils Wear Prada. where the, uh, Anna Hathaway was, uh, being the assistant for the lady, uh, who was her boss at the time that, uh, Anna Hathaway was literally just whispering to her boss, like, who exactly this person is, her favorite, when was the last time that you talked to her, what does she like to hear, and then the boss literally just used those information to help communicate.

That Is probably the, 

Deep: that's a funny, that's a funny movie to hone in on. I remember that the devils were what devil wears Prada. It is a capability that very well endowed organizations have had for, for a long, long time, right? All the way back to thousands of years ago, you know, somebody going in to speak to the king or the queen or something was prepped on what they like to hear, what they don't like to hear, you know, all this kind of stuff.

Final question for our folks. A lot of our listeners. You know their product managers or product builders, whether they're developers or executives or entrepreneurs. Um, but not all of them maybe know how to get started with like leveraging AI and integrating it into their product features. Is there any like advice that you might give, uh, you know, those folks in terms of how to figure out what AI can do for their product and their users and how they should get started.

Jason: Absolutely. I really would say is the first thing first is how do you embed the entire AI into the product, or how do you embed that entire AI into the customer experience right within their workflow, because that is where your AI only become extraordinarily and extremely valuable. So what exactly what I mean by that is like, imagine.

With the Engage AI before we, uh, engage ai, um, people can actually copy and paste the content to go to ChatGPT, write the prom and ask chat g b t to provide the answer. And then you copy and paste back to comment, and then you update it and change it, right? Yeah. Now, in theory chat, g B. Really do that.

So that is not really no room for engaging. But what I'm doing right now doing is we literally say people find clay and writing some pro. So the point that I'm trying to make a AI is good, but it is only as valuable as it would be, and its value would only amplify when you are embedding the AI into the existing workflow, into the existing experience, into the existing product where users usually just use 

Deep: it.

Yeah, no, I agree. I think a lot of times folks think that they have to inevitably like completely revamp their product, which can be the case sometimes depending on how deeply they're integrating. But a lot of times it can be about finding efficiencies and, uh, more encapsulated, uh, improvement vectors. So, well, again, thanks so much for 

Jason: coming on.

Thank you so much for having me. I really enjoy it. I think you asked really, really good question that, that wrap around the, the entire topic in a very fun way, but also in a way that. Probably the things of the question that I am yet to think about. 

Deep: That's all for this episode. I'm Deep Dhillon, your host, saying Check back soon for your next AI injection. In the meantime, if you need help injecting AI into your business, reach out to us at xyonix.com. That's X Y O N I X.com. Whether it's text, audio, video, or other business data. We help all kinds of organizations like yours automatically find and operationalize transformative insights.