Speaker 0 Hey everyone, welcome to on product. I'm Paul Adams and let me today as always is des. Hey, paul. Hey.
Speaker 1 Off you go.
Speaker 0 Good. Alright. So today, we've a great topic. We're gonna talk about Ai and product strategy. We're gonna talk about what that means for people.
Speaker 0 People are in a range of different positions on this. We've seen what's possible from this kind of first wave of companies. And like any big technology where it's cloud or mobile. Now it's Ai. We've It's don't unclear people to start how it's all gone pan out and so on.
Speaker 0 I think when you look at the landscape today with people who are all in, they're like bet the farm, bet the company, and then you've got people who are still? A little unsure. Is this really a big deal? Is this just more ko aid? So silicon Valley Cool aid.
Speaker 0 So looks like a ride range there where do you think you're on it?
Speaker 1 I'm am definitely in bet the farm, bet the company, get better to Ko, go to your neighbors, that their firms. I think it's huge. I think it's... Like, I I understand that the cause for skepticism because it does seem to have conveniently arrived at a time when very silicon valley of investors were like sky for something needed to talk about. Whenever ever you have the experience, like that sort of the Ai stuff is delivering at about.
Speaker 1 It's quite clear that something massive this happening. And that it's like we're still in the sort of embryonic stage of seeing it like, you mentioned the dose to settle. It is like the first wave of those. Like we've we're starting to now see whole companies again, like, a series a or series b off the back of. You know, being an Ai native, applied company Like, mean, I I wanna talk about this, what I mean is, like, not, like say, open Ai or around, like, providing the actual Ai, but, like, people who are building whole, like workflow products that are entirely powered by Ai.
Speaker 1 As in if Open ai and that's all that didn't exist, company also wouldn't exist. Right? So it's like people are really leaning on as a platform. I think when you look at some of the capability out there, it's a straight high uncertainty for me to, like, the this whole industries, whole categories of soft will be abandoned.
Speaker 0 Yeah. Sometimes in tackling, you can talk about, like, extinction events.
Speaker 1 Mh.
Speaker 0 You know, like, mobile came along. Mobile first companies killed companies who weren't mobile first and couldn't adopt, prior to that cloud, same cloud first companies, Like you mentioned Ai first companies? Yeah. Do you think this is like an extinction of event type thing?
Speaker 1 I think in certain pockets certainly and certain a lot other pockets, I think if it's not an extinction event, it's because of it 1 new dynamic, which might be dot air. Because of a lot of the power actually sits on... Let's just say an open Ai server. Right? The power is kinda on top access to true an Api like, May summarize this 5000 word incident for me.
Speaker 1 And you ping that over to a third party and you get back the response. That is not the same thing as rebuilding your entire company to make it ios native. Yep. So as a result, it there will be areas of software where I think the incumbents actually make use of this again a lot of the value. So Think some areas will be extinction events, but I don't think it's a it's a white.
Speaker 1 It's not an asteroid or at least maybe it's a small asteroid or a small asteroid share in certain areas. It's not gonna take out at the whole industry. I think you're gonna see a lot of the big companies actually just get bigger.
Speaker 0 Yeah. Which else have order order other...
Speaker 1 Yeah with
Speaker 0 Mobile 2. Yeah. Like Google Facebook Event eventually did figure out had how to do it. They did. Yeah.
Speaker 1 That's right. Yeah. That's an interesting. Like. They figured out to do it quicker than the anyone could figure out like come how to be great.
Speaker 1 I'd say search. Like... And we'll come back to this idea of a ratio in a second. But learning objective c in deploying a objective c, powered or whatever ios paired sort of interface on a mobile phone, to an insanely powerful search engine. It turns out the hard bit of all that is sd the insanely powerful search.
Speaker 1 Yeah. Not... You know, I'm, like... So if that ratio is important of, like, how much new work do you have to do? How much of the incumbents work is still valid?
Speaker 1 Yeah. Effectively, Google's back end is still today extremely valid. But and the front end might change and they were seeing barred instead of the old sort of search box or whatever. But like, the, you know, turns out crawling the entire Internet is not something that, like, 2 Rand falling out of Yc you can do in the evening.
Speaker 0 Here. Yeah. Yeah. Yeah. But let's talk about both sides are that.
Speaker 0 So Yeah. There's like, all the kind of we some to call these, like table features.
Speaker 1 But it's
Speaker 0 like the core features of product needs in in a certain category.
Speaker 1 Mh.
Speaker 0 There's like new things. You know the new things that can do, new technologies enable new things. Let's start with, like, the new things that Ai can do. Right. So like you a whole list of things that make you bullish on it.
Speaker 1 Yeah. That's true. Like, if we go back to say, November 20 ninth, when we saw Chat Gp 3.5 or Turbo or whatever. If I remember various naming but became obvious or at least the first thing we've had seen it from that was, like, well, this thing isn't is very, very good at being conversational. It's very, very good to understanding humans and very, very good at replying back really well.
Speaker 1 It takes prompting really well. It takes instruction really well. It was very good at which is just basic text writing. So expand upon this, summarize that, re rephrase this, re tone that. It was also very, very good at a like, deduction or inference.
Speaker 1 So you can give it a complex scenario and ask it like, you know, like, if you say, like I don't know, somebody is struggling with the long term illness inside of the burning building, which is the bigger issue. Like, it'll like it'll work out answers those questions, like, which, you know, you know to human these sound insanely simple, but like, to get a machine actually she understand that is like, is un making inference and then suggest an action, is quite powerful Or, like, you know, given the state of this project based and all the, like, updates you've read, what do you think the most important issue is. Yeah. And it actually do a really good job at that. So the idea of, like you know, deduct inductive reasoning is is like pretty powerful there as well.
Speaker 1 And that's also far we just talking about like, the text domain. The other stuff we've seen... We saw some of this you slightly prior, which was the, like say Dolly and Dolly 2. Yeah. We're both prior to that.
Speaker 1 The ability believe to given a piece of text render and image. And that's getting insanely good right now The latest major journey stuff is just breathtaking. People often ask like, why is that useful? Well, like, there are loads of scenarios where people aren't creative, but they know what they want. So, like, I'd like to send this email and I'd like it to be sent and I'd like to be a l font on a dark textured background.
Speaker 1 Yep. And creative it can give you 27 versions that on the screen, you can buy this 1, but I wanna be more rich or more high end or more luxurious. So all of a sudden people who can't do art can do art. Right? Massive expansion to the accessible market of in this boring case like, we can email creators or whatever, you know?
Speaker 1 What like, being able to like, generate imagery is like not to be sniffed that I think a lot of these things get like like, ty by the 40 use case. Like, Xiaomi, you know a cheeseburger eating a planet and like, it does a really good job of that. But people are like well, never wanted that. Yeah. That's fine.
Speaker 1 But I guarantee, like, give me a really nice header background for my new website is gonna be a... Feature in squarespace or wix or something like that. You know? So like So that's imagery. We have voice.
Speaker 1 So this is been coming along which is both the ability to parse voice. So like, you know, pool pretty much real time audio transcription and it can generate voices as well. Like the latest breakthroughs there. So if you look at say S easier or is 1 called Play dot h t, there's a lot of them like where you can literally... You can give it like, mission and impossible shit.
Speaker 1 Like, you give it like, 90 seconds of you speaking, and it'll do a passing impression of you. For like a single sentence. Yep. Give it, like an hour of you speaking, And it starts to get... You can trick it or you can force it to make it clear it's definitely not Paul, but you could certainly get away.
Speaker 1 So generating voice, then generating video. So again Ci easier does this like a fake video avatar thing, like, you can like record you and some manner of view and it'll be able to like, basically, make it look like you're talking. But we're gonna be able to generate full on video in the same way we can generate full on like imagery. I I think the mistake I was making initially and a lot of folks make initially is like, they think right. Well that's ground.
Speaker 1 Sounds really important if we're working in Adobe I should be all over this. Think people don't realize. How much disk has a way to creep it here normal life. Like, that voice tech can be literally the same thing that, like, would imperative to future of messaging or comparative future of product interaction where you just talk to your product while you're driving or whatever. Yep.
Speaker 1 All of that that's now possible. And similarly like as I said, like, the imagery isn't just like, you know, hot dog eating planets or whatever. It can literally be like you design the tie background and res skin this product that I'm using right now to look prettier here. You know, all that sort of I could keep going with, like, order cool stuff that's that's like now possible. But when I look at the sort of collective weight of all of that potential.
Speaker 1 And I think about its applications to, like, specific software domains, to to creativity, obviously, to Ui to just how humans might interact to humans. What jobs can be automated, what what parts of jobs. We. I think, like, you couldn't, like, pressure me into being an Ai skeptic to at this point. I'm just like it's not possible.
Speaker 0 Like yeah.
Speaker 1 You know, it'd be like trying to push back the tide. Just... It's pretty obvious me massive transformations coming, and you're better off getting on the right side of them.
Speaker 0 Yeah. I mean, on there too. And even in some of the things you said there, like, imagery, for example, you know, the entire the entire industry of advertising. Would probably be turned up to them with. Absolutely.
Speaker 0 Certainly, if you work in any kind of creative agency or media agency. Mh. I know people work who work in Creative she's already using,
Speaker 1 it Yeah.
Speaker 0 Ai to generate all their work. You know or most of their work. Let's talk about the other side of it. So like even someone of met the startups you mentioned there, I haven't heard it before. It's like, it's just an explosion.
Speaker 0 Right? I think on again anyone could keep up with all the new types of things built on this new generation technology. Meanwhile, you have, like huge companies, hundreds of millions of dollars in revenue who've built a business over a decade, 2 decades, And I think we in the earliest come a little bit naive. You know, we were kinda coming in, like hot startup, like, taken on the incumbent. Yeah.
Speaker 0 You know, giant killing, tiger your
Speaker 1 salesforce Yeah.
Speaker 0 Like, you know, chip on your shoulder giant killer. Right? Yeah.
Speaker 1 And
Speaker 0 then you realize, oh, you know, an an area like reporting and so if you're like, oh, this is a big, deep. Thing yeah.
Speaker 1 These are big for reasons.
Speaker 0 Yeah. Years of, years of product development requires. Just to have the table 6 just to be there. How do you think companies think should think about that?
Speaker 1 I So Think you can look at this from both sides or like... If you're like a scrappy startup and you're picking an enemy? Like... So if you say when it's go after workday, what is the attack angle on workday? That Ai permits.
Speaker 1 Well, you look at all the capabilities that we have. Like, you could try and generate performance reviews and try and pass that sort of stuff. But ultimately, when you... You know, let's say you find a few examples where you can sprinkle and dart in, like bits of Ai magic to simplify existing workflows. Think I even use to use Workday would have to admit.
Speaker 1 I think I even gives a shit a bit to come complexity to workflows inside that company. That's not their like, Roi. That's not the reason people buy Workday. And in fact, reason people by workday there is I think because it's the largest sort of, like I guess, e or p for humans so you could imagine and they have a massive enterprise sales team and they've built a huge brand that's all about, like, we are, like, the final boss you know when it comes to like Systems and that's what they care about.
Speaker 0 And I think an incredible bigger. Like exactly almost infinite bigger.
Speaker 1 Like workday doesn't... I even get assumed A these people are humans. Yeah. You know what I mean Interesting you map things to things and the 1 role is call this 1 roll called So the question, if you were to rebuild all this in an era of Ai, what would change. If people are buying stream config ability.
Speaker 1 It's not obvious to me that the attack angle is going... Is basically there. Like, I think people are really buying a glorified like, almost W to a database where they can connect ting to Ting by manager relationship and say, Ting has report and ting has home address and ting has salary. And on all that, you know, I have to say, like, they'll have you hiring, like, full time work workday they like, you know, at people. I don't think any of that really, like changes massively in the near term.
Speaker 1 So I think you could you could have a far more beautiful workday and it's Ai powered. Just don't think I even would give a shit. I think you'd be doing it at what other Series a Series B startups, who are probably still more mature in you. Right? But give you maybe a sexy example like, if you and me say, hey, we're gonna go kill Stripe, but we're gonna use Ai.
Speaker 1 Okay. Job 1, you start working the Ai. I'm gonna stick on a suit and go meet with 7 banks and Visa and Mastercard to see if, like, get permission to go and charge credit cards. You know, like like, that is the option... Will task.
Speaker 1 So... Yeah. And and then how do I go and build a brand the people trust and all that sort of shit. Right? And you realized they're like, yeah, Your ai might be amazing fraud detection, even veterans stripes Ai for fraud detection.
Speaker 1 And your Ai might be amazing like detecting the right... Pricing points for like, the B b Saas companies are like, that's probably, like, 15 percent of proposal. And the order 85 percent of the proposal is where I'm, like, you know, 10 years. Behind stripe, Chase and bank seem if they'll giving me here like merchant provider I can't too. If you're a startup, you have to really believe If you were to build this entire product category from the grand ground up today given what's now possible as of this kind of Ai revolution, you would do it substantially different.
Speaker 0 Mh.
Speaker 1 The definition of the words substantially, I guess means that like, how much of the incumbent products? Technology is still relevant in the future. And if it's like a very, very small amount. Like, maybe they're their login system and shit like that still robin? Yeah.
Speaker 1 Does blood in the water get going. However, if it's like if we take a mailchimp. Let's go after mailchimp, and we're gonna use Ai to write the emails and style the notes. That's cool. Most people like mailchimp because they have really high deliver ability rate, a really good email newsletter analytics and list management and subscription management and they've good spam detection and all that sort of shit.
Speaker 1 Probably have to build all that. Yeah While, you're building all that. Let's say that's that's like, you, that's like 30 months worth of work. Mailchimp will probably work out how to build your. You're...
Speaker 1 You your little Features somewhere along those 30 months. And then you just have what they have, but they have still a far more mature and well known brand. 1 big differentiator you were bringing to the party they have. Because Mh. And it's especially true if the core engine of differentiation is actually on...
Speaker 1 The other end of an open Call and open an Api call. Because in that world, I'm sure it'll work at the prompt to, You know, Yeah. So like, so that's the startup up angle. You really have to sort of say, hey, I think if this area was to be built again today, you do it. Fundamentally different.
Speaker 1 I'll give you, like, an example, there is a product or money products that like you connect to all of your different advertising platforms, and it kinda houses all of your sort of central advertising inventory. And it runs analytics and it'll tell you things like, hey. Our most effective odds are like the following and we're gonna produce we're gonna run Ab b test of this against that 1 and you can go and you can configure and tweak and re upload new versions and all that stuff And you can look at charts and dashboards and all the sort of shit you show your boss. That says, okay. I'm doing a great job here.
Speaker 1 I think that entire product category would be built entirely different today. Yeah. I think the idea would be... You would ask Da ai to generate the. You'd ask the Ai to run the out.
Speaker 1 You'd ask the ai to measure the Lt tv c of the ants. You'd ask the to suggest all of the different bake offs and like Ab b tests. And to optimize the ads per channel per person. Mh. I just think it would just run all down the backward.
Speaker 1 And when I think about a product like that. I actually don't even know what the interface is. It could be like, 1 of those, like Shell script so you just run and you never actually see what happens in the background and you just, you know, trust in the Lord that, like, the money's gonna start coming in. You, you maybe maybe like the Ai also learn so in order to justify divide own volume it spits out at Pdf every now and then just to make you feel like you're doing your job, you know, Yeah. Or help you get promoted or whatever.
Speaker 1 But like I think that type of product category where it's like create optimize, explore exploit, iterate whatever. But think all of those tasks are individually doable. And if you're sitting in 1 of these companies today, go, shit, maybe does as a point. The temptation is say, well, let's just do 1 of them. I But the reality is the actual future is gonna be doing all of them and all be, like knitted together.
Speaker 1 So they're won't. Like, you'll convince yourself, that, like, hey, you surely know it's gonna automate all this. But like I think in a lot of these cases it actually can't be done. When you see how good the the re reasoning of say Gp 4 is, You're like, okay, It's not obvious to me why a human would want to log in here every day and eyeball a list and see the red flashing number Be like, let's turn that out off. You know, or like let's generate 10 versions it is bright green 1 because it seems like it's really good and we should see if we optimize it like all of those decisions can be by Ai.
Speaker 1 So I think like, that's an example where, like, there's probably a massive startup up opportunity. That is worth pursuing in some sense.
Speaker 0 Yeah. So like startups need to clearly understand. The actual business that they're trying to attack?
Speaker 1 Yeah.
Speaker 0 And understand who our customers actually care about? Your value. Mh. Is is is it like the kind of front end stuff, which is much easier for us to kinda see and recognize think about? Or is it actually like a workday case, the back end stuff or swipe case, the regulation, you the lawyers and the chats or whatever.
Speaker 0 Yeah. So I think there's a good question that that you and Have talked about and I know you've taught a lot about that are very useful for bigger companies to think about whether or not they have an opportunity to be attacked. Yeah. Legitimately attacked by a startup, before that though, I'd love to just talk a little bit you touched in different categories there. And I I think we've a couple here that kind of we've been talking about.
Speaker 0 I then we should go through real quick because they make concrete for me. And I'm sure if other people, how things might change. Yeah. And it's not only... It's not like a hop of jump.
Speaker 0 It's like, no Not today. Yeah. For example, like, people talk about, like, you mentioned, like, know, multimedia type things like injury, video and voice and so on and forth. Does a whole bunch of categories, sales? Tools, project management tools, reporting.
Speaker 0 Like, let's start with a with a couple of those. Like, example, sales. Yeah. You know, today, lot's gonna try hire sales people, They spend a shit ton of money training them. Yeah.
Speaker 0 Like, How think that would change.
Speaker 1 Like every aspect, I think is susceptible to significant change. Yeah. The training of salespeople under our training tools. All of that, like, no can be training where word Ai sits live into Gone call and provides, like real time, updates on like, hey, they ask you pricing. Here's pricing.
Speaker 1 Hey, they ask but this here's a slide. Hey they ask but here's a video to play. Here's a customer to reference. Here's a testimonial Like, so Think all of your training, it's gonna be much more like in ear versus after this call, Johnny, we're gonna sit down and talk to you, but all the things you should've have said. Yeah.
Speaker 1 Right. It's much more in into moment. Right? So that's just training. That's before we can get to your desk.
Speaker 1 Sales, like, 1 not motorola of sales will be the same prospecting. Right? This idea of here's a list. We're gonna go and go through this list, try and to find to people who are credible, kinda try and make contact with them, reach out to them We're gonna try and ring them. We're gonna try and email them.
Speaker 1 We're gonna maybe, like, target ads against their specific email address so hopefully you can follow them around with the Internet. I haven't set a single thing human needs to do. Mh. Right. Look at this list, ai I can do.
Speaker 1 Leads score assist, Ai can do. What are dots directly or by api over to like a Zoom info getting leads score backward. Email these people Ai can do. I recall these people, Ai can do. You know, prioritize, run outs against these people, Ai can do.
Speaker 1 You know, targets specific testimonials as use cases and Pdfs and sales, that specific to this person in this industry, Ai I can do. So, like, that's 1 example, like, there are there are companies phone call reg dot ai, we'll call that n or nukes dot ai that are that are just like really looking at real specific value points in the sort of sales workflow and saying, Right. Draw a line around this, we can do all of that. Yep. And and like that's...
Speaker 1 By the way, that's awesome news for salespeople like, it's like a lot of the differentiated heavy lifting. Will get taken away an everyone's path to being what they wanted to be, which I presume with without, like, you know, senior sales leader or a senior, like, you know, sales rep like, deal dealing with higher deals at higher values. It's almost like, we've taken away a lot of the, like, the training courses. Mh. When it said, hey, turns that no needs to do any of that you anymore.
Speaker 1 So like let let's get you into the mixer straight away
Speaker 0 Yeah. Yeah. I think for a lot of these things, it's... Well, 2 categories things. 1 is for some people, but sales.
Speaker 0 It's the same job selling. But, will make the job much easier.
Speaker 1 And more phone.
Speaker 0 More fun. Yeah. Sure. And then the other category of of thing is where people's jobs might change. Like project management another category where probably people's jobs will change.
Speaker 0 Yep.
Speaker 1 Because of Ai. I think so. Project management quite nuanced. So think this is 1 area where you're seeing a lot of Ai applied. And I think a lot of it's like what I call like, the condiment style ai.
Speaker 1 It's... Salt and pepper. It's not the dish. It's just a little bit of kimchi on top. I'm wary of the whole Right, the first sentence of a of a status update and press cut to expand where it's like I take this project is on chorus tub.
Speaker 1 But the following risks were made, you know, like, because I'm like, I rather I was actually comb of your head done, like, Gp and. It. Because I need you to sign it over because you're putting your name against, it actually tells me that you profession you think that we pay you to understand these things.
Speaker 0 Yeah.
Speaker 1 You know, so Purring I do worry a little bit that sometimes you might get overuse in these areas. But to say, like, think about something like an Asana Jira or a base comp or whatever ever and say, how could Ai help? The organic activity of water are the things that happen here like, so let me know what's going in this project. I again, I can do it right? Yeah and you can sort of base basically ask, Gp 4 and say, read all the most recent for.
Speaker 1 Depend out to your most recent knowledge on see what are the semantic differences that an executive would care about to the status of this project and if it's still
Speaker 0 course. Yep.
Speaker 1 And send me that every day as a slack message. You know? That that that's like, you know, I, again, just thread there as a project you'll notice again we're moving away from the Ui to just being a push versus a pool. Right Yeah. Rather and logging in in every day.
Speaker 1 You're just gonna be told shit's ever go wrong. Mh. Find the root cause of all these issues. Why is this project running late? You know, maybe other stuff like, who has contributed, like, the most to this project in terms of making concrete decisions?
Speaker 1 What decisions have been made to solve? What was the biggest reason this project was late? There's a lot of stuff there that can actually change where I think the current workflow for trying to work this out is, honestly, and you you've probably don't it yourself, not not to your team are imperfect. But, you know, you probably have to just know and end, like, sit down and, like, read an entire like like you know, 4 Google docs and tree based comp or whatever to try and work out, like, what happened when Was away. Yep.
Speaker 1 You know,
Speaker 0 does it even matter to me? Yeah. Yeah know.
Speaker 1 Yeah. Yeah Yeah.
Speaker 0 A lot of stuff could have happened and a decision was made and that we're good at the decision. Yeah. Context actually is so unnecessary.
Speaker 1 Yeah Totally, But be sometimes you're just digging for the decision almost right like... Yeah. And so like, imagine a world where you can log in and say, like, the reason I'm actually logging in to base come today is because I need to work out, Are we on track for August 11. Mh. Be able to get to that level of, like, here's the thing I actually wanna know.
Speaker 1 And like, the words don't really matter to that can be quite powerful. I have yet to see that, like, doing well to be clear by I suspect it will happen. Yeah. So like, but the nature of a Pm tool will change. From not point of view, Think this shift that we haven't...
Speaker 1 We don't necessarily... You know, we're not a consultancy. Like, I think identifying conflict resources and so I thought like, hey paul across these 7 things and these actually books to be here. That sort of stuff could be pretty useful too. Mh.
Speaker 1 So I think, like, in general, the Pm tool is definitely right for it. But I think I'm I'm a little personally allergic to, like, the tab to complete massive paragraphs of writing and judgment because III prefer if that actually comes from someone's brain at least right now.
Speaker 0 Yeah. Yeah. Another 1, let me final we'll cover today. We'll we'll go back to some practical things for to think about how's takeaways. Is reporting and reporting too.
Speaker 0 Yeah. Right? So, like, for example, we hear Income have spent the best part of a decade. Mh. Building deep reporting.
Speaker 1 Yeah.
Speaker 0 And at you know, editing reports, creating reports, all all the kind of typical things, you know, from from a crow point of view, like create, you
Speaker 1 know, create new report like, change of filter rise a facet it. Yeah.
Speaker 0 The more we build, and the more research to the customers, the more we learn... There's more to build.
Speaker 1 It's like never an story.
Speaker 0 Yeah. More config, more customization, etcetera. Now though, you realize Ai can probably do a lot of that. Yeah. And there's no need to build all of these things
Speaker 1 Yeah.
Speaker 0 Or use them if the beam built already. And we finance ourselves a position who we're kind of thinking, like, we're still building reporting features. Yeah. But also wondering like. Should we also be building the need for our customers to never use them?
Speaker 0 Yeah. Instead, they have some kind of, like, literally a a field yeah. And the type in the question, like, you know, is... Lt tv up or down or, you know, is my customer's support volume gone down or what would everybody say this week? Right?
Speaker 0 It's all chop based Ui. Yeah. Ai will clearly be good at that. Yeah. I think it will do things like uncover correlations in data.
Speaker 0 Mh. That humans never would. Purely because there's just so much data.
Speaker 1 You know Like... And it's so much more powerful than He anyone press like.
Speaker 0 Yeah. It's exactly. Yeah. I know he's just do so much more. Yeah.
Speaker 0 Yeah. Yeah. I think the role of humans might be less about, like, digging into the data and the analysis. Yeah. And much more of a judgment.
Speaker 0 Yeah. And then beyond... This come here and, like, do the analysis, apply human judgment, then make decisions. Yes. Think humans will move away from the analysis spark I would do that.
Speaker 0 Yeah. But they'll apply the judgment to make the decisions, but but you said, and I agree, The Ai would do a judgment too.
Speaker 1 Yeah. Yeah.
Speaker 0 Can you explain that of it?
Speaker 1 Yeah. Like, I think... I'll get this wrong, but like, there's is education educational psychologist called Benjamin Bloom, who like... Was trying to describe how how you get to know something an area of any sort. And he has this thing called Blooms taxonomy of educational objectives.
Speaker 1 And at a very, very, very low end his recall, right? Just... Can you list to 26 counties of ireland right Like, there's no depth today. You're just looking at a Roast Common Mona and whatever. And it's a very, very, very high end like, synthesis.
Speaker 1 Can you create new stuff based on existing stuff Right? Yeah. So you go something like recall, recognition, comprehension analysis and synthesis. This. I'm skipping 1 or 2 there and we'll put we'll put a better diagram in the show notes.
Speaker 1 But but I think, like, a lot of people are are only really comfortable with Ai as like as a house pass they they they like it at the low end. Right? It's a. Like, in the same way if people are cool but like, typo correction or whatever right. But I think Cam...
Speaker 1 Yeah. We have to get more comfortable with Ai is appear in a sense. Right? Yeah. I think Ai will be to apply George because like, even if you take our own ba fin, like a lot of our is like given this answer that.
Speaker 1 So like rewind dot ai is a customer fan I a rewind user it's an awesome product. And I was asking... Rewind us this thing where it wants to record every meeting and I didn't want to do that. So I was trying to disable this pop up. And I went to rewind help sir.
Speaker 1 And I, how do Disable to pop up and fins. Oh, here's how do. And a linked open article where it never... The article never directly said. To disable this pop up.
Speaker 1 Here's you do it. Yeah. What the article said was something along lines of... Oh, if you wanna turn on this, you go here to do it By the way, when you do it, it won't be always on. It's gonna pop up every time and Finn inferred that having read that article.
Speaker 1 It was, like awesome. If that's the thing and that's the preference for it must be a this screen, and fan like basically gave me a perfect answer. That to me is like a a very I'm amazing that just because knocked promote, which is because it's a very live in the wild, like, the it is here today type example of, like, deduction or judgment and un suggestion. Like, it it was confident enough to tell me that was the answer. Which I just saw was a in a a simple, great example where, like, no 1...
Speaker 1 Like, no 1 in rewind Tata she write that answer. Mh. Didn't work to do. In the case say reporting. Imagine we go from, like, you know, show me which the s reps get the highest scores.
Speaker 1 Right. I think that's a pretty simple task. Right? Yep. You know.
Speaker 1 Then you could say, like, show me what topics correlate with the highest... Scores. That's probably pretty simple. You can say, like, show me which Cs sr reps tend to perform the lowest on on which topics. And maybe I could be like, where you have better better training course or whatever.
Speaker 1 And then you could say, like, prioritize that list and suggest the type of training they should do. And then you could say, mail those people and tell me to go into our training. Right? All of that is like is like judgment in a sense. And And I think the...
Speaker 1 Like, it's not clear to me, like where the Ai stops in its capability here. What is clear to me is that a lot of people are... Does a does a human comfort level in terms of, like, you can go that her, but, like, I need to be the person who fix this. And it's it's not too far off, like, the old Gilbert cartoon of the pointy here boss who just like, likes to feel important. So, like, gets to be the person who presses the launch button or whatever like you.
Speaker 1 Like, fundamentally, they're actually irrelevant but many these scenarios, but they... Their needs to feel important and still represents itself. And I think there's a lot, like, a lot of our first pass attempts that, like, using Ai will be like that. Like, well, hang on a second. Like, sure all the low level shit can go away.
Speaker 1 But I still need to be here for the importance. So if you know... Yeah. So I... But I do think you're gonna see the, you know, the like, did does some like dark sort of futuristic cartoon where like, you know, there's a gil humans on a factory store I'm, like, they're all there to like, you know, do certain things into the button on a switch that they can click in case at.
Speaker 1 Anything's that ever going wrong. And then the other side of the wall, those things aren't wired up tiny anything.
Speaker 0 Mh.
Speaker 1 It's just there to make the humans feeling important. Yeah. Give them a sense that they're part of this process.
Speaker 0 Like traffic lights western traffic like
Speaker 1 as your roads.
Speaker 0 Yes. The button doesn't do anything.
Speaker 1 Yeah. You're. But it gives you a sense of economy your matter in this world. Yeah I hop into the world. So, like, I I think there's a bit of that.
Speaker 1 That we all kind of apply, which is just like, well, surely, the the ai I won't be able to go offer. I understand all of the natural human desire to cling onto the well shirley but in practice, I think we're gonna just see we're gonna see that bar, just creep up and up and up. Yeah. Especially given the, like, you know, the reality of the stitches, like... It it tends to be pretty right and it tends to be quite accessible and probably works 3 06/05/2407.
Speaker 1 You know, Yeah. So, like, I I think you're gonna see judgment, like, what people define as Judgment just creep open and up and up. Yeah. I think the stuff they work gets more funky is like, and under our values reasons for this. Ai is not perfect.
Speaker 1 Neither our humans, but Ai is not perfect. Under are some decisions really like, right. Let's not launch the email campaign without a human eyeball. Totally valid. So what you can imagine might happen there is all the work up to the last like step of the marathon, might be doing by Ai and any human moves in who goes.
Speaker 1 Yep. Click. Like, that's sort of stuff, I think makes sense. You know, In fact, that's just logical.
Speaker 0 Yeah. Let me ask you, like an incredibly perspective question. And we'll quickly move on to give people some very pragmatic tips on on questions to ask themselves, think about how much this matters to them and how fast. Sure. So the second question is about how fast you know we're talking about, like, from analysis to synthesis.
Speaker 0 Mh. Then there's judgment and making decisions and humans, for sure, we all will feel in... Need to control it, Mh. And I hit the red button? And so the decision making?
Speaker 0 Yeah. Do I do do do we or don't we hit the red button? Yeah. Is left to us. How far away do you think we are from tools, really great software tools that are very act...
Speaker 0 That are accented judgment and not pushing us on, like, maybe they should just make the decision to.
Speaker 1 I think it's gonna go the same way as, like, you know, like, the features we've built in come like role based access controls. Yeah. I think it's gonna be like that. I think we're gonna have, like, literally be buildings. Preference dialogues into Intercom and order tools, where it basically says, like, you'll have you'll have a lot of settings that begin with, allow a 2 dot dot dot.
Speaker 1 Because I think there is like AAA spectrum we spoke with as a start of, like, ai skepticism. Yeah. I think you kinda want, if you're in any real sort of market, you wanna serve all ends in a market. So like, you could imagine, like L allow Ai to, like, you reply, requests, Csa sat scores, l Ai to ping my own support team under Csa sat scores are dropping. Hello Ai, you know, like all the way down to, like, some probably slightly bigger things like a Ai to, like, you know, what be extremely here, like, post job opening on indeed doc come because we're clearly under stuff.
Speaker 1 You. You know Like... Yep. Like, it's all of that, like, you know, there's a spectrum of like, what are the king's humans would do there on, like, you know, almost like an if and out type workflow player. So, yeah, the...
Speaker 1 That's basically how I would... How I I think we're gonna end up. Have to... So then to the case of, like, your question was maybe how long before we see this. I think, you know, there won't be some waters at moment where it's like it is here.
Speaker 1 It'll just be like what what might happen is we sit down next year, and we're talking about... Like, if today we're talking about maybe ai I can see correlation and reporting. The next conversation we might have might be, like, maybe the Ai should be sending around and suggested next steps. You know You know, like, as we're we're past the disclosing correlation, not ship sale. I just think this conversation will be like the the like the continuous incremental creep of what we believe to be possible on what we're coping with.
Speaker 0 That makes sense to me too. And it's a similar pattern, you know, history is the best of the future and love these cases. Similar pattern and things are a about. Where, you know, Like, the first I told is very, very basic. Yeah.
Speaker 0 And then every release, so maturity more power more. Yeah.
Speaker 1 You're totally you're right. Just as a sign. When I was like, Web 2 consultant. Discussion the time was like you'll never do x on in the cloud right? I was like, you never have a word processor in the cloud.
Speaker 1 You never have like, a video editing tool tear. You never blah unlike every single now you can play counter strike in cloud. Like, it's like literally full on proper desktop gaming in cloud. Mh. It's old done true browser and then similarly, you'll never do x on a phone.
Speaker 1 Like, yeah Phone's good all up. Like, you're not really gonna... Whatever the thing in there was, you've done it. You know, you're not gonna find a date. You're not gonna apply for a mortgage you're not gonna buy a car.
Speaker 1 Yeah. It's like it turns out you do. It turns that you do all of these things. So when people tell me you're never gonna really do blog with Ai. I'm like, may, I've been you know, I've done this rodeo many times, and I'm telling you you're probably will.
Speaker 0 So absolutely. To give people like, practical questions for them to think about how quick this might happen Yeah to them in their industry. There's a couple of ones that that that that we have the... I know you've used a lot here to talk to our team and our product, org. Things like, you know, how can this Ai technology be applied to create new features?
Speaker 0 Or how can they be applied to make existing features easier.
Speaker 1 Yeah. Better, more powerful.
Speaker 0 Yeah. You wanna talk us through that, like
Speaker 1 Yeah. Yeah. I mean, like, the the core point I always come back to would all new, like, capabilities what are, like, whether it's Ai or, like, you know, chat or messaging or anything new that's ever come along, It's like, what is a product. The product is like usually platform of feature features that let a user get a certain job doing or a certain set of jobs doing. The questions you ask yourself as, like a product manager as a product leader is like, given the available.
Speaker 1 What is the best way to our users can get this done. Right? And the the idea error is like, it's kind of the core jobs to be done idea. Right? Which is kind of fixated in this, Jobs don't change, technology change.
Speaker 1 Solutions change, but the job is the same run. And I think generally with these things, what you're trying to do is either, and make it so that more people can do the job. And a great example that will be gonna say equals to spreadsheet company where like, I don't know excel functions. What I do know what I want to add of them. Yeah.
Speaker 1 I want to see the average growth rate of to start over the last 6 months if you exclude organic traffic. And I don't I have to do that. But I can I know Can write it into a box and equals a little work at what I mean? And Write up the formula for me. No.
Speaker 1 I don't. I don't actually know if the form is right, but it seems to be most of the time. Yeah. Or if it's wrong, it's so egregious wrong it's not a problem because I can correct it. But like, that's a great example where it's made it possible for more people to do to sing.
Speaker 1 Alright? And if your tool involves either arc languages, complex query stuff, or a creativity, like, as in... I know I wanted to have a fancy black image, but I don't know how to. Design mh our designer. Like, so can you make...
Speaker 1 Can Ai help us help Sort of more people can do this job. Yep. Like, an example might be like, hey, we want to let all of our English being support staff, be able to support all languages in Europe. Can Ai help probably. Right.
Speaker 1 You know. So can Improve many people who can do to job. Like, increase it. Right? Now that's usually.
Speaker 1 That's massive the impact for your market size. Yep. It means more people can use your tool. Yep like, more people can use equal they can use excel.
Speaker 0 Tools for for narrow markets that require.
Speaker 1 Yeah. Yeah.
Speaker 0 Become tools for general markets.
Speaker 1 Yeah. Because because you... You change 1 core thing, which is even the people who know what they want to do and even if people who can do it are near the same thing. Yep. Right.
Speaker 1 So that that's 1 huge thing. So kind Ai all of this you kinda make it set of more people, can use your product ultimately and shi Ui, the huge part of that. Literally 1 is kind make people kinda, like, increase the power of their work like you the the anatomy here... Sorry, analogy geez here would be like, a crane. Right?
Speaker 0 Mh.
Speaker 1 Is that if I jump into the into a crane, I am now much stronger than I was before. I can move stuff at a far greater rates. Still me still medium to work. Right? Yep.
Speaker 1 But like, now I'm lifting heavy heavy stuff. Right? If your 9 was capable of. And so similarly, if if a human can summarize 1 conversation at the time, can Ai summarize 1000000 conversations of the. Or whatever right.
Speaker 1 Like that that type of technology or, like, you know, you referenced looking at correlation across all datasets. Mh. Like time a human can do that 1 by 1, Ai does not need to, like, act 1 by 1. Yep. So does that type of thing, which is just increased the capability of the human.
Speaker 0 Crane is a great example. Because, you know, when you're saying, like, you know, 1 I you in this case gets in the crane left the the the the volume of thing, that 80 people would have had to, like, manually... Exactly. Right? It's any...
Speaker 0 I I think you can apply that very mh clarity to Hey, what are things that lots of people required to do Yeah. Where Ai could make it So that 1 person overseeing
Speaker 1 do do
Speaker 0 it by itself.
Speaker 1 Absolutely. And like, fin snippets in intercom is like, when 1 person answers a question property, Ph say, hey, is that the right answer because if it is, I'll take it from here. That's 1 person fact doing the work of all future people for the future. Know. Yeah.
Speaker 1 Is a a type of crane. Right? And then the third category, I think, like, you have to look out for it. It's the 1 like... Yeah nearly ironically, it's the 1 that people tend to.
Speaker 1 It's just like, other things we can just get rid of entirely. Like, because in things were like, it's not even to do it in to create anymore you know, like, it's like we we've actually just taken away a need for that in its entirety. Yep. And if you're call say the advertising example I talked to the ad platform whatever It's like Johnny logs in every day to look at all the various charts and tables in my you. Like, there's definitely an argument where yes.
Speaker 1 You just don't need that done at all. Like that's just you you just assume from this point onwards in the same way you assume the electricity works in your building, you assume the ads are optimized. You know? Or if they're not optimized or getting optimized. You remember like, like, there's easy need to do better.
Speaker 1 So, like, there's definitely has third counting error as well. So, yeah. The... Just to Zoom pockets? Like, what are the new capabilities?
Speaker 1 Like, what are new things that people can do at what are the things that like the 10 x human capability? What are the things where you can expand the addressable market. And then lastly, 1 of the things we can just remove entire chunks of work. Right? And like all of that is is is like, generally how I think...
Speaker 1 Be thinking about this. Which again, goes back to the said, this is why I'm not nano skeptic. I just... I see too many opportunities even in a in a pretty prescribed domain like customer support. It's just so clear to be like, all of their ways in which, like, we we could use 10 times, you know, if, like, Ai and ml people to go after somebody many if there's so many opportunities in the space.
Speaker 1 And that's, like, every time I get ping by, you know, we're doing Ai for customer support, type startup. I... I'm I'm the quite frustrating Like that's a brilliant idea. Like, we have... We either have or haven't taught of it, but like, does so money brother.
Speaker 1 I can't you know Yeah. And like, that's just in 1 little domain. You know.
Speaker 0 Yeah. That's great. That's really really good. I think practical advice. These are good questions.
Speaker 0 I think all companies what are your startup and comment. We've talked a lot today about. How starts you think about entering categories, how I can... Ai can just disrupt that category or not as the case may be. And I'm on the incumbent side, I worry more about those bigger companies because I subjected this myself at times where I'm like, hang on a minute.
Speaker 0 We're like domain neck first, We've been here 10 years doing this. There's no possible way. Yeah. Ai could ever know the things we know. Totally.
Speaker 0 Right? Nonsense. Of course. Yeah. Yeah.
Speaker 0 It can will.
Speaker 1 Because The kids don't know what they're doing. Yeah. It's nice.
Speaker 0 And the order you getting the longer and the are that the stronger the feeling gets. Yeah. So they will leave it there for today. Yeah any kind of last pressing advice for the startups or incumbents or even investors which we haven't on covered today.
Speaker 1 We've covered the the the gist of how I think about the space. I would say like it's a good time to re reread the innovators dilemma and to remind yourself the true nature of disruption. Like it has to be a it has to be a new attack vector that the incumbent businesses can't easily take. Yep. And I think a lot of people are gonna say that they're gonna disrupt industries with Ai.
Speaker 1 And I think if you are ever tempted to say those words at all. Just do yourself a favor and even just read 1 of the 6 page or Harvard business review papers on it whatever, really refresh exactly what it means. Be properly disruptive or it's low end disruptive order It's like new use case disruptive for new market disruptive. But like, really just make sure that you know what you're saying, because I think there's a lot of businesses that will build a really cool piece of product, but it'll ultimately end up being unpaid R d for the much bigger company because they're lockdown down and got like, that's clearly the right thing. We should do that.
Speaker 1 Mh. And that will be it. Right? Like, you might have a cool new way of doing some specific task in, accounting in surveys and time tracking and, like, whatever expense tracking. And you might have a really cool little feature that's dripping an Ai.
Speaker 1 And I might get you, like, honestly, you might a product on feature of the of the day, like you, you might have a sexy landing page. I might even tweet about and say check out just dope shit, right, like, It could be stunning. Question is, is it a enough of an attack angle to be truly disruptive or will some principle engineering and principle designers sit down in mega big corp and be like, we should probably copy those. Yeah. If they copy, it might take them a year, but in that year you're unlikely to have built at a fully matured platform, and that's that's the challenge.
Speaker 1 And maybe that's okay. Maybe maybe you're okay being like, hey, we're gonna go after the low end of the market. We don't Compete with Mega corp. That's fine. But just make sure you're making all of those decisions together.
Speaker 1 And that which is black. We're going to kill salesforce because we've have an Ai based leads scoring algorithms. I thought salesforce are going to work with now.
Speaker 0 Yep. Yep. Yep. That's great as I'll see with there for today, and I'll see you maybe 12 months, so we can figure out. Strategy.
Speaker 0