Subscribe
Share
Share
Embed
IIIMPACT is a Product UX Design and Development Strategy Consulting Agency.
We emphasize strategic planning, intuitive UX design, and better collaboration between business, design to development. By integrating best practices with our clients, we not only speed up market entry but also enhance the overall quality of software products. We help our clients launch better products, faster.
We explore topics about product, strategy, design and development. Hear stories and learnings on how our experienced team has helped launch 100s of software products in almost every industry vertical.
AI agents shouldn't replace staff, but instead should augment employees. It's really kind of leveraging their intellectual capital to guide and optimize agent functionality.
Speaker 2:If an agent can do your job and can't transition into a creative or strategic role, then what is your value in the economy? This is the uncomfortable multi trillion dollar question nobody has an answer for just yet. All right. Hello, everybody. Welcome back to another episode of Make an Impact podcast.
Speaker 2:I'm your host, Makoto Kern. I'm the founder of Impact. We've been in business for the last twenty plus years, launched hundreds of digital products, and provided a cheat code to successful software product launches. We turn chaotic b to b and enterprise projects into predictable wins through proven strategic workshops and derisk road maps. Welcome back to the show.
Speaker 2:And I've got with me as as my cohost, Brinley Evans.
Speaker 1:Our lead
Speaker 2:he's our lead UX designer, been with us for since the beginning. And, yeah, we've got a exciting topic for today.
Speaker 1:Yeah. Looking forward to diving into it.
Speaker 2:For sure. We're diving into AI agents and not just the buzzword kind. You've probably heard the term agent thrown around, but what is an AI agent and how soon will you be working with one?
Speaker 1:Yeah. Well, how about yesterday? Because it's I think that the real question as well isn't like when are gonna be working with one? But, you know, really, you know, if you're gonna be the one giving the orders or are you gonna be the one made redundant by an agent that does the job faster, cheaper, and pretty much without lunch breaks, coffee breaks, or even toilet breaks, without sleep breaks, It's definitely a new era coming. And, yeah, it's gonna be very interesting to see.
Speaker 1:So, you know, with that said, you probably are working with an agent already. A lot of products have started putting the initial agents in place. We're obviously gonna see a lot more development, and I think that's why this is such an exciting topic. The fact that agents are set to explode in the near future. And, yeah, really an agent is going to be one that's, going to not just be able to perform tasks, but I think what makes it unique is it's going to be able to reason and then take action on that as well.
Speaker 1:And then obviously work across apps to get all the things that you need done done.
Speaker 2:Let's talk to the audience about what the heck is an AI agent. Let's imagine you're in a kitchen. Old software is like a recipe. You follow it step by step. An AI agent is your personal chef.
Speaker 2:Tell it, I want dinner for six, gluten free with dessert, and it figures it out. Or think of Jarvis like from Iron Man.
Speaker 1:Yeah. Or you could say, Howl 9,000 from 02/2001: A Space Odyssey depends on which side of the coin you kind of feel this technology is gonna land. But I guess either way, agents, again, perceive reason and act. They're not just those chatbots that you may have interacted before.
Speaker 2:Alright. Let's let's kind of be real about this. Perceived reason and act sounds great in a sales pitch. But, right now, most of these so called agents, they're just glorified chatbots hooked up to an API. Think of Clippy even.
Speaker 1:Be controversial there. Nice.
Speaker 2:Yeah. I know. They can book a flight, sure. They can also fail spectacularly by, like, booking the flight to the wrong place or, you know, the same city but a different country and then deleting the confirmation email. The potential for for chaos is just as real as potential for productivity.
Speaker 1:Yeah. And I think if we look at them now as well, it's early days, but we're looking at, I mean, GPT-five was just released, that's got an agent mode with it. Quite interesting if you've played around with that because that starts sort of almost having its own sort of browser window that it starts performing actions in. And I think we've got to look ahead to the predictions of AI agents really taking off in sort of 2026, and sometimes even surpassing what a human level employee will be able to do. They'd probably pick up a lot of criticism for maybe poorly implemented ones.
Speaker 1:But you've got to remember this sort of technology that they're built on has only really been made official in the last few months when we look at things like MCP servers. So there's a lot that is going to be formalized and improved on and refined, but they're really going to be a really sort of powerful addition in the future. And if you look at, again, the way they can move between applications, they'll be able to know what your goals are, what are you looking to achieve? What's the background on that? And then being able to sort of take action at each step.
Speaker 1:What I like is Bill Gates mentioned that agents will upend computing like GUI once did. And GUI is obviously the graphical user interface. So, Bill Gates with Windows, anyone that's not familiar with one of the probably most famous people. But that was another massive jump where he had to think of operating systems that were purely in text and think, well, how are we going to get a graphical user interface? Like, what are we going to do?
Speaker 1:And that's where windows came with the buttons and the links you can click with the dropdown. So a lot of those elements were forged from this. You've got the sort of visionary also identifying that they're really going to And this, I think, was even back in 2023, that they're really going to be a lead, a transformative shift in just changing how people interact with technology.
Speaker 2:Yeah, I think Sam Altman said in 2025, we'll see the first AI agents materially change company output.
Speaker 1:Absolutely.
Speaker 2:Yeah, and why now? I think the massive jump in model performance in kind of the recent GPT-four point zero, Claude three, and integration with tools and APIs.
Speaker 1:Absolutely, I mean, that's what we were just mentioning as well. There's the infrastructure for this, the fact that it's only being standardized now and rolled out is that you probably have even interacted with agents that aren't potentially even built on a lot of those protocols. So things are set to change and they're going to change rapidly.
Speaker 2:Yeah. I think this kind of is a good segue into kind of like knowing the AI agents are already at work. And so, you know, we think of AI agents as it's all futuristic and it's not here, but it's actually already here. And it's not really the humanoid robots, but it's the digital workers really working behind the scenes.
Speaker 1:I think that's a good way of thinking about it as well, is that they don't have a physical form, but it is sort of digital workers that, you know, are are interacting and processing through things. And you can think of, you know, everything from sort of handling legal reviews in in legal firms to writing up the the soap notes for clinics to managing inventory at Walmart. It's already here and it's doing high value skilled work. And this is just the starting point.
Speaker 2:Yeah. And there's a few different agents that I've actually played with already. One of them is Manus and and, you know, DeepSeek is another. But it's amazing to see because you actually see it working in front of you going to the websites, doing the tasks. And some of it does take longer versus, like, trying to pull information from GPT or Claude, it spins it up very fast.
Speaker 2:But when you see them going to the websites, actually scanning it, booking things, creating reports, all that in real time in the applications, you're like, wow, this is actually amazing. I think some of them do a great job. I almost call it like a brute force way of doing things, but it does it pretty effectively. Yeah. It's exciting times.
Speaker 2:So I think, you know, one of the good things, and I'm sure people in the audience wants to know, how do you prepare for AI agents kind of in your business? It's not just an IT problem. It's a boardroom strategy conversation. I think a McKinsey adviser calls this a business strategy moment, not just a tech one.
Speaker 1:Yeah. And I guess, I mean, if people are looking for practical steps and roles, you know, that they can you know, what do you do now? You're interested in agents, you've discovered them. How do you get them into your business? And probably the first thing you want to look at is maybe updating your software architecture.
Speaker 1:The agents need the access to tools and data, and those are two very critical things. Like how are you structuring your data? Is it accessible? How are you handling the security and privacy behind that? Square all those basics away and also put the tools or protocols in place to actually give these agents access to start working with whatever it is that makes your business unique or what your business deals with.
Speaker 1:Also, there's probably a lower barrier to entry on starting with more pilot friendly platforms. So some of your big names like Microsoft Teams, Slack, even Salesforce, they're starting to put, whether it's MCP servers or just really simple interfaces to start automating processes. So start with those and get handle on that. And then I would say it's also about really sort of appointing, I think what they're calling AI integrators. Can you put in place people that are responsible for embedding and managing these aged workflows?
Speaker 1:I think there was a recent posting by Cora actually looking to hire the sort of dedicated role of AI integrator. So it's starting to sort of think about those new roles and what they can do in your organization.
Speaker 2:Yeah, this is kind of where the corporate UFC fight begins. You're going to see this kind of turf war between not just IT versus marketing anymore. It's going to be the AI augmented workforce versus the AI automated workforce. You know, the integrators are going to you appoint basically will hold a lot of the power. They'll be deciding which jobs get supercharged by AI and which ones get replaced entirely.
Speaker 2:So this isn't really a technical role. It's a kingmaker role.
Speaker 1:It's true. And I think that's where a lot of projections go as well. You know, how will society and roles be sort of split? Where will the workforce go? Where will it disappear?
Speaker 1:And that's where they're saying there'll be this sort of upper echelon that anyone that's controlling these will be in a position of definite power, I think, and to be cognizant of that really. And this is something we'll look at in this podcast as well, how you can really empower your employees to utilize this. So it doesn't necessarily mean replacement, it just means how do you drive efficiency more? And I think these sort of roles are going to have to be carefully filled to make sure that people still recognize the value of human contribution. Yeah, looking at other things, other points we looked at addressing your software architecture, appointing AI integrators.
Speaker 1:Next point would really be looking at training your staff to work with AI. And that's sort of building on the point I just mentioned is, do they understand AI? What is their prompt fluency like? And how do they learn to really work alongside, promote this human AI collaboration? These are almost things like, back in the day, you'd say, well, you're going need to generate reports.
Speaker 1:You should probably, no matter what you do, have some basic Excel skills. Pretty much the same thing. These are tools, if you're not using them already, which you should be, if you're watching this and you're not using AI, please sign up with any of the platforms. Get going with it because otherwise it's going to be more and more difficult to join with the speed that it's going.
Speaker 2:It's amazing. I've talked to front end developers who just not not even really played around too much with it. And, yeah, it's amazing. I mean, it's it's, you know, maybe there's certain things they've tried as far as just basic tests, but they're not really leveraging it. And it's interesting where even you know, I I'm encouraging my kids and everybody to to, like, really embrace it and just get used to it and see what it can do.
Speaker 2:Yeah, there's certain things that you want to watch out for and always double check, but not being on it at all, you're going to be left behind.
Speaker 1:Absolutely. Especially when, I mean, just talking about that with software development, the tools are incredible. The agents that can literally go through your code base and find any issues and vet your PRs. And, I mean, they're just opportunity's huge. And I think there's a nice quote by someone by the name of Darryl Keeling.
Speaker 1:He's kind of a senior VP and CISO at a place called Parkview Health. And what he said is, The real opportunity lies in training staff to thoughtfully integrate these agents in their daily workflows, not only to automate repetitive tasks, but to enhance the value of their expertise. And I think that's so true. You've likely employed people with certain skills and they've got valuable experience and opinions and insight that they can provide. And it's really just plugging them into the right tools to accelerate that.
Speaker 1:And he went on to say, well, you know, AI agents shouldn't replace staff, but instead should augment employees. It's really kind of leveraging their intellectual capital to guide and optimize agent functionality. The last thing is If
Speaker 2:they if they give you any lip, then you get rid of them and you just hire the AI
Speaker 1:think you read on to say true success lies in leveraging the intellectual capsule of our employees to guide, shape and optimize these technologies, not replace them. And then the next one would really be, one of the final points is think about guardrails as well, establishing those. Things like ethics, compliance, approval thresholds, auditing. And this is particularly critical in areas like finance, legal, even healthcare. You obviously want to prevent harm from any errors in high stake decisions.
Speaker 1:So these agents are able, are empowered to make the decision. So put those safeguards in place. Make sure you're compliant with any regulations and standards. And then something we've discussed a lot on these podcasts is protect against bias as well. Remember, these agents are trained with human data.
Speaker 1:They can be biased and they can be through the content. There could be discrimination or even ethical violations. So just make sure that you have that in place to guide these agents and make the experience one that's fair and valuable for your business. Just, there we go, I like it.
Speaker 2:So what happens if you don't prepare? So Gartner released their stats and they said 80% of businesses not experimenting with AI agents by 2026 will fall behind competitors. Totally believe that and we keep pushing that with our clients that are saying, hey, we could just do AI and we're moving so fast. Like so is your competitors, it's a high tide raises all boats. If you think you're using it and you're getting an advantage, so is your competitors who are using it as well.
Speaker 2:So not only do you have to have speed, have to have direction and you have to continuously improve.
Speaker 1:Absolutely. I mean, that couldn't be more true because speaking of barriers to entry as well, I mean, those are just dropping, especially with the code space. Getting to the five coding now, so many people are building applications. If you're in a digital space, you're going to have competitors coming up and you want to make the most of the lead you've got and ensure that you keep that lead.
Speaker 2:A lot of people there are thinking, there's an idea for this product that I have that helps me do X, Y, or Z and I can vibe code the whole thing. So sure, you can do that and then you could try to sell it, but then so can somebody else. They'll just vibe code their own solution. Our products like SaaS products, are they going to be basically all internal? You're going to have a hard time selling it.
Speaker 2:Yeah, it's a great idea, but then somebody is like, Oh, let me just vibe code this myself. I don't need to pay you a monthly subscription and then I've got it. Or your monthly subscription starts to dwindle very quickly because other people are now vibe coding your SaaS platform. So your solution has to be something that stays relevant and stays up to date and is super frictionless, easy to use because, yeah, if I see something that looks nice and it just there's a few features, I can vibe code that really quickly. And if it's internal, I don't care about security so much.
Speaker 1:Mhmm. And I think identifying what it is that makes your business valuable, is it the business processes? Then you've really got to identify what those are because it's not going to be your interface. It's not the fact that you offer standard features in a web interface because, you know, as you're saying, just code that needs to be everything behind the scenes. There's a lot of thought and value in it, that's what's going to probably drive your product.
Speaker 2:Mhmm. There's a gentleman, and I will get his last name right, Eric Brynjolfsson. He's the director of the Economy Lab at the Stanford Institute for Human Centered AI. He's a senior fellow at the Stanford Institute of Economic Policy Research, and he's a former professor at the MIT Sloan School of Management.
Speaker 1:Sure. He's he's gonna have a busy day. That's a lot of
Speaker 2:Yeah. So he might know a little bit about AI. I'm not quite sure yet, but he's saying that professionals using AI will replace those that don't. He also advocates that while AI will display some jobs, the greater opportunity is in augmenting human work and creating new forms of value.
Speaker 1:Yeah. That's it. I mean, how do we adapt to this technology as well? We really I mean, you can see for ourselves, like efficiency using AI tools. I mean, I don't know what I would pick my efficiency increase, but if it's not 100 or 200%, I'd be surprised because everything is quicker.
Speaker 1:You know, if you if you have all the the skills, then it's really just, you know, getting to decision points faster and, you know, having a sounding board for ideas and so much value.
Speaker 2:Yeah. There's definitely you know, from running a business, I'm looking at it from just that marketing side, sales side, content creation side, you know, depending on who your audience is. You know, people are gonna be able to tell if it's AI slop. So I think the amount of noise is going to increase, and people will be able to tell if you don't put any effort into it. If it's all AI and your audiences, they're not simping for some AI, you know, girlfriend and there's actually like, they want some, you know, humanized content that really took some effort, you can use AI to help you get, you know, maybe spread the word or do things a little bit quicker, but you still have to be original.
Speaker 2:Even my, you know, my kids can tell immediately if it's AI slop and it's devalued almost instantly. I'm sure the same thing goes for content. If you see anything and it has the tell telling signs of AI has written it, they're like, I could just do this on my own or redo this research on my own. So again, you can use it for efficiencies, but that human creative element is still gonna trump all that other stuff that's there.
Speaker 1:And that is probably one of the tenants of being human is, you know, this focusing on the skills of like creativity and all the other sort of points that we've mentioned that make, we believe make us who we are. But, you know, we'll have to see where this road takes us.
Speaker 2:Yeah. So we know that they're here. And so I think the next thing is, you know, what will they actually change in the industry? This is where things get a little little spicy. You know, AI agents aren't just cost cutting tools.
Speaker 2:They're industry remodelers. And so with, like, tech and software, AI agents embedded across apps, Salesforce, ServiceNow, Microsoft three sixty five, you know, Google. ServiceNow, I think they said it's about 400,000 labors hours saved and 300,000,000 in productivity value. Yeah, it's pretty amazing.
Speaker 1:I can believe it completely.
Speaker 2:And then you've got about a billion agents promised by Salesforce by the 2025. That's quite a bit.
Speaker 1:That is. I wonder which data centers they're also going to be looking at running all these on because I think that's the next big thing is Hopefully it does
Speaker 2:a little bit better than their software currently is. Maybe 82,000,000,000. Sorry, Salesforce. Don't mean to dig at that.
Speaker 1:There are not enough data centers for them. So yeah, looking at then other industries like healthcare. I mean, already, I think they've been the creation of sort of virtual nurses and even these sort of triage agents. And I think the Mayo Clinic, which is really popular, that's integrated agents into its kind of more clinical workflows. And I mean, those agents are handling symptom collection, document summaries, scheduling, even mental health chats.
Speaker 1:It's right in there for healthcare.
Speaker 2:Yeah, and legal and finance, there's definitely some big ones out there that are already really helping with that, review contracts, summarize case laws, and then draft responses for law firms. And I have friends that are lawyers and they they tell me that if they know that you've used AI, but you haven't done your due diligence to check because it still hallucinates some things, They'll immediately dismiss it. And so you, you need judges are really on top of like, Hey, did you do your due diligence? Fine. You can use it to help you, but really check your work.
Speaker 1:Cause you don't know.
Speaker 2:Yeah.
Speaker 1:There was that funny case. It was probably a year or two ago when ChatGPT was really just on the scene, it was someone generating their whole sort of defense with claiming two hallucinated sort of cases that they were basing their conclusions off, which is really funny. Yeah. But now you've got more of the deep research, which is also agentic in terms of going out and finding details and building and reasoning you know, that I think would probably do a much better job.
Speaker 2:Yeah, definitely. And in fact, for financial services, you can have that fraud detection done in like milliseconds. The virtual investment advisors to help you with just the basic setup, and it can help you. It could probably follow you along your life plans and and immediately if there's anything dynamic that happens along the way, it can tell you, hey. Let's do let's move money over here or do this or that.
Speaker 2:So Mhmm. I mean, that's gotta be pretty powerful and any kind of regulatory compliance agents as well.
Speaker 1:Yeah. That makes sense. Then you look at other you know, you think of, like, retail and and manufacturing. You're starting to have these smart shopping assistants. So, you know, like, alright.
Speaker 1:I'm looking for this type of backpack and this color for this budget. And then can go to track all that down. Could be a much more intuitive process, I guess, than having to dig again through sort of different websites or even through big catalogs with some of the main shoppers. A few things like supply chain agents, being able to predict when there's stock outages or kind of reroute logistics based on specific disruptions. And I know there've been some interesting, I know you've heard of the hyperspace protocol that's being We had hypertext at the moment, but it's moving to hyperspace.
Speaker 1:And that'll affect things like logistics, which will be where everything is sort of coded in spatial dimensions. So things will know, all right, there's an ambulance coming along and it can then send out its route and everything through this sort of hyperstase protocol, cars can move out of the way, traffic lights can adjust. That's the next big sort of change in sort of internet architecture, which we've got to unpack in another. Yeah. That's,
Speaker 2:that's going be interesting with, especially as, self driving cars become more and more prevalent. I mean, I'm seeing more and more of those on the road just in Scottsdale this weekend. And so many of those Waymo cars are around. And obviously now Robotaxi Mhmm. Has taken over as well.
Speaker 2:And so we're gonna see that happening. And it'll be interesting to see, like, when they move all in one ecosystem to to be able to see that. It almost reminds me of the scene in iRobot where the self driving cars are going at, like, 200 miles per hour because they can because they're all, like, self they're all Exactly. From a high level network.
Speaker 1:Yeah. That's
Speaker 2:amazing. Can't wait. No more traffic. Exactly. Hopefully.
Speaker 1:No more driving. Just relaxing. Yeah.
Speaker 2:I'm curious. In video games, cheat codes let you skip months of grinding to unlock special abilities instantly. Have you ever wished for something similar for your software challenges? What if there's a way to instantly access twenty plus years of specialized expertise instead of developing it all internally? What if you could solve in weeks what might otherwise take months or years?
Speaker 2:Would you agree that most organizations faced a steep learning curve when implementing new software solutions? At my company, Impact, we serve as that cheat code for companies looking to transform complex software into intuitive experiences that users love and that drive real business results. Would it be valuable to explore, and how might this work for your specific situation? Visit impact.i0 for a free strategy session focused on your unique challenges.
Speaker 1:Then yeah. Things like manufacturing. So I don't know if you've heard of cobots.
Speaker 2:Yeah. We actually you know, that's one of our clients we worked with. They're corobots. So that's But are are
Speaker 1:are they cobots in in terms of the collab I guess they yeah. True. They are. Yeah. Yeah.
Speaker 1:Just had to get them referred. Yeah. Are you thinking
Speaker 2:more like a humanoid or anything?
Speaker 1:Well, I I guess, I mean, cobot being collaborative robot. I wasn't so aware of the term. And that's Mhmm. You know, anything designed to work alongside human work. So that's right.
Speaker 1:All the work we did there, is obviously cobots. But Yeah. Yeah. Which is anything sort of outside of a cage, you know, that's more sort of
Speaker 2:They work in conjunction with the human. They're safe. They won't they won't touch. They won't you know, if there's, any type of hit because back in the day, you'd have to surround them with, like, big cages. I remember working at Motorola, and these massive robots would help, program phones off a conveyor line.
Speaker 2:And so you'd have to surround them with safety cages. But then when we went to the client Yasukawa, they created a cob robots where it would immediately if there's a human, you could work next to it. There's no safety cage. If it touches you, it may immediately stop. So it's more user friendly to work alongside with that kind of semi automation situation, which is great, but we still had to program those manually.
Speaker 2:So to be able to have any type of AI agent that immediately like adjusts, reprograms itself, I mean, that's going to be pretty powerful to be able to do that.
Speaker 1:Yeah, exactly. And I think that's the key is agent managed. The fact that as you're saying, like there was these ones that are all sort of manual setup and opposed to saying, well, we've got one agent that's controlling 50 robotic arms or mobile carts that kind of follow workhouse workers around and sort of carry parts.
Speaker 2:That's going be interesting. That's gonna be pretty interesting too, because the fact that you could be so dynamic and changing and do it immediately. Because that's kind of the thing with manufacturing is, is like maybe a part is known as defective or you're switching, like there's a huge demand for a specific widget. Widget a versus widget b. So you've got to make a quick switch because, you know, all of sudden there's an advertising campaign and everybody wants said widget.
Speaker 2:You have to be able to just kind of push on the fly. Hey, we've got, you know, part a, part c, part d needs to go together. And so it could immediately do that and spin it up and or there's, you know, something from the from the supplier, it ran out and we have to switch it. There's, you know, different, size dimensions or whatever. Like, all these things could be done by AI.
Speaker 2:I mean, you know, being in robotics for about ten years before starting Impact, I mean, it's just I can't even imagine how exciting that is to be a robot robotics engineer and where things are going.
Speaker 1:Yeah, that'd be fascinating as well to see how they sort of handle the hybrid factories initially, until they're all robotics. Do you have sort of trackers on the human employees? And so the agents sort of go, all right, this is where each of the human employees are. You know, I can route things to, you know, supply them with something or, you know, work around where they are. It's gonna be really interesting to see how that goes.
Speaker 1:But then looking even further or, you know, for full automation is what's called another new term I hadn't heard, lights out factories. And that's, you know, thinking of traditional factories, you obviously need lighting, need climate control, and you're keeping everyone able, all the human employees able to work because they need to see what they're doing, they need to be fairly comfortable. Whereas depending on the product, a lights out factory doesn't need any of those. So humans certainly need to be on-site. So you've got this running 20 fourseven, lights off, climate control off, HVAC systems can either be reduced or turned off completely as well, and AI agents obviously controlling these completely.
Speaker 2:Yeah. Google Gemini on my phone just woke up and was asking me a question. So it's like, no. We're not that smart just yet. We want you to lead you to believe we're not that smart and listening.
Speaker 2:Yeah. And it's not just factories either. It's, imagine a lights out hedge fund run entirely by AI agents. It operates twenty four seven. It's making billions of trades based on data no human that no human can process.
Speaker 2:No employees, no board of directors, and its only goal is to maximize profits.
Speaker 1:Crazy.
Speaker 2:I know. What happens when its goals conflicts with market stability?
Speaker 1:Yeah, exactly.
Speaker 2:I mean, there's already human beings that are influencing, know, for the crypto world, influencing trades, market manipulation, especially when you hold a large amount of a certain stock or crypto coin, you can easily do that. Or when, you know, when two of these AI funds start a financial war against each other. You know, we're not just talking about job losses, but we're talking about creating autonomous economic entities that we can't control.
Speaker 1:Yeah. Interesting. Yeah.
Speaker 2:So I think the next kind of big thing is what changes are for the economy and the job market. You know, if we jump into a time machine and look ahead, you know, what's happening or what's going to happen in the next two, five and ten years?
Speaker 1:I think, again, if you're sitting there going like, Oh, agents here, I'm sure they'll run out slowly. Maybe my kids will be working with more in the future. I think it's happening very quickly. So if we look at the next sort of two years, like going up to 2027, most jobs on a computer are going to be impacted. So agents, you can think of them more as going to be your coworkers for millions of people.
Speaker 1:And you're going to see entry level roles pretty much disappear or at least start having a major shift to your data entries, your schedulers, those sort of things are going to be impacted. And we're really going to start feeling the mass effect of agents with productivity and just seeing the boost show up. If we fast forward to five years, getting to kind of 02/1930, you can have the end to end processes handled by agents. So think of anything you're doing now, like you got an insurance claim, you're taking out a loan or there's a procurement process, completely automated. The mass market for these personal agents are going to be almost your second brain.
Speaker 1:Consumers, employees, students, everyone's going to have access to a really powerful, always on intelligent digital assistant. And it's going to augment your memory, your decision making, your productivity, and potentially, we can't measure yet to some extent your creativity or at least be a facilitator to allow you to be even more creative and allow you to use tools that unlock more of your creativity. Like we mentioned before, lower barriers to entry as well. You're going to see a lot more startups and there's going to be a lot more competition. What do we know from that?
Speaker 1:You know, price reductions, potentially new companies appearing and, you know, putting traditional slow moving ones out of business. And I think from the PwC McKinsey, they're projecting, say, like a 13 to $15,000,000,000,000 addition to the global GDP due to this AI, which is huge. So then jumping the furthest ahead, if we look ten years out to 02/1935, we're going to have almost autonomous companies or autonomous departments, which is mind blowing. So you think a department that may be filled with people today. So look at a call center or all these people replaced by AI with really good and I don't know whether you saw, I think GPT's got its new voice model coming out, which its current one is impressive, but that's being sort of tabled for, you know, their new voice model that's coming out September.
Speaker 1:And you can just see that the shifts already are massive. So, yeah, it's going to be, looking at the future, if you look at a company, you probably find 90% of the work is going to be done by agents.
Speaker 2:I mean, I think they said that they anticipate the one person kind of first billion dollar company by a single person happening within the next ten years. So that should be.
Speaker 1:Yeah. Yeah. And, you know, all that knocks on to a massive workforce shift. So, you know, you've got to look at creative, strategic and human roles only, you know, so creative work, obviously, designing, storytelling, innovating, writing original content, you know, that's still something that, you know, that humans are gonna be able to do. So examples are kind of, you know, UX design.
Speaker 1:So hopefully, we're in it for a while. Yeah. Like we have been. We're one of the last to go product innovation, brand strategy, that sort of thing. You know, there's got to be more emphasis on strategic work as well.
Speaker 1:Know? And these, again, are skills that you want to foster in yourself, in your kids, in your students. Creativity, strategic thinking, high level planning, decision making, evaluating scenarios. So things like business strategy, market positioning, competitive analysis, all things that are going to be more driven by human. And then obviously the more human centered work.
Speaker 1:So things that we have that are unique to us, empathy, being able to negotiate, building trust and ethical reasons. So think of areas like therapy, leadership, teaching, client relationships. And then also they're going to be AI aware roles. So roles that sort of supervise, train and manage AI systems. Examples of those would maybe be, you know, almost the AI integrators that we were talking about earlier, AI product leads, AI workflow designers or, you know, prompt engineers, which, are already sort of becoming more and more valuable.
Speaker 1:And I think there was a quote from Accenture that or an estimate that AI could double growth rates in some countries in the next ten years, which is amazing.
Speaker 2:Wow. And I think this brings up a terrifying question that economists and philosophers are starting to whisper about, that's, you know, what about the people who can't reskill? There's a historian, I'm gonna butcher his name, but Yuval Noah Harari, he warned that about the rise of the useless class. And so it's a population rendered economically irrelevant by AI. It's a brutal term, but it forces us to ask, if an agent can do your job and can't transition into a creative or strategic role, then what is your value in the economy?
Speaker 2:This is the uncomfortable multi trillion dollar question nobody has an answer for just yet.
Speaker 1:That it is, that it is. Even when we shift to see more actual humanoid robots in the workforce as well, Everyone that is working on production lines or cleaning or sweeping the streets or anything like that, it's immediate effect. I was trying to think, especially in countries where it's obviously a luxury to have someone service your house and do things like that. The question is going to become, well, I'm paying someone X amount per year. When I'm paying them in a year, I could potentially buy a humanoid robot and that has a six year lifespan or a ten year lifespan.
Speaker 1:Is it up to us to say, no, we're not going to do that? But then when you do the math, you're like, well, I can have someone that could fetch me a cup of coffee at 1AM. That's amazing. Whereas you can't get that with, you know, someone that maybe comes in and, you know, helps do a bit of cleaning for a few hours a day. So there are going to be some tough decisions.
Speaker 2:It gets into, you know, the movie, the the cartoon, WALL E, where you got all these people just floating around in their little thing, watching TV, eating all the time, and that's all they're doing. They've got the the screen in front of them and just floating around and and and and just doing that. I mean, that's you're gonna have a class that is maybe doing a lot of, you know, just nothing. And, but you also, you know, if you can get the replicators from Star Trek where it just makes everything for you, then you don't have to worry about, like, food, water, and all the basic necessities. And maybe AI can figure that out.
Speaker 2:Hopefully, we'll get AI not only to help replace the menial tasks. I mean, if you're doing something that is just not satisfying, but then AI allows you to create a life that is a lot more satisfying and doing what you want. Mhmm. Hopefully, we can use leverage it for that as well.
Speaker 1:Yeah. Agreed.
Speaker 2:So I think the the next thing is obviously there's so many different software agents out there. And so it's kinda like the battle of the agents. So we we wanna take kind of a more comparative look of what's out there. We've talked about, you know, how they're changing industries, but who are the key players in this space? And so let's dive into kind of a head to head comparison of the leading AI agents out there today from today.
Speaker 2:I mean, this can change, you know, in the next week or two, but from today, which is, you know, we're looking at August 2025.
Speaker 1:Yeah. It's, I mean, it's definitely a crowded field. It's, you know, I think everyone's trying to one up, you know, the other company and especially create that hype to, keep everyone interested. And, you know, like I was mentioning, we just saw GPT-five release. But there are a lot of major releases as well.
Speaker 1:The big names are OpenAI's ChatGPT, Google's Gemini, who's listening to you actively on your phone. We kind of probably don't know, maybe we need to pause to see whether there's a comment. Anthromix, Claude, you've got XAI's Grok, and you've Manus as well. And I mean, even more, like we've spoken about DeepSeek, you've got a lot of different open source models as well.
Speaker 2:So I think let's start with what makes them different and how do these agents stack up against each other?
Speaker 1:Yeah, mean, of them is different. I think if you've been involved enough to look at the for anyone listening, looking at the benchmarks, I think whenever you see the new models come out, they sort of have benchmarks with reasoning and context windows and and all different things. You've got if anyone who hasn't been or wants more information, got ChatGPT, which is probably what most one of the most popular ones, and it's, you know, known for its just it's really good at, you know, its conversational abilities and it's obviously got a growing and and quite a large user base. You got, you know, Google's Gemini. It's obviously you've got the advantage of being deeply integrated into Google's ecosystem of products, but I always feel it was kind of slow, slow to the start, and it seems to be sort of behind the curve in terms of catching up to things like ChatGeePut tea.
Speaker 1:Then you've got Anthropix Claud, which is often praised for its focus on things like safety and its large context window, meaning you can just take in and process a lot of data, which is obviously a great task, or great for tasks involving things like long documents. And you've got Grok, you know, from x AI. It's obviously one of the the newer releases. And, you know, I think, like, everything that Elon Musk kind of touts, it's got that more sort of, rebellious streak and obviously allows to, you know, have sort of access to real time data from X as well. And I think that was probably I mean, keep me honest here, Mikoto, probably one of the reasons he, he acquired X as well to sort of mine that data and those interactions.
Speaker 1:Oh, sure. A lot of sort of training material. And then, yeah, there's Manus as well, which has been designed from the ground up to be you know, what's really a true agent, you know, capable of not just talking, but, you know, doing things across all different applications.
Speaker 2:Yeah, so I think it's not just about which one's smarter, but which one's really right for the job. I use all of them. And there's definitely some some that are better for, you know, whether it's coding or just research to ones that are better at writing copy, things that kind of more conversational. I I just noticed with certain things for Grok, it it's really good at kind of more of that conversational interaction and and what you should say. I used it to negotiate.
Speaker 2:You know, I was looking to buy a, used car and and, you know, negotiation. Like how do you do that and what should you say to the dealership? And it was telling me step by step what to say, what to research, going back and forth. I mean, was pretty amazing that Rock was doing that.
Speaker 1:That's cool. Yeah.
Speaker 2:You know, Manus is really great at the brute force type of, tasks to to really you know, I had to do like, I had VO Google VO do an animation sequence
Speaker 1:Mhmm.
Speaker 2:From a single character, like, picture and have it do a single animation sequence. And then but it doesn't create the JavaScript for what you need. And so I had Manus actually go in, and it created a 120 frames of this animation and created the java JavaScript file for it. Wow. And you could manipulate the speed and everything and created controls so you can actually see how fast or slow you can make it.
Speaker 2:I mean, it was pretty amazing.
Speaker 1:Yeah, it is. What did they have with the GPT one recently? Just looking at the GPT five where it rendered, I think, three d castle that was actually a video game that was just described these little characters, things were firing at them, really impressive, like just from sort of chatting to it and building it.
Speaker 2:It's amazing. With all this hype, there's also a healthy dose of skepticism. There's a recent report I came across from MIT that found a staggering 95% of generative AI pilots at companies are failing. So we really want to know what's going on here.
Speaker 1:That's a decent question. I think it's a million dollar question because we've seen only those funny memes, the fact that, Oh, it's AI. Companies will just put AI in everything. The users are like, oh, we didn't ask for this. Like, you're getting AI, just take it.
Speaker 1:That comes down to, I think what we're The processes we follow impact as well, not basing things on an assumption, but what is the real value that a technology is really addressing? And if you can identify that, I think you can make a much more successful product because you're actually working on a tangible result of increasing something opposed to, well, everyone else is putting this in there. So it's gotta be good to add this in. And that's probably where there's that huge gap between, you've got a great demo, but have you got a real world scalable solution? And I think it's understandable that people jumping on the bandwagon would find that companies struggle to integrate agents into their existing workflows and actually prove a real kind of return on investment.
Speaker 2:Yeah. It sounds like we're in the the trove of delusionment for AI agents. And we've seen the initial hype, and now we're facing kind of the harsh reality of implementation. I mean, I've I've seen Claude where I ask it to do research on certain people on LinkedIn and it'll make up completely different employment, you know, where they worked at, what company it is. And I'm like, it's right there in front of you.
Speaker 2:You're telling me the wrong thing. And yeah, you still have to double check a lot of things. It's just weird that it just makes up things even though the data and information is right in front of it.
Speaker 1:Yeah, exactly. The technology is not the problem. It's just, has there been the correct oversight in the project? And are you getting that technology to actually deliver something that is valuable? The one MIT report, I think you're mentioning, highlighted that obviously many of the failed pilots lack those clear business objectives.
Speaker 1:It's just like, all right, technology and let's run with it and see what it takes us. So also the necessary talent to see them through. Do you have someone that is aware, is thinking about the future of this technology, not where it is now? Because a lot of it, coming back to the points earlier, is thinking about, have you got your foundation right? If you're wanting to plug in agents, but you don't have the infrastructure set up or there's no roadmap plans to have this, you're going to hit a wall pretty quickly.
Speaker 1:It's avoiding that classic case of a solution looking for a problem.
Speaker 2:Yes. What's the takeaway for our listeners? Are AI agents doomed to be a failed experiment?
Speaker 1:No. I mean, that I feel strongly about. There's so many leaders that are seeing and recognizing that this is the way of the future. And it doesn't matter what your experience has been, they are going to increase in capability and doing more things, being able to integrate more things, and we're just going to see this. This is a part of our future.
Speaker 1:So the wake up call is probably that those 5% of successful AI projects are those that have been strategic. They're not just throwing this technology at a solution. I mean, at a solution, at a problem. And they're obviously redesigning processes, they train with people, and they're starting with everything, starting with a clear business case. These may seem like a revolution, but it's more about the evolution of this and building, getting that foundation in place, growing with the technology and unlocking all the new features as they become available.
Speaker 2:Great. I think this is a a good good place to kinda close out and, you know, ask the question, what do you do now? The agents are not just coming. They're here, and and they're coming in faster, and they're evolving. And so think of them like electricity.
Speaker 2:They're foundational, they're silent and powering everything very soon.
Speaker 1:Yeah. And I think just start simple, like look at a single pain point. What can the agent assist that is going to be valuable for your users? Start forming a team or even a culture of AI champions. Share within your organization how are these tools working for you?
Speaker 1:How are you using a tool and someone else may be using it differently? Share that and get an understanding of how this is actually going to change your industry. And start a pilot project. Start just again, that simple single pain point, just work on that and scale from there and just have measurable, trackable metrics so you can kind of quantify the impact that you're making.
Speaker 2:Yeah. I think it's very important, especially when you're trying to implement this AI within organizations that are very technical and it becomes this black box of magic. And if something breaks or starts to push them in a direction that's incorrect or wrong, they won't know that. And they won't know to to how to fix it or change it or how to correct it. They just know that they are being driven in a certain direction or given certain information.
Speaker 2:And this can cause a lot of problems, especially, you know, think about AI in schools. If they're giving the wrong information and kids are being affected, you know, that can be a problem. Or if it's health, you know, that's another thing. If if AI is telling you to do one thing, but then it's you're supposed to be doing another, you could get sued. I mean, there's there's a lot of repercussions by I call it the cliff notes of reading a book.
Speaker 2:You know, you're this is that on steroids where, yeah, it's gonna get you to the what the main point is faster. But if you don't know how it got there, is that point correct? So, you know, these are the kind of things that I think from our organization, what we do at Impact, we really try to understand the strategy, understand the users, really understand what is it there because, again, don't be so feature centric. Yes. It's tech, but you have to understand your users' empathetic goals and what are you trying to do to make things, you know, better for them in the long run.
Speaker 2:So I think if you're a leader, this is your Jarvis moment. You know? Will your business be Ironman or will it be obsolete? So again, thanks everybody for tuning in. Just subscribe, share, and hit that like button, and, yeah, start preparing today.
Speaker 2:And it's been a good one. Have you ever played a video game and discovered a cheat code that instantly unlocks abilities that would have taken months to develop? I'm curious. What would it mean for your business if you could access a similar cheat code for your software challenges? What if you could bypass months of trial and error and immediately tap into proven expertise?
Speaker 2:You know, I've noticed that many organizations spend years developing specialized software expertise internally, often through costly mistakes and setbacks. Would you agree? That's a common challenge in your industry as well. At my company, Impact, we function as that cheat code for companies looking to transform complex software into intuitive experiences. Our clients gain immediate access to twenty plus years of specialized knowledge and the experience of launching hundreds of software digital products in many different industries without having to develop it all internally.
Speaker 2:You might be wondering how does this actually translate to business results? Well, companies we work with typically see go to market times reduced by up to 50%, their overall NPS scores rocket up, and their product to development team's efficiency significantly improved. Instead of struggling through costly mistakes, they accelerate directly to solutions that work. This is why organizations from startups to Fortune 500 partners with us for years. We consistently help them solve problems in weeks that might otherwise take months or years.
Speaker 2:If you're responsible for digital transformation or product development, wouldn't it make sense to at least explore whether this cheat code could work for your specific challenges? From boardroom ideas to code, this is what we do best. Visit our website at iiiimpact.io. You can see the link below to schedule a free strategy session. It's just a conversation about your unique situation, not a sales pitch.
Speaker 2:And you'll walk away with valuable insights regardless of whether we end up working together. Thank you.