Episode Categories:
Resources:
Join the discussion on LinkedIn: Got something to say? Let us know on LinkedIn and network with other AI leaders
Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
Connect with Jordan Wilson: LinkedIn Profile
Start Here Series in our Inner Circle Community: Join for free access
The 11 AI Tool Categories: An In-Depth Roadmap for Business Leaders
AI adoption in business is no longer about chasing every shiny new tool. Instead, understanding the landscape — specifically, the 11 key categories of AI tools currently defining enterprise operations — is fundamental for those who want to maximize return on technology investments and maintain a competitive edge. Below, each category is unpacked to spotlight what matters for businesses, why most only leverage a fraction of AI’s potential, and how to target AI decisions for practical outcomes.
Text Reasoning Assistance: Foundation for Enterprise Knowledge Work
The first parent category — text reasoning assistance — includes tools like ChatGPT, Gemini, and Claude, built to transform typed queries into actionable recommendations and insights. Recent advancements mean these tools offer economically viable outputs indistinguishable from expert human work, provided users are patient enough to let models complete their reasoning cycles. Choosing the right model and workflow is critical; opting for speed over depth often results in missed opportunities for superior deliverables (11:19). Businesses using only stripped-down, instant-response tools miss the full advantage of nuanced reasoning.
Multimodal AI Platforms: Unified Solutions Across Modalities
Multimodal platforms, typified by Google Gemini and Microsoft Copilot, go beyond text inputs and outputs. These solutions accept and produce images, videos, code, and more, making them ideal for teams seeking centralized AI functionality (12:44). Google leads this space — Gemini Pro 3.1, for instance, can “see” and analyze uploaded videos in detail, though such tasks are resource-intensive. Most users only tap text functions, bypassing broader capabilities that could streamline workflows across departments.
AI Search and Deep Research: Verifying Data and Industry Insights
AI-powered research and search tools are essential for fact verification, market analysis, and strategy. Products like Perplexity, Notebook LM, and u.com are complemented by sector-specific tools (legal, academic, etc.) that integrate deeply with industry data (15:15). The source and recency of information are crucial here — leaders must understand whether their outputs are based on static training data, user uploads, or live web browsing (17:16). As agent-based browsing becomes standard, preparing business content for AI discoverability will soon outweigh traditional SEO.
Voice and Speech AI: Capturing and Deploying Oral Knowledge
Voice and speech AI covers both speech-to-text and text-to-speech. Tools such as Eleven Labs, Fireflies, Otter, and Mistral’s Voxrel are making voice interfaces standard in meetings, training, and global communication (20:13). With institutional knowledge at risk from layoffs and retirement (“silver tsunami”), capturing expert discussions and transforming them into accessible IP is gaining strategic importance (22:03). Moreover, AI’s improved natural language understanding allows teams to operate up to four times faster by speaking rather than typing (22:56), and synthesized output shortens comprehension for those who prefer listening to reading (23:22).
Image Generation AI: Rapid Visual Asset Creation for Multiple Use-Cases
Image generation tools, powered by diffusion models, transform text prompts into detailed visuals. Leading players — Midjourney, Flux, Ideogram, OpenAI’s GPT Image 1.5, Google’s Nano Banana Pro — are not only for designers. Slide creation, infographics, and product visuals are accessible to anyone needing enhanced marketing, sales, or learning materials (25:33). Businesses should note that younger audiences increasingly prioritize visuals over text (26:06), and standalone image generators are being eclipsed by multimodal tools integrated across platforms.
AI Video Generation: Efficient Content Production and Simulation
Video generators (e.g., Google VO 3.1, Runway Gen, Seed Dance, Pika 2.5) rely on models that understand physics, character consistency, and environmental context, producing output indistinguishable from original footage for most use cases (28:35). The technology is intensive and can be costly, but the gap between amateur and professional content is closing quickly. Effective use cases include social media production, training materials, and even simulations for product development.
Music Generation AI: Automated Audio Solutions for Content and Branding
Tools such as Suno V5, UDO, Eleven Labs, and Google Lyria 3 Pro enable prompt-based creation of fully mixed songs and custom audio assets (30:35). These solutions support video content, advertisements, and royalty-free branding elements for businesses looking to avoid copyright headaches. The reported revenues — Suno alone at $300M — underscore demand from both creators and enterprises (30:53). Granular customization (altering instruments, vocals, etc.) allows teams to differentiate audio assets quickly (31:33).
Design and Visual Content AI: Complete Artifacts from Text Inputs
Design platforms (Gamma, Canva AI Magic Suite, Beautiful AI, Symphyzia, HeyGen, Google Stitch) merge layout, images, and content into slides, websites, avatars, and app prototypes (32:13). Microsoft Copilot’s PowerPoint integration now automatically generates slide decks including custom images. The ability to build digital twins and export visual assets directly into code accelerates both marketing and product development (34:08).
Vibe Coding App Builders: Nontechnical Teams Creating Business Apps
No-code, natural language app builders (Lovable, Replit, Bolt, Vercel’s VZero, Base44) empower marketing, product, and entrepreneurial teams to ship live apps without direct development resources (34:50). Previously limited to disposable prototypes, these platforms now support full-stack applications with user authentication and management (36:01). Investment in this segment is driving traditional software solutions into obsolescence (36:33).
AI Coding Copilots and Agents: Automated Software Development and Desktop Control
AI coding agents (Cursor, GitHub Copilot, Claude Code, Windsurf, OpenAI Codex, Google’s Anti Gravity) handle codebase comprehension, bug fixes, feature shipping, and general desktop actions (36:48). They extend to tasks beyond software — automating any command accessible via terminal. Knowledge workers outside engineering roles should upskill in using these tools, as they accelerate automation across all desktop workflows (38:29). Integration with local files and apps makes this category pivotal for future agent-led operations.
Autonomous AI Agents: Delegating Tasks via Agentic Browsers
General and specialized agents (Manus, Genspark, ChatGPT Atlas Browser, Perplexity, Claude Cowork, Microsoft Cowork) are now able to plan, sequence, and execute work via browser or desktop environment (40:36). This category is dominant among technically adept workers seeking to offload repetitive processes and manage complex schedules proactively. Agentic browsers will eventually replace reactive chatbots with scheduled, goal-oriented task completion (41:14).
Building Strategic AI Stacks: Prioritizing Tool Categories for Business Outcomes
Understanding AI tool categories rather than accumulating dozens of platforms enables organizations to build a stack aligned with specific workflows. Leaders should select two to three categories matching their business priorities, maintaining high proficiency in AI search, multimodal, coding copilots, and agents. The next 18–36 months will see agent-driven work as standard practice, providing early adopters an operational advantage (46:27). For example, marketers might pair search with design tools; entrepreneurs can combine app builders with agents (46:55).
Organizations must avoid “shiny object syndrome” by choosing tools that fulfill nagging requirements in critical workflows. As most AI products fit under these 11 parent umbrellas, focusing on ones with real business impact ensures sustainable value rather than superficial upgrades (48:38).
Conclusion
Mapping the AI landscape by parent categories clarifies which toolsets provide the most business value. The full map gives decision-makers not only a guide to the current spectrum of possibilities but a strategic filter for what to integrate next. Early proficiency and targeted stack creation now are poised to set the most operationally advanced organizations apart for years to come.
Topics Covered in This Episode:
- 11 Essential AI Tool Categories Overview
- Text Reasoning Assistants Explained
- Multimodal AI Operating Systems Breakdown
- AI Search and Deep Research Tools
- Voice and Speech AI Tech Trends
- AI Image Generation Diffusion Models
- AI Video Generation & World Models
- AI Music Generation Tools and Players
- Design and Visual Content Automation
- Vibe Coding App Builders Functionality
- AI Coding Copilots & Local Agents
- Autonomous AI Agents and Browsers
- Building a Competitive Personal AI Stack
Episode Transcript
Jordan Wilson [00:00:16]:
You shouldn't use as many AI tools as I do. It's actually a recipe for disaster, and I think that shiny object AI syndrome is one of the biggest problems in today's enterprise landscape. But although I don't think most business leaders should be using 20 plus AI tools or systems every week, I do think it's imperative to understand the landscape. And obviously, it's ever changing because in just about any field, sector, or modality, there's probably a juggernaut unicorn AI company that has created a top notch AI tool that's redefining work in that given space. So on today's show, I'm going to give you a lay of the AI land. Because I've been there. I've wasted thousands of hours. So you can just learn from me as I tell you what types of tools are good for what reason.
Jordan Wilson [00:01:10]:
So on today's show in volume 16 of the start here series, we're going to walk from chatbots to AI agents and everything in between and quickly and simply lay out the 11 different AI parent categories where tens of thousands of AI tools mostly all fall under. Because, yes, there's a lot more to AI than chatbots and agents, so let's just go ahead and connect all the different dots and categories in between. Alright. Let's get into it. If you're new here, this is everyday AI and our start here series. So here's the big picture of what you need to know for today. Well, there's hundreds of AI tools literally launching every single week, and most people feel overwhelmed. Well, I do too.
Jordan Wilson [00:01:58]:
I try to keep up with it, and I don't recommend you try to keep up with the hundreds or thousands of tools, in the AI space that are released every week because most of them are, well, kind of garbage and don't hold a lot of utility. But almost every single one of those tools falls under, you know, maybe 11 of these parent categories that we've created for you today, And there's actually very few exceptions. And, yes, some of these categories are extremely, broad, but I did that for a reason. And I think once you understand these 11 categories, you can kind of stop chasing tools and start building a stack that makes sense for you. Because like I said, even within all of these individual categories, the functionality is changing constantly. So, you know, you might say, oh, there's 50 categories, but, they're changing so much and adding so many new features and functions. Well, it's I think it's easier to just understand just a couple less than a dozen of these categories. So this episode is your map to every type of AI tool that exists right now.
Jordan Wilson [00:03:00]:
So on today's show, stick with me for the next twenty ish minutes, and you're going to learn why, well, well, why you're probably only using one or two of these AI categories and you're probably falling behind because of it. And again, you shouldn't use tools from all of them. You're also gonna learn the single framework that makes sense of every AI tool from ChatGPT to Cursor to Suno and how to build a personal AI stack that matches your job before your competitors find out. Alright. Welcome to our start here series. This is the everyday AI essential podcast series to both learn the AI basics and to double down on your AI knowledge. I created this series, well, because I didn't have a good answer when so many people said there's, like, 700 plus episodes. Where do I start? Well, you start here with the start here series.
Jordan Wilson [00:03:49]:
If you are picking this up midway through, that's okay, but I highly recommend you listen to all of these episodes in order. They're short ish. They average, like, twenty nine minutes or something like that. And you can go to starthereseries.com. That's going to give you free access to our inner circle community, and you can go check out our start here series, channel in there where you can go listen, read, and just connect with other people who are going through this series in order. So if you have any questions, make sure to hit me up there after you join the Inner Circle community. And if you missed our last start here series episode, we talked about how everything is fake and how your company can leverage human expertise properly and fight AI work slot. That one was a good one.
Jordan Wilson [00:04:35]:
It was actually a fun one to do. Alright. But let's get into the roadmap here in 11 categories explained. Here they are right away. Number one, text reasoning assistance. Number two, multimodal AI platforms. Number three, AI search and research. Number four, voice and speech AI.
Jordan Wilson [00:04:52]:
Number five, image generation. Number six, video generation. Number seven, music generation. Number eight, design and visual content. Number nine, Vibe coding app builders. Number 10, AI coding tools. And number 11, AI agents. Alright.
Jordan Wilson [00:05:07]:
And let me just tell you this. I took a little bit of liberty and played around with these how to categories all these for tools. Right? Because as an example, something like Gemini, well, they fit in basically all of these. Right? And in theory, we could just have combined number one, two, and three into large language models. But there are, literally thousands of actually pretty good tools, right out there. Right? Even though I think that 99% of them are pretty garbage, there's literally hundreds of thousands of AI tools. But so there's actually thousands of great ones that fit in to these, 11 different categories. And so I did kind of take a little bit of leeway, and maybe this will make sense as we unpack this a little bit more.
Jordan Wilson [00:05:52]:
But you might be wondering, well, what's the difference between a text reasoning assistant and a multimodal AI platform? So, yes, in the beginning, you might just be saying, well, isn't that just chat GBT or can't Gemini do, you know, nine out of these 11 categories fairly well? Well, yes. But as you'll see as we unpack this a little bit more, there are some, unique, tools that only fit in maybe one of these categories. And I think as we start to to think about how you can practically apply these different categories and kind of create that, you know, essential kind of tool category stack, it's important to understand the differences. Yes. There's crossover, but I think that we need to kind of properly go through this. So don't forget. AI is not new. Alright.
Jordan Wilson [00:06:40]:
And I know if you're a long time listener, you're like, okay, Jordan, I get it. Right? But I'm trying to keep this very basic, and I think this is gonna be one of those episodes that a lot of people listen to, and I'm gonna point people toward, because, yes, there's there's always an AI tool that can do something. Right? And even in these 11 different categories, all of these also have, sector specific tools that really shine. Right? So as an example, there's, you know, even image generation. There's image generation tools that are essentially just rappers of, you know, Google's nano banana or ChatGPT's image tool for different sectors. Right? So, also keep that in mind as you're using these different tools. A lot of them are just based off of, well, one of the core platforms from one of the big players. Not all of them, but a lot of them are.
Jordan Wilson [00:07:28]:
So also keep that in mind. But also keep in mind, AI is not new. Right? I wanna give a very, very quick, kind of, recap of the last sixty years. This would be the fastest suburb. Right? But AI has been around since the fifties, and the first AI chatbot was actually Eliza in the sixties. And, kind of the big boom of AI was actually in the eighties. Right? A lot of people think it started with Chai GBT. No.
Jordan Wilson [00:07:54]:
You had expert systems in the eighties. But, essentially, these were very rigid, if then else logic, pieces of, you know, old school AI algorithms. And they weren't really meant, for general use cases. Right? A lot of them, as an example, you know, you could look at, an old school artificial intelligence in banking. And all it was, it was a very complicated, rules based decision tree. Right? That when it maybe an inquiry came into one of those old computer systems that were, you know, the size of a school bus and literally only did one thing very slowly. All they did was, kind of traverse this path of this very rigid if then else logic. And, usually, if there's one thing wrong, right, one extra comma, one extra space, the entire thing broke.
Jordan Wilson [00:08:43]:
And that's obviously very different from today's generative AI. Right? And a lot of that, happened because of a very famous, paper essentially, from Google called Attention is All You Need in their 2017 transformer architecture. And that really unlocks, well, AI that's not as rigid. And that's what, you know, kind of Chad GBT exploded this scene, which is generative AI. So, you know, when we had for decades AI that was very rigid, now it's moved into gen generative reasoning and, intuition. And I think that chat GBT really helped launch this new category. Yes. There were other generative AI tools before chat GBT, you know, maybe by a year or two.
Jordan Wilson [00:09:31]:
But I do think that it was the, surgeons of ChatGPT that popularized, AI tools. Right? It brought in a lot of funding that really helped pave a lot of these different categories in all the different tool, kind of parents' umbrellas that we're gonna be going over today. So let's talk about category one. That's just text reasoning assistance. This is you type a cat you you type a query into chat g b t, into, you know, in Propic Claude, into, Gemini. You type in text, you get text out. And, well, who is this for? It's for everyone. This is kind of the gateway drug of AI.
Jordan Wilson [00:10:08]:
Right? But it's important to know by default, everything is reasoning now. So if you maybe were a a heavier user of AI in 2022, 2023, and you're like, this thing stinks. I'm never using it again. And you're just waking up now today. Yeah. The models are a lot better. They are smarter than most human experts, when it comes to creating economically viable work. But they do reason by default.
Jordan Wilson [00:10:33]:
They think by default. Right? So some of the models are slower, and I also right? We've gone over this multiple times in the start here series, especially when you are using tools in category one, which is thing is where a lot of people spend time. Do not sacrifice speed for using the right freaking tool. Right? Don't use, like, an instant version or a non thinking version of any of these tools. Right. We literally have insanely helpful technology that can create outputs, deliverables, artifacts that are indistinguishable from human experts. Yet so many people aren't getting that just because they're impatient. Right? And they don't wanna maybe wait three, four, five minutes for a response.
Jordan Wilson [00:11:23]:
And instead, we'll, well, they'll just say, I'm just gonna get this fast one, or I'm gonna use this model that doesn't think or doesn't take as long. So number one, stop doing that. Right? There's a a reason, why the, people who are getting the most out of these models are, you know, usually have multiple screens and, you know, multiple platforms because you have to wait. You know, you have to wait. You have to be patient. You have to have a good workflow. This is why I usually recommend, yes, pick your AI operating system of choice, but have backups, right, or be working on multiple projects at once. You do have to be a little bit, you know, okay at, multitasking and bouncing between different projects because you should be taking your time when working in the category one, which is our text reasoning assistance.
Jordan Wilson [00:12:07]:
Alright. Category two, this is multimodal AI platforms. And for the most part, it's the same quote unquote tools in category one as they are in category two. And I think it's actually important to differentiate these, use cases because I would say the overwhelming majority of the billions of people that use generative AI every single week are using the same tools, but they're only using them for, for the most part, category one, text reasoning. They're not using them as a multimodal AI platform. Alright. So what that means is, well, multimodal in, multimodal out. Alright.
Jordan Wilson [00:12:44]:
So the main players here are the same as the main players in number one. So that's ChatGPT, Gemini, Claw, and also Microsoft Copilot. And they really are expanding into multimodal operating systems. So who is this category for? Well, anyone that wants one AI platform, and across multiple modalities. So, I think Google is by far the leader in this space. And as we go on, you'll probably see Google tools mentioned more than any of the others because they are absolutely dominating in multimodal AI. But what does it mean for a multimodal, AI platform? Let's say Google. Right? A lot of people don't know in Google's AI studio, which is a kind of more customizable or, you know, developer friendly version of Google Gemini, you can upload videos.
Jordan Wilson [00:13:34]:
Right? And Google Gemini 3.1 pro doesn't just, you know, transcribe videos. It can see them. Right? Yes. It's very token intensive. You know, you might burn through, a little bit of your your budget, but it can actually see and understand the videos. So that's the difference, between and my thing is important to have two different categories for a literal multimodal input. And a lot of people still, you know, don't take advantage of that in a chat gbt cloud copilot, right, that it can, you know, upload screenshots. You can, you know, output, you know, depending on which system.
Jordan Wilson [00:14:13]:
Right? You can output code. You can output graphics. You can output images. In some, you can output video. So it is important, even though a lot of people spend most of their time in category one, those the three or four big boys in the room, they play in category one and two, which also leads me to, yes. I know I said we could, in theory, combine categories one through three because category three is AI search and deep research, and you have a lot of the same players, but they are kind of different products. Right? They have their kind of dedicated, you know, features or modes. So in Chatt GBT, you have deep research.
Jordan Wilson [00:14:52]:
In Gemini, you have deep research. In Claude, you have, extended kind of the extended reasoning, kind of, version of this to take longer and do, deeper research. And there is also a research tab inside clause. So it is kind of a different platform, but then you have a completely new set of players in here as well. One of the ones that most people are probably most familiar with is is perplexity. You also have, you know, notebook l m, I think, technically fits in this, category as well. Although, you know, there's completely so many different categories where you could place notebook l m. You have tools like, you know, u.com, and then you have literally hundreds of great industry or sector specific researchers as well.
Jordan Wilson [00:15:39]:
So who uses this? Well, I think your average everyday professional that just needs to verify facts, but also researchers, analysts, strategists. And like I said, some of the more, you you know, especially on the academic research side, legal research side, Some of those tools are probably gonna be a little bit more familiar, for people working in those sectors. So I'm not gonna confuse everyone with all those names. Right? The the Harvey of the legal worlds. You know, there's so many on the academic side. But essentially, AI, you need to understand. And I think people always skip over this. The source of your answer is extremely important.
Jordan Wilson [00:16:15]:
So when you talk about large language models having AI search and deep research, it's it's it's actually not quite a novel concept. They've been doing this for multiple years, but you have to realize there's really three different main sources, for where a large language model gives you its answers. Right? Number one is its own internal training data. So as an example, if you're using, an offline, you know, model that you downloaded and you're not connected to the Internet, which a lot of people now are doing as we're moving into, you know, more powerful computers, more capable models, things like OpenClaw where people are, you know, downloading open source models and running running them locally. Those models are running off of usually very old training data. So you have old training data. You have data that you can upload, right, both individually in chats, via connectors and integrations, you know, via projects, GPTs, etcetera. So there's training data, data that you can upload or connect via, you know, different mechanisms within these large language models.
Jordan Wilson [00:17:22]:
And then there's the ability to browse the web. So you really have to have a handle and an understanding and make sure you're always checking the summarized chain of thought to see and ensure that they're looking at the right types of websites and they're not making things up or hallucinating anything. And I think this matters because, I think humans aren't really gonna be using the Internet a lot in the future. That might sound weird. I find myself personally using the Internet less and less every day, and I'm spending more and more time in these different platforms. Right? Like your perplexity, notebook LM, you know, Gemini's deep research, ChachiPT deep research, etcetera. And I do think that as agents, become more prevalent, it's just gonna be agents really browsing the web. That's why we have now all these agentic protocols that are making websites more, accessible and, well, the data in the website's more readily available to agents.
Jordan Wilson [00:18:20]:
Right? I could see a day where, you know, obviously, we're already starting to see it. You know, even from a technical, kind of SEO or AIEO or GEO, whatever you wanna call the new SEO for AI. Right? Where there's multiple versions of a website. Maybe there's one version that's for humans and one version that's for AIs. And probably the version for AIs is ultimately gonna become way more important for most businesses that rely on discoverability online. Alright. Category three, category four, another big one here. And it's, it's an emerging one and I think a trending one in 2026, and this is voice and speech AI.
Jordan Wilson [00:18:59]:
So this is both speech to text. It is text to speech. So yeah, this is, if you didn't know, yeah, there's an AI for that. And they're good. Right? I've even been building a secret little project in this category, and even the open source models. Right? So not just the frontier models that you're paying for, but this, I think, is going to be really important moving forward. Not saying that people aren't gonna be able to type or that everyone wants a text to speech voice, narrating something, but, you know, I usually I usually see where I'm spending more and more of my time, and I'm kind of forecasting that out and saying, okay. I think a lot more people are going to, be using this.
Jordan Wilson [00:19:42]:
Because like I said, I use, hundreds of AI tools every, year, and I've used thousands of them in the three plus years I've been doing, everyday AI. So, you know, I I fancy myself the average knowledge worker. Right? I'm not overly technical. I'm fairly, you know, technical. So I'm kind of in the middle. Right? So I think things that, at least for me personally, stick and make sense. I think that they will make sense for many other people. And I think that's why, voice has really been a trending category.
Jordan Wilson [00:20:13]:
So some of the main players here, Eleven Labs, Mistral actually has a great new, model that they, announced, which is Voxrel, Fireflies, you know, for, you know, AI meetings. You know, you have a lot of those within Copilot, within Google Gemini. They have their own, meeting assistant that transcribes meetings. You have Granola. You have Otter. Right? There's so many big players built on the text to speech and on the speech to text side. So who uses these? Well, anything from podcasters like myself, right, video creators, global teams, corporate training, accessibility. There's so many different use cases.
Jordan Wilson [00:20:49]:
But one of the reasons why I think this category is gonna become increasingly more and more important is there is a literal knowledge, layoff coming. Right? So both as companies are unfortunately, we're seeing because of AI, laying off hundreds or thousands of workers for some of the bigger tech companies. I do think that's gonna unfortunately sprinkle into, kind of corporate America throughout the 2026 and into 2027. So I think you are not only companies are losing a lot of their institutional knowledge when they're laying people off. Right? They they're doing it public companies are doing it because it helps their bottom line. Right? And they can become leaner. And a lot of people don't know what the future AI jobs are gonna look like, so they're like, well, we might as well get rid of the old jobs now, save up money. And so when we need those new AI jobs, we can go hire people, and well, we'll be leaner in the process.
Jordan Wilson [00:21:42]:
But you also have the silver tsunami, which I talk about a lot. Right? You have more and more, people who are retiring. And a lot of these, more senior workers maybe have decades of institutional knowledge, subject matter expertise in their head. And I think that's another reason why some of these platforms are becoming more and more important. Right? Documenting this institutional knowledge. Right? Doing Zoom meetings and just talking to senior people, right, that maybe they're leaving in a couple of years. You know, trying to understand their thought process and then using that as, well, your company's internal IP. I think, these type of tools are extremely important, because and now it's also very common for you know, it's starting more in the coding space, but it's more and more common for people to just be talking instead of typing.
Jordan Wilson [00:22:32]:
Right? As large language models become better at understanding natural language, even for me, I find myself I'm a decent typer. Right? I'm in front of the computer ten, twelve hours every single day. Large language models are very good now, I think, at understanding natural voice, and you can speak about four times as fast as you can type for the most part. Right? So this is another reason why, this category. And then on the flip side, right, text to speech. I know this one is maybe me more personal. I hate reading. I hate reading huge, lines of verbose AI generated text.
Jordan Wilson [00:23:12]:
Right? Because that's where I spend the majority of my day. Right? So everything I read for the most part is AI generated and extremely long even with custom instructions and, you know, certain settings. Right? So I think, you know, being able to use these, text to speech tools or, you know, they're baked into a lot of, you know, ChatGPT and Gemini as an example, you can just click the read aloud button. It's it's a big category in and of itself. Alright. Number five, AI image generation. So how this works, it's a little different. So, for the most part, most of the things that we've talked about so far in the first categories were all based off some of that original, you know, transformer models.
Jordan Wilson [00:23:49]:
Right? The, original research from Google. You know, you can say it's next token prediction, but like on steroids. AI image generation is a little bit different. So these are something called diffusion box. So, So, essentially, they start with random noise and they've been trained on a large data set, and then it kind of refines it into coherent images. So you could kind of think of it like, you know, I don't know if you have that that little petri dish of milk and you, you know, drop some black ink in it and it eventually diffuses into a photo. Right? That's kinda what it does behind the scenes. I actually like I really enjoyed the older versions of, like, you know, OpenAI's DALL E when it it was, like, really slow because you could actually kind of see, right, in some of the earlier mid journey versions, right, when the images are real slow, you could understand the diffusion process because you could see it and watch it.
Jordan Wilson [00:24:41]:
Right? One image might take a couple minutes, and it turned out absolutely terrible. Right? Like, back in the Dolly two days. But then you could actually understand and see, oh, okay. I kind of understand what this model is trying to do. Alright. So some of the main players here, Midjourney was one of the OGs, even though I don't think that they're any longer, like, a top three name, but you have to, kind of, tip the cap to them. Then you have Flux, Ideogram. And then as always, you have the big players, right, which is, GBT Image 1.5 from OpenAI, Nano Banana Pro, Nano Banana Pro two.
Jordan Wilson [00:25:15]:
And then also Microsoft has recently, re announced some pretty decent rides, some top five ask, AI image models. So who uses these? Well, I think anyone can. Right? I think especially, with using something like Nano Banana Pro that can make slides. Alright. We're to kind of get into, that category, a dedicated category. Right? But, you know, things like GPT image 1.5 inside of ChatGPT, Nano Banana Pro inside of Google Gemini, being able to make infographics, things like that. So, yes, I think people think that these are only for designers, marketers, content creators, but I think they can really be for anyone. Right? Because we also have to understand.
Jordan Wilson [00:25:58]:
Right? I say this as someone, I'll say mid career. Right? I think the younger generation doesn't really care about text. Right? They care about videos, and that's also the base for creating AI videos is both AI image generation, but also just more interactive graphics and, just better, visual elements. I think, businesses should be starting to experiment not just on social media, on their website, on traditional marketing materials as well that are just, you know, blobs of thousands of words of text. This is a great, a great kind of tool category to start exploring if you haven't already because, I think stand alone image tools, are losing ground. Right? As now we have these AI image generators that are really sweeping the field. Alright. Category six, this is AI video generation.
Jordan Wilson [00:26:50]:
So here's how it works. Well, it's kind of like the in image generation, but then like times 60 for all the different, you know, frames per second or the, you know, 30 frames per second. So it's slightly, similar technology as they have these diffusion models, but they are also then extended into the time dimension. And then there's denoising frame sequences. The other thing with good video mod models is they are extremely expensive to run. Right? So as an example, Sora, right, depending on when you're listening to this, but so technically, it was just last week, OpenAI shut down Sora. One of the reasons why, well, video generators are extremely expensive because they also understand the world. Right? I do think that video generation and world model generation are going to start to blend kind of like what we've seen out of runway, because the video generate, generators, good ones, they also have to understand gravity.
Jordan Wilson [00:27:47]:
They have to understand, you know, shadows, reflections, you know, sun or right? They they like, they have to understand things. You know, character consistency. Right? Like, if if if it's a video generation of, you know, someone walking toward a pencil on the ground, but from the pencil's point of view, right, the object is gonna shame, change shape, dimension, all those things in real time. So, the video generation and I think the, the gap that was closed in 2025 because at the beginning of 2025, AI video is not good at all. Right? And then fast forward to the 2025, and AI video is actually, to the average eye at the 2025 indistinguishable for the most part, if it's done well. And I think where we're at now in, you know, mid twenty twenty six is even better. Right now it's even people like me. Right? I have a background in videography.
Jordan Wilson [00:28:45]:
It's getting harder for me to even realize. So some of the main players here, you have Google v o three one. You have runway gen four five. You have claim three o. You have seed dance. Right? Seed dance gone. It's gone mega viral recently because they're kinda making Hollywood videos and kinda got in trouble. You have Pika 2.5.
Jordan Wilson [00:29:01]:
So who uses these? Obviously, if you are working in or around video, social media, I think these are great tools to use. And I think that video is one of the hottest and biggest growing categories of all of them. Alright. Moving into the next category, a category that's actually not super competitive, and I'd say it's kind of one of the last defensible creative AI frontiers right now, and that is category seven music generation. So this works a little bit differently than some of the other models. And, I I won't even get into it tech, the technical side because it technically uses like image generation to visualize what music looks like. But it is, for the front end user, how it works, not too worried about, technically how it works under the hood for this one, But you can essentially, in text, describe the genre, the mood, the lyrics, and then you can receive a fully mixed song in seconds. So some of the main players here, well, you have Suno v five five, which just came out.
Jordan Wilson [00:30:03]:
Yes. I know. Depending on what you're listening to, I don't know. Maybe we're on Suno v seven by now. But right now, at the time of this recording, Suno v five five, you have UDO. You have Eleven Labs now. So Eleven Labs, was originally a big text to speech player. Now they're in the AI music space as well.
Jordan Wilson [00:30:21]:
Google yeah. Like I said, Google in, like, almost every category because they're a leader in almost every single category. Their new Lyria three pro, actually really good, being able to create three minute songs, and you don't even need a separate subscription. So there's a lot. So, you know, who uses this? Again, content creators, anyone that's, looking for, you know, music to accompany a video, anyone's needing royalty free audio. And I think that, again, it's actually a shockingly, a lot of money that these companies are making. I think the last I saw that Suno, was at something like $300,000,000 in revenue. So, yeah, these companies, even if you haven't heard of them, they are really big.
Jordan Wilson [00:31:02]:
They are really good. And the the the cool thing, right, because I've always been, kind of impressed. I've used, Suno since the very early days, and they're the biggest player, in the space by far. We had their CEO on the show, you know, many years ago. I say many years ago. It feels like like 10, but it was, like, two and a half. They've been the platform that's been on fire. Right? But you can literally describe anything.
Jordan Wilson [00:31:26]:
You can, pull out, you know, just the guitar, and you can say, hey. Just change the guitar on this, but keep everything else the same. Now you can experiment with your own voice. Right? So there's so many cool things that you can do. So, you know, maybe as you're understanding these categories a little bit more, maybe there's some more use cases for your company that you maybe didn't think of before. Alright. Let's get to category eight, and that is design and visual content. This, I know I could have technically broken this one down into, like, three or four separate categories, but instead, I decided to keep it a big parent category because I do think that, like, as an example with, like, notebook LM and, you know, we're gonna see a a a lot of this in Google with Nano Banana.
Jordan Wilson [00:32:10]:
The lines on this are really gonna start to blur. Like, as an example, if you look at Nano Banana, yes. Originally, it was okay. It it uses, it's it's for photos and infographics. Right? But then all of a sudden, it's doing complete slide decks, and then it's the base for video slide shows. Right? But so this category, it is AI that combines content generation, layout design, and imagery into finished artifacts, whatever whatever that artifact may be. So some of the big players here, were, you know, kind of earlier in the game, and they're doing things like decks. Right? So obviously, I think you, now have Microsoft Copilot, which is pretty good, in creating, decks inside PowerPoint.
Jordan Wilson [00:32:57]:
Actually, I was surprised how good the new task feature in Microsoft Copilot was as at creating decks. Alright. So even outside of the tools that we talked about, you also have to understand. I'm not mentioning, you know, Chat GPT and, Claude for each of these because they can all create slides. Right? But I think this is a little bit different because Gamma, you know, Gamma is one of the big players in this space, you know, very unique tool that can create, you know, a 20 page slide deck can create, you know, websites. It can create anything visually, and I think it is, a really does a really good job. Similarly, you know, Canvas AI magic suite, falls in that category, beautiful AI, but then you have a whole another, kind of genre of, you know, AI design and visual content, which would be AI, AI avatars or digital twins. Right.
Jordan Wilson [00:33:48]:
So you can upload your, you know, they're actually fairly easy to make way easier than they were, like, two or three years ago with tools like Symphyzia, HeyGen, etcetera. Then you also have other platforms that are, you know, visual content such as Google Stitch. Right? So it helps you and, you know, I Figma, AI. Right? Like, there's so many so many tools that could fall under here, but, you know, there's just tools that you can type in a sentence, and it will create an entire app layout. And then you can export that, to, well, something in category nine. Alright? As we move on to category nine, and that is vibe coding app builders. So here's how it works again, natural language. Are you seeing the trend here? Y'all for the most part, natural language is a very important skillset, right? Because for the most part, this is a starting point for all of our categories, is just natural language.
Jordan Wilson [00:34:41]:
So with vibe coding, right, if you slept through 2025 and missed vibe coding, I think being the word of the year for, like, dictionary.com or something like that. You know, you describe an app in plain English, and it can get deployed with a live URL. So again, between nine and ten categories, nine and ten, the, you know, Vibe coding app builders and the AI coding co pilots, there's a lot of crossover similar capabilities. Right? Google AI Studio could be in either. But I think the main players that are specifically in the vibe coding app pillars would be like Lovable, Replit, Bolt, v zero by Vercel, Base 44. Right? There's there's a lot of ones that are really just made, to be a complete, infrastructure for an entire app. Right. And I'm gonna talk a little bit why I even separated, vibe coding app builders into AI coding copilots because, yes, they do have a lot of similarities, but they're definitely different.
Jordan Wilson [00:35:40]:
Right? So I'd say people using this are more of the nontechnical people. So, you know, nontechnical founders, entrepreneurs, marketers who are building tools or anyone with an app idea. Right? They are great platforms. I think, you know, earlier on in the, you know, 2024, early twenty twenty five, I wasn't a big fan of these tools, to be honest. Right? But now they're much they're much better. Right? Now that they have complete, you know, front end, back end, auth user management. Whereas before, they were just, you you know, you could make disposable apps and they were kinda disposable. Right? Very similar to what you can make inside chat g v t, Claude, Gemini.
Jordan Wilson [00:36:18]:
Right? But now they, have turned into kind of fully functioning, tools. So, I think this is one of the most heavily funded categories of 2026, and it's rewriting who builds software and how it's used. Yeah. Like, literally, if you look at The US stock market, traditional software has gotten squashed from category nine and category 10. So great segue to category 10, which is our AI coding co pilots and agents. So here's how it works. Well, AI just can understand your entire code base. It can write code, test it, and ship new features, but it is not just coding.
Jordan Wilson [00:36:57]:
It is also any work that can be done on your desktop. Right? Because there's this magical thing that, you know, most people, unless you're a dork like me, I've been playing around with different terminal tools, for twenty years. Right? But, computers themselves are controlled by something called a terminal. Right? So you can run any command, on a computer, including accessing local files on your desktop, creating files, running code, all of those things. So that's why you kind of have a little difference, in the kind of vibe coding tools that are ultimately about, like, okay, you're creating a piece of software, where AI coding is a little different. They can create software, and, yeah, you can vibe code in all of these tools. I do all the time because the main players here, you have Cursor, Claude Code, GitHub Copilot, Windsurf, Codecs from OpenAI, anti gravity, from Google. Right? But I think that we're starting to see it.
Jordan Wilson [00:37:52]:
These main players also dip their toe into nontechnical knowledge work as well and just becoming agents in and of themselves. So who uses this? Well, professional software developers, engineering teams at every level. But like I said, I think with Claude Code, and Claude Cowork, which I think I have on, the next one here, and I think codex as well. I think more and more, and I know you've heard me say this a lot. And if you are not a technical person, if you're not a software engineer, if you're not a dev, you don't fancy yourself any of those things, you need to be using cloud code. You need to be using cloud co work. You need to be using, codecs. You should probably start using anti gravity.
Jordan Wilson [00:38:35]:
Right? Because as we start talking about the future agent layer, which right now is, well, computer using agents, but I think that's not the ultimate or the end layer. But I think you really have to understand these tools. Right? And we're gonna get to kind of my takeaway advice on that here in a minute, but every single coding tool right now is racing toward fully automated agent driven development, which is why I think if you understand AI tools, right, yes, there's great agent capabilities that can happen in the front end AI chatbots, but you don't have every single program, every single file, every single app that you use, you can't connect those. Right? A lot of those maybe live on your local machine, which is why I think it's important, to pay attention to the AI coding, copilots, and agents category. It is important. And then last but definitely not least, and, yeah, this one could literally be seven different categories. I've actually broken down, AI agents into the seven different types of agents. But for ease of today's episode, in understanding all the different AI software, not just agentic software, We're gonna say category 11 is autonomous AI agents in, agentic browsers.
Jordan Wilson [00:39:48]:
So this is how it works. Well, you state a goal in the AI plans, the steps, it uses real tools, and it completes the work. Right? So there's a lot of main players. Alright? And, you know, we already mentioned some that you might think were in this, like, replit, you know, maybe clawed code. But, I'm saying for the most part, these are more general agents, or nontechnical, non dev type agents. So players like Manus, Genspark, obviously, OpenClaw, definitely fits in this, ChatGPT's Atlas browser, Perplexity Comet's browser, their Perplexity computer, personal computer, Claude Cowork, especially their new computer use that can click, use a mouse, can open different, files, different programs on your computer, which makes it really unique. You have Microsoft's new cowork, you know, you have Claude Dispatch. There's so many.
Jordan Wilson [00:40:43]:
Right? And I think I would even probably throw codex into this platform as well or into this category. But this is, I I'd say for more, technically proficient knowledge workers, right, who are, looking to delegate as many of their time consuming mundane tasks as possible. Alright. And, well, I think the main reason why this matters is the shift is happening right now, right? Where AI is no longer just something that you ask a question to, and it works reactively. Right now, agents are working proactively on schedules. Right? Like if you're an old school, you know, technical person, right? Heard about, you know, cron jobs and, you know, RPA robotic process automation right now. That's what AI agents are doing. Right? They're running computers on schedules, and they can do anything that us humans can do if you're using the right tool at the right time for the right purposes.
Jordan Wilson [00:41:40]:
Alright. So those are the 11 categories, but I wanna leave you with this. I wanna leave you with some advice and, you know, we talked about, you know, why if you're using only one of the categories, you're probably failing and, you know, a framework to make sense of every tool and how to kind of build the right AI stack. Right? So here's some pieces of advice as we wrap up. So Gemini is an absolute beast. Right? We've actually seen in the recent weeks, OpenAI intentionally stepping back and pulling some of their modalities off the table. Right? So as an example, OpenAI, you know, kind of killed Sora, very popular, you know, AI video generator. And I wouldn't be surprised if OpenAI maybe, slows down progress on some of their others, to really just start focusing on, more profitable, sectors, within the enterprise, and just focusing more on their core models.
Jordan Wilson [00:42:39]:
But Gemini is an absolute beast. It does. I'm pretty sure as I'm looking at all my different, categories here, yeah, it does every single one. I'm just double checking. Yeah. Image generation, AI video generation, AI music generation, design and visual. Yeah. It does it does everything in here.
Jordan Wilson [00:42:58]:
Right? So, you can't talk about all these different categories of AI and not absolutely say that Google is dominating this space. Right? Anthropic, great models. Coworker Claude, Claude Code, some of the most, widely used tools over the last three to six months. But Gemini is crushing it across the board in terms of not just multimodality, but in terms of having a very powerful tool in every single category. Alright? The other thing, chat gbt is winning the user's race. Right? Because we did, in in earlier show on who's winning the AI race, but I think it's important to know as well as you're choosing, oh, you know, what should I be using for AI images? Should I be using, you know, chat g p t or something else? Well, I like we said, you know, OpenAI seems to be cutting back and focusing on some of their services, but there's a lot of other great tools, that can lead the way. But still chat g p t is absolutely dominating, users across the board. And I think understanding the players, even at a base level like I just went over today, is extremely important because I think that you have to understand where the platform does enough versus where the specialist wins.
Jordan Wilson [00:44:18]:
Right? Because what can happen is, well, you can start drowning in tools because you can look at all these tools and say, oh, wow. I have all these capabilities now that I didn't have before. And depending on what your job is, maybe that's part of your job. Right? But I think some of the best things that the average everyday knowledge worker can do is say no to certain AI tools that would give them new capabilities that they didn't have before, unless that is a nagging requirement. Right? And and it's it's it's hard to stay away from that. Right. But I think you have to stick to where your specialty lies, and what category of tools are going to help you, expedite that. So here's some advice.
Jordan Wilson [00:45:02]:
You don't when putting together your ideal, you know, tech stack of the right categories, you don't need every single, you know, oh, I'm gonna pick one tool from all these categories. Don't do that. Like I started with right? Like I started the show by saying, you you you shouldn't use as many AI tools as I do. You shouldn't be using for the most part. Right? Unless you're a one person company solopreneur, entrepreneur, you know, maybe you're a jack of all trades marketer that has to do way too much. But for the most part, you shouldn't be using, one tool from all 11 categories. You know, maybe if you wanna learn the basics, you know, set aside a weekend, you know, spend an hour in each category shirt, but you should just be choosing two to three categories that most align with your daily work. But I will tell you this, you need to be extremely proficient in well, technically categories one, two, and three because they're all largely which model related, and you at least need to have one tool, that you are very good at in category 10, which is the coding co pilots, and category 11, which is the general super agents.
Jordan Wilson [00:46:05]:
Because when I talk about this shift from humans doing the work to agents doing the work, I think there's gonna be a period of probably, I don't know, eighteen to thirty six months where if you can commit to those categories or your department or your company, you are going to have an unfair headstart on everyone else. I think in eighteen to thirty six months, these things are going to become the defacto way to work, but we have superhuman AI capabilities that are here now. And it's only a matter of time until that just, well, everyone learns to work that way, But they're here. They're live. They don't require, you know, 30 pounds of duct tape. They're there. Alright. So, you know, as an example, stick to, the two categories that make the most sense.
Jordan Wilson [00:46:55]:
So, you know, a marketer might pair AI search with design tools. An entrepreneur might, you know, pair vibe coding with agents. Right? Really stick to the lanes of that align with your expertise and what your outputs need to be. But I think the real divide this year is going to be, between those who build a real AI stack versus ones who just uses one tool casually versus those who get distracted by a shiny AI syndrome. So you can either be the one that just sticks in one category, which is probably not enough. You could be the distracted one, the learner that sticks in every single category, but you're not, excelling in any, or you you gotta find that middle spot, that sweet spot. Right? Knowing the, you know, three to four categories that I mentioned that I think are essential for any knowledge worker, and then finding the right tool or maybe one or two additional categories in there that really align with your day to day work. And understanding the full map is where your advantage comes in because I think most people are just using one AI category and completely missing the others that are sitting around them.
Jordan Wilson [00:48:04]:
Right? If you're kind of new ish to AI, after listening to this, your mind might be completely blown. Alright? You're like, oh, I didn't know that Google had a tool that could do this, or I didn't know, you know, Suno could create all this music with your voice. I didn't know Gamma could, you know, take a blog post URL and create a deck out of it with your own image assets. Yes. Right? They can do all these things. But you have to ignore shiny object AI syndrome because I think for the most part, almost any AI product in the world can fit under one of those 11 parent umbrellas. So pick your categories wisely, the ones that you need to compete in, choose your AI operating system or your AI tool of choice, and try as much as you can to ignore the others. Because you need to pick the categories that are gonna move your work forward starting today.
Jordan Wilson [00:48:53]:
Alright. I hope this is helpful. I know this is kind of a longer one, but I think it was an important show that we had to do because I get these questions all the time. What tool should I be using? Oh my gosh. Like, what are AI's capabilities? Well, there you go. Obviously, you know, when it comes to, you know, April, May, June 2026, I think capabilities are gonna change. Maybe these categories are gonna change. But right now, if you're listening to this, at least in the 2026, you have a great advantage.
Jordan Wilson [00:49:20]:
You understand all of these AI tools, where they fit in the landscape. Now it's up to you to do something about it. I hope this was helpful. If so, please, if you're listening on the pod podcast, please make sure to subscribe to the show. Leave us a rating if this was helpful, and then make sure you go to starthereseries.com. That is going to give you free access to our inner circle community. It's literally not listed anywhere. You can't find this anywhere on the Internet except well, unless I email it to you, but otherwise, by going to starthereseries.com, make sure to go, join and check out, all of the different episodes in the series that you can go listen to watch for free now.
Jordan Wilson [00:49:59]:
So thank you for tuning in. Hope to see you back tomorrow and every day for more everyday AI. Thanks, y'all.