You've got seventeen AI tools scattered across your browser tabs right now, don't you? Yet somehow you're still manually doing the same repetitive tasks that drove you to purchase those licenses in the first place.
Whether you're a solopreneur building your empire or leading a team trying to eliminate admin drag, here's the brutal truth: most entrepreneurs are picking AI tools like they're shopping for shiny objects at a tech bazaar, grabbing whatever looks innovative without any strategic thinking behind it.
The result? Digital chaos that creates more work instead of saving time.
This is exactly what Kevin Williams, CEO and Co-Founder of an AI enablement company, discovered when building his own essential tech stack. Kevin has cracked the code that most of us struggle with - creating a daily workflow using AI tools that actually simplifies operations instead of complicating them.
The Strategic Framework: Build, Buy, or Delay?
Before Kevin adds any new AI tool to his arsenal, he asks himself a critical question: "Is this a build, a buy, or a delay decision?"
As Kevin explains, "Just a few months ago, people like me were having to build pretty much any solution that we wanted to, and that gets really expensive. Most organizations don't have the capacity to do that."
The buying lens is now your friend. If you pour hundreds of billions of dollars into Silicon Valley, amazing tools start emerging. The marketplace is improving rapidly, which means most problems you're facing likely have existing solutions.
However, Kevin only considers building custom workflows when they'll create a genuine competitive advantage - something that won't be readily available in the market because of how his organization operates uniquely.
The Essential Three: Kevin's "Can't Live Without" AI Tools
After testing dozens of AI tools, Kevin's daily workflow centers on three essential applications that work together like a well-oiled machine:
1. n8n: The Workflow Revolution with Intelligence
Think of n8n as "Zapier's smarter cousin." While most entrepreneurs are familiar with basic automation tools like Zapier and Make, n8n revolutionized the space by being first to add an intelligence layer using AI models.
"It's quite another to use an AI node to deconstruct an invoice, analyze it qualitatively, bounce that off Slack to inform the bookkeeper, and then log it into a file," Kevin explains. "It's basically a Lego version of workflow building with an AI layer sprinkled on top."
The game-changing advantages of n8n:
- Cost Structure Revolution: Instead of paying per interaction like Zapier, you can self-host n8n and pay for processing power rather than individual actions
- Community-Driven Templates: Access 5,000+ pre-built workflow templates that you can instantly import and customize
- Seamless Shareability: Workflows download as JSON files that can be shared and imported with 95% of the work already done
2. Gamma: The Presentation Engine That Thinks
Gamma transforms how Kevin creates presentations, proposals, and client communications. Starting with just an outline, Gamma uses multiple AI models to generate complete slide decks with associated images and charts.
Kevin recently adapted an entire speech for a new industry: "I went from one of my existing speeches and said, 'Adapt this material for industrial machinery instead of insurance,' and it ran through the entire presentation - changed graphics, callouts, discussion items - in a blink."
Beyond presentations, Kevin uses Gamma for:
- Converting lengthy documents into digestible 20-page decks
- Creating professional client updates in minutes
- Information synthesis and rapid content consumption
3. Wispr Flow: The Productivity Breakthrough
Kevin calls this "the smallest technological change that has made the biggest difference in my entire life." Wispr Flow is an LLM-layered dictation tool that eliminates the productivity bottleneck of typing.
The math is staggering: Kevin's brain operates at roughly 1,000 words per minute, he speaks at 200 words per minute, but typing limits him to just 50 words per minute. That's a 75% productivity tax on communication.
"I'm now dictating 26,000 words a week," Kevin shares. "Why should we be putting artificial filters that limit our output to 25% of what we could produce?"
The "Adventure, Education, Automation" Philosophy
Kevin's company operates on a powerful principle that directly applies to tool selection: "The first time you do something, it's an adventure. The second time, it's an education. The third time, it's an automation."
This framework prevents the costly mistake of automating processes before understanding them. "If you try to automate right out of the box, you're going to automate in ways that misinterpret what the goals are," Kevin warns.
The lesson: You need to manually struggle through a process at least twice before you're ready to automate it effectively.
Red Flags: The Biggest AI Tool Mistakes
Kevin identifies several critical errors entrepreneurs make when building their AI tool stacks:
1. Building Independent Agents Too Early
Current agent development hasn't caught up to orchestration capabilities. Instead of isolated agents, focus on intelligent workflows that chain together different AI models for specific tasks.
2. Falling for "One Platform to Rule Them All" Promises
Kevin prefers specialized tools that excel at specific functions. "A lot of these agent companies are jacks of all trades and masters of none."
3. Productivity Pornography
"A lot of organizations do things that make them feel busy - like making spreadsheets and reports they'll never use again. That's not valuable. You've just filled your time with stuff."
The Audit Framework: Cleaning House
Kevin's process for evaluating and streamlining AI tools focuses on identifying friction points and measuring ROI:
Step 1: Friction Analysis
- How often do you perform each task? (Daily, weekly, hourly?)
- Where do you feel the most resistance in your workflow?
- Which activities create the biggest bottlenecks?
Step 2: Usage Tracking
- Maintain a sandbox for testing new tools
- Track actual usage versus promised benefits
- Look for tools with per-use pricing models when possible
Step 3: Subscription Hygiene
"I look at my P&L and go into that AI tools section to figure out when was the last time somebody actually used it for something that wasn't just dinking around, but was actually ROI-based."
The Strategic Question That Changes Everything
When evaluating any new AI tool, Kevin returns to one fundamental question: "What is the output I'm trying to solve?"
This simple question cuts through the hype and forces you to:
- Focus on specific problems rather than general capabilities
- Consider whether existing LLMs with advanced prompting might solve the issue first
- Evaluate true ROI potential before committing
As Kevin puts it: "Where are the places where you spend low-value time? Where are the low points in your day where you think, 'I really don't want to do that'? Those are your automation targets."
Building Your Essential Stack: The Right Way
The path forward isn't about accumulating more AI tools - it's about thoughtfully selecting the right tools that work together to eliminate your biggest productivity bottlenecks.
Start with Kevin's framework:
- Identify your highest-friction, repetitive tasks
- Test solutions manually first (Adventure phase)
- Refine your process (Education phase)
- Only then automate (Automation phase)
Remember, the goal isn't to feel busy with cool AI toys. It's to reclaim your time for high-value work that actually moves your business forward.
The question isn't whether you need AI tools - you do. The question is whether you'll choose them strategically or let digital chaos choose for you.
What's the one repetitive task that's stealing your time every single day? That's where your essential AI tool stack journey should begin.
Ready to build your strategic AI tool stack? Download our free AI Tools for Content Creation guide to start eliminating admin drag and reclaiming your time for what matters most.
You've got seventeen AI tools scattered across your browser tabs right now, don't you? Yet somehow you're still manually doing the same repetitive tasks that drove you to purchase those licenses in the first place. Whether you're a solopreneur building your empire or leading a team trying to eliminate admin drag, here's the brutal truth - most entrepreneurs are picking AI tools like they're shopping for shiny objects at a tech bazaar, grabbing whatever looks innovative without any strategic thinking behind it.
Kevin Williams, CEO and Co-Founder of an AI enablement company, has cracked the code that most of us struggle with: building a daily workflow using tools like n8n, Gamma, and Whisper Flow that actually simplifies operations instead of complicating them. This guy has taken his company philosophy of "adventure, education, automation" and applied it to create a personal tech stack that's so effective, he literally can't imagine working without these specific applications.
In this episode, Kevin pulls back the curtain on exactly which tools made the cut and why - plus the strategic thinking that separates useful AI from expensive digital clutter. You'll discover his decision-making framework for evaluating new tools, how his essential stack works together to create seamless automation, and why fewer, better-integrated tools beat sprawling collections of disconnected AI applications.
Because at the end of the day, it's not about having more AI tools - it's about having the right tools working together like a well-oiled machine to reclaim your time and focus on what actually drives results.
The AI Hat Podcast host Mike Allton asked Kevin Williams about:
✨ Strategic Tool Selection Framework - Learn the decision-making process Kevin uses to separate genuinely useful AI tools from shiny object syndrome purchases.
✨ Essential Daily Workflow Stack - Discover the specific tools Kevin can't live without and how they work together to create seamless automation.
✨ Quality Over Quantity Approach - Understand why fewer, better-integrated tools beat a sprawling collection of disconnected AI applications.
Learn more about Kevin Williams
- Connect with Kevin Williams on LinkedIn
Resources & Brands mentioned in this episode
- Ascend AI Labs
- n8n
- Zapier
- Gamma
- Wispr Flow
- Magai
- Apify
- Scalability Audit
- Explore past episodes of the The AI Hat Podcast podcast
Full Transcript
(lightly edited)
Why Most Entrepreneurs Pick the Wrong AI Tools: An AI CEO's Essential Daily Stack That Actually Works
[00:00:00] Kevin Williams: A lot of organizations, solopreneurs and beyond do a lot of things that make them feel busy, and it's sort of like productivity pornography where they're just like, yay, I'm, I made this spreadsheet and I made this report and I did this thing. Or, heck, I spent all day doing a workflow that I'm never gonna use again.
[00:00:17] Like, that's not really that valuable. You're just filled your time with stuff and you felt good at the end of the day. Okay, great. But how can you put yourself in this place where you're going to be more productive in alignment with the things that you're good at? And dare I say it, the things that you actually like to do.
[00:00:35] Mike Allton: Mike's got The Hat, the AI kind. Unlocking power in business, you'll find. Cut through the hype, explore the tech. Every week ,new ideas to check.
[00:01:00] Allton leads us on a digital quest. AI secrets, he'll always invest. Tune in tight, don't miss a beat. Knowledge, groove, keep you on your feet. Listen up! To the AI flow.
[00:01:23] Business growth, here we go. Cuttin' edge edge insights, that's the plan. With The AI Hat you'll understand.
[00:01:44] Greetings program. Welcome to the AI Hat Podcast, the resource for established entrepreneurs who are ready to stop being the bottleneck in their own business. Each week we architect the simple AI operational systems that eliminate admin drag, and reclaim your time, freeing you to focus on the high value work that truly scales your revenue.
[00:02:01] I'm your host, Mike Alton, subscribe and let's build your empire together.
[00:02:06] But I bet you've got 17 AI tools installed on your computer right now, don't you? Somehow you're still manually doing the same repetitive tasks that drove you to download those apps in the first place. Here's the brutal truth. Most entrepreneurs are picking AI tools like the shopping for shiny objects at a tech bizarre, grabbing whatever looks cool without any strategic thinking behind it.
[00:02:28] It's the digital chaos villain convincing you that more tools equal more productivity. When what you actually need is the right tools working together like a well-oiled machine. That's why I'm absolutely pumped to introduce Kevin Williams, the CEO and co-founder who doesn't just talk about AI tools. He lives and breathes them every single day to run his AI enablement company.
[00:02:52] Kevin's cracked the code on so most of us struggle with building a daily workflow using tools like n8n, Gamma, and Wispr Flow that actually simplifies his operations instead of complicating them. This guy has taken his company philosophy of Adventure Education Automation and applied it to create a personal tech stack that's so effective he literally can't iMagaine working without these specific applications. And today he's pulling back the curtain on exactly which tools made the cut and why. Plus the strategic thinking that separates useful AI from expensive digital clutter. Hey Kevin, welcome to the show.
[00:03:28] Kevin Williams: Hey Mike. Thanks so much for having me.
[00:03:30] Mike Allton: So glad to have you here. I wonder if we could just start with that strategic thinking, like when you're evaluating a new AI tool, what are some of the questions that you're asking yourself before deciding whether or not it might make the cut into your daily workflow?
[00:03:46] Kevin Williams: First, I like to look at, uh, whether or not it is a build, a buy decision or a bit of a delay decision.
[00:03:54] Fact is just even just a few months ago, people like me were having to build pretty much any solution that we wanted to, and that gets really expensive and not most organizations don't have the capacity to do that. So looking at it more from, uh. A, a buy lens, if it's something that is solving a problem for you that's already out there on the marketplace somewhere.
[00:04:19] And believe me, the marketplace is definitely improving. If you pour a few hundred billion dollars into Silicon Valley, uh, amazing enough, good stuff starts popping out. So it's pretty likely that a lot of the choices you make, you're, you're making are not a build decision, but they're a buy decision.
[00:04:37] However, from a build perspective, and I'm really not talking about building something that's super technical or even vibe coding my way into something. I'm talking about constructing some sort of a, a workflow. I really only go about that if it's something that's going to be a competitive advantage to my business.
[00:04:55] So it's something that I know isn't really going to be out there and readily available in the market because of the way that I do things in my organization. Only then am I really looking to, to consider building something again, either internally using lightweight tools or maybe using lightweight dev.
[00:05:16] Mike Allton: So just to help folks understand what, where you're coming from, when you talk about building something, what's a, what's a use case? A workflow that you have in mind. That kind of a direction.
[00:05:25] Kevin Williams: Well, we'll get into n8n in some detail. Yeah. And I'm sure that a lot of your audience has at least experimented with it, and if they haven't, they totally should.
[00:05:34] It's not that, uh, workflow tools or anything particularly new. Um, all of us know about Zapier and Make, used to be called Integra mat. Back in the day, being able to move a document from an email into a folder and then move that somewhere else and maybe do some basic transformations to it but. The unlock.
[00:05:53] And the reason that n8n sort of came out of nowhere, which by the way they've been around for a long time, they were just sort of an, also ran to the other guys, is they were the first to the table where they were able to add, uh, an, an intelligence layer using, uh, AI models in order to interpret the steps.
[00:06:12] And that was a huge unlock because okay, it's fine to move an invoice from one side of your desk and put it into a drive folder somewhere else, and maybe trigger an alert that that happened in Slack. It's quite another to use, uh, an AI node to be able to deconstruct the invoice analyze it qualitatively.
[00:06:33] Bounce that off of Slack to inform the bookkeeper of what's going on here, and then log it into a file and then do something else down the line. It's basically a Lego version of workflow building with an AI layer that's been sprinkled on top of it. So that's, that's different from doing a dev version of something.
[00:06:52] And it is important to note that that tools like this have limitations in that they're not necessarily built to be scaled tools. Uh, if you're doing hundreds of thousands of actions. If you're a direct to consumer company who's trying to do something with every order and you have millions of orders, it's probably not the best long-term approach, but you can absolutely use it to demonstrate minimum viable product of how a workflow is supposed to work.
[00:07:17] Test it out, make changes to it. And then if it's not scaling particularly well then consider using an outsourced dev department. Or you know, if you have the gumption trying to figure out how to build it yourself, I'm not actually recommending that for everyone. Because you need to be focused on your business, not focused on trying to vibe code it your way out of a challenge.
[00:07:36] But for people with the right aptitude, it's easier than ever to build things.
[00:07:40] Mike Allton: Yeah. I'm, I'm laughing 'cause just this morning I was researching Google's Vertex AI studio. And, and trying to decide, is that something I want to share with my solopreneur and creator audience? And the answer was a resounding no because it was just, the learning curve was too steep and the output and the game was going to be too little.
[00:08:03] But I love that you made that convention, that comparison to Zapier. That makes complete sense to me. I've been using Zapier for years, but now Zapier also has layered in that ability to make a sequence or an automation, Zapier kind of AI enhanced. Mm-hmm. So like in my AI SOS membership, i, I teach folks how just as a basic workflow to connect their website contact form.
[00:08:29] Does Zapier have some AI analysis happen inside of Zapier? What's the sentiment of this thing? Is it spam or is somebody actually reaching out properly? Right. And then funnel that response into Notion where they can be doing some basic CRM capability. I hadn't messed around with n8n very much, so I love that, that you're coming on the show and you're telling us a little bit more about this alternative, the Zapier.
[00:08:52] And for some of the solopreneurs like myself who maybe haven't used it too much. Is there another problem that it solves in your daily operations that maybe other tools couldn't? Oh, there's
[00:09:03] Kevin Williams: so many. I'm in love with this tool. I really am. It's gonna be my BFF for a while. First, uh, it has a business model difference from Zapier or Make traditionally, um, you're paying per interaction in Zapier.
[00:09:16] Or Make. And that can be problematic. If you're running a lot of automations or you're doing a lot of complex things you're doing audience vetting or whatever it is, it can get pretty expensive in a hurry. One of the just sort of gangster unlocks from a business perspective that n8n put forward is the ability to self-host it.
[00:09:35] So you can actually. In up your own instance of n8n on hosting or, or something like that on A VPS. And if they, if that's not resonating with, with, with folks, YouTube is your friend. There are a hundred videos out there on how to do this. And it allows you to basically pay for process instead of paying for interact interactions or um, actions.
[00:09:58] So we have. Tens of thousands of actions that are either running from our own business or client businesses, and we can spin up our own instance of it. And we're really just paying for the CPU use on the backend of the hosting as opposed to paying for each interaction. So it scales much, much, much better from a cost perspective.
[00:10:19] The second unlock, which also like, I just have a ton of respect for these guys. I don't know them. I have no in there, whatever, but, is the shareability of the things that you build and the community that they're building around it. The tick that Zapier in particular missed, I think is the community building aspect.
[00:10:37] And what n8n does has done is it's to rec, recognize that workflows, once you break them down to code, are essentially just JSON. With n8n if I wanted to share a flow with you, all I have to do is download it in JSON. You would just upload it in JSON. You'd have the flow instantly, and then all you'd need to do is create the connect credentials.
[00:11:01] So if you're going to Gmail or you're going to your calendar, or you're going to Slack or HubSpot or whatever it is, you'd have to do that layer. But all of the architecture and all the work that I put into it, it's 95% done for you. Instantly shareable. And the result of that is a how cool to, you know, be able to share things amongst your friends, but also how cool as a a business person to be able to super easily replicate your flows and even figure out how to monetize them. There's some nuances there as far as the way that n8n works, but there's also this huge community of n8n Builders who are all doing the same thing. So I think there's a straight template library that's on the n8n site has 5,000 different templates on it.
[00:11:48] So you grab the template, you click on it, it opens, and then you adapt it to whatever you need. Maybe you just like the top part, highlight it, grab the JSON, dump it into another one. So, so, so easy to use. And then the other piece that I think might resonate in a lot of the rest of this conversation that I find super powerful when I'm thinking about tools is thinking about where I operate.
[00:12:13] And in my case, I'm kind of crazy busy right now, and. I spend most of my day living in Slack, and my marketing team needs my input into a, uh, into a blog post, or I need to look at this proposal, or, or I have a new lead coming in, or I have a new meeting coming up. I live in Slack, so I want to operate in Slack, and I have no respect for SaaS platform boundaries at this point.
[00:12:43] I want everything to be operating in Slack, so I can use n8n to trigger flows that either start, end or sit live sometime in the middle in. In Slack where I'm already paying attention, that reduces my switching cost, uh, as far as bouncing between a whole bunch of different platforms and makes it super easy for me to do like an edit to a blog post or approve something or whatever it might be.
[00:13:09] And, you know, I don't wanna overstate. It's so easy. I mean, by all means, hire me to build this stuff for you, but it's pretty easy to grapple with in terms of a technology, and there's a lot of information out there. Uh, sort of a pro tip with Zapier, Make, actually anything you're using is to explore, uh, the new Claude skills and be able to download the skills and create basically your own LLM guide to whatever tool you're using, and between the native assistance that's in L, in n8n and using Claude skills. Even the non-technical would be just blown away by what you can accomplish. The trick as a business person is to understand the desired output and decide and understand what the right trigger is.
[00:13:55] Everything else is sort of free game to create in between and how you elaborate, where you store it, how you move it around, how you notify, et cetera. But you as an operator need to be clear of what you want to happen and what starts the action.
[00:14:13] Mike Allton: Love that. Now, for folks listening, you know, if there's something Kevin or I say that you're not sure what we just said, like J first of all, Perplexity is your friend.
[00:14:22] Uh, have Perplexity open while you're listening to this episode and just drop it in there. Perplexity will happily explain any of those terms to you, but the other thing that I often do. I think this speaks to your, your Claude skills point, which was great, Kevin. If there's something that I want to do and I'm not sure how to do it, and I don't wanna spend a couple hours researching how to do X, y, or Z, I'll run a Gemini deep research on that topic.
[00:14:43] I give Gemini all the context. I'm a solopreneur, I'm helping other solopreneurs. I wanna learn about n8n, and I want to build workflows to help my community. As an example, right? And I'll let deep research cobble together all of today's best practices. I did this when I was building my Circle community.
[00:15:01] I'd been on Circle a couple times. I'd never built a Circle community from scratch, so I didn't know what paywalls were. I didn't know what best practices for spaces were, all these little things, right? I let deep research figure all that out. Downloaded the, uh, reporters of PDF fed it back into my LLM and said, okay, let's build this thing together and have that step-by-step guide.
[00:15:21] But even then, when I've done that with Zapier workflows, I still have run into lots of problems. So I love that n8n has that ability to basically copy and paste, uh, workflow that somebody else has built right into your own n8n instance. The question I had for you real quick was, i'm assuming n innate N supports MCP.
[00:15:44] That's something I've been playing around with with Zapier recently. I used a tool called Magai, which you guys will hear more about in a second. But that's a wrapper for all the different lms. So I'm using Claude within Magai, and I've got a Zapier MCP and connects it to. My notion, for instance, and I'm able to pull transcripts of customer calls into a chat via MCP.
[00:16:05] And that may be the long way around. I'm fairly confident in a year or two. I, I would never even do it that way. But as does any n have similar capabilities?
[00:16:14] Kevin Williams: A hundred percent. Um, you can do almost anything. So we do embeddings chains to create RAG. You know, again, we gotta be careful with the terminology.
[00:16:22] Utah cp, which is even worse than JSON, so that's not me. But yes, you can connect MCP servers. It's e you can have API connections. You can have direct connections, you can have webhook connections. So it gives you a ton of different flexibility in order to make things happen right within that environment.
[00:16:40] And. Going back to the, using your LLM as your guiding friend part, sort of the pro tip is whether you use a Claude scale or whether you just try and contextualize like, I'm using n8n, what do you know about it, et cetera. Is to start the conversation and the context by dumping the JSON in, so now it knows the entire flow as opposed to just trying to fix one little piece.
[00:17:04] And, uh, it's sort of funny because we've been at this for a while, uh, you know, some folks have memories of, of not really being able to work with big pieces of data. With the LLM, so they kind of, I'm gonna snip this little piece and I'm gonna try and fix this little piece. Fixing the little piece doesn't make a ton of sense to the LLM if it doesn't understand the whole picture.
[00:17:26] So by, by adding the whole JSON to the context, it can now look at, it's not the code base, that's not the right term for it, but it can look at the patterns and the arrays that are evolved in all the logic and it can help walk you through the different pieces.
[00:17:44] Mike Allton: Love that. It's so true. I mean, obviously whether we're talking about code or business strategy or marketing or sales or anything like that, we gotta give the LM.
[00:17:52] Lots of context. I, I love this part of the conversation from an automation and systems perspective, that's something I talked to solopreneurs about all day long. You guys have to have workflows and systems and processes in place and as much automation as possible to save yourselves time. Let's switch gears a little bit though, 'cause I want to ask you about Gamma.
[00:18:11] I know that's one of the tools that you talked to me about. I. I think I've used Gamma maybe once in my life, and it's been a long time, but apparently it's essential to your stack. So how does that work? What are the specific workflows or tasks? What does it do for you and why couldn't you replace it with something else?
[00:18:27] Kevin Williams: I'm not saying I couldn't replace it with something else. It's been to some degree, there is, there's something to be said for developing a good enough workflow and sticking to it unless you have a good reason to switch. Um, there is, uh, I spend a lot of time exploring new tools and new approaches to things and yes, indeed, if I do find something that is a sea change in the way that I'm approaching it, I might be tempted to change it up.
[00:18:51] But good enough is good enough for a lot of applications. And I think for us, Gamma falls into this category. And first, what Gamma is, is a, a deck creator. It also has the ability to do sort of one single purpose landing pages and presentations and things like that. And you can start with uh, an outline is typically the way that we start.
[00:19:14] And it uses a, a, an a, a. A variety of different models to produce associated images and charts and whatnot. And basically in a couple of clicks, you go from a detailed outline to a detailed slide deck that then you can manipulate further right in the platform. So I literally just came from a speech and I needed to adapt the speech for the, the, the company that I was working with.
[00:19:41] And, uh, it was dead simple for me to go from one of my existing speeches and say, Hey, um, adapt this material for instead of being insurance, let's make it about industrial machinery now, and here's the brand name. And it just ran all the way through the, the presentation change, the graphics change the call outs, change the, some of the, the, the discussion items in there in a blink.
[00:20:07] And what's interesting to me beyond just the presentation aspect, is how we use it as a tool. And in terms of like client communication, client proposals uh, information synthesis. So one of my own pet uses of it is I'm constantly getting really big documents, and if you put the entire text of the document into Gamma, it will turn the text into a slideshow.
[00:20:32] So you end up with a 20 page deck instead of a. 40 page document and if I can, if I wanna fly through it, it's going to grab onto the key points. So yes, you can use Notebook, lm, there are a bunch of other tools you can use, but I'm all about not breaking my flow and Gamma's already open, so why not use it for that information synthesis.
[00:20:52] But, uh, you know, typical, uh, consultant work. You need to update your clients on deliverables and progress and things like that. And why not make it pretty as opposed to being a kind of a, a lame email. So, uh, now our updates are done in just a couple of minutes. Uh, we link those in with transcription tools.
[00:21:13] Of course, such that there is human in the loop to refine whatever the outline is. We can use an automation that we built in it, and again, to feed it to the gamma API, and then, you know, when you're ready to go, send the deck, it's already sitting there and it, it's ready for you. So, is it as easy to use as PowerPoint for cool animations and things like that?
[00:21:35] No, absolutely not. But if I do need to do that sort of production deck, I can start in Gamma, output it to PowerPoint, and then do the tweaks that I need there in PowerPoint if I wanna do make it pretty, because as you know, Microsoft still hasn't figured out this whole AI thing, so it's not like PowerPoint's particularly useful.
[00:21:55] Mike Allton: I'm laughing 'cause that's absolutely true. Uh, and I was crying a little bit because I do three hour AI workshops for like larger manufacturing companies where they wanna bring in, you know, ten, twenty five, fifty folks onto a call and have them all learn the basics of AI and how to, you know, do their own custom AI assistance altogether.
[00:22:16] Which is, which is great, but I'm always changing some of the workshops and examples and those kinds of things in that three hour presentation deck to meet each individual industry. One night I'm talking to a medical manufacturing company. The next, it's, it's a different organization, right? And I'm doing all that manually.
[00:22:35] Some input from my AI assistant, but I'm, I'm the one who's clicking the buttons and, changing everything in Canva, what I typically use these days to build decks. So clearly. I need to check out Gamma and give that a second second look.
[00:22:50] Kevin Williams: Don't underrate, Canva. Uh, you know, if you are a creator who's using Canva and that's where you're living, Canva does great decks.
[00:22:58] I am not as proficient in Canva, Canva as I am in Canva, gamma sort. Annoying that they're similar or sounding I don't know if this is the case, but one of the key unlocks of gamma is it has a presentation prompt plus individual slide prompts. So you, you can, if you wanna change one slide to reflect industrial machinery or if you wanna change the entire deck.
[00:23:23] It's pretty cool how it goes about it. So at the, at the, at the top level you know, going from, again, I, the, the actual thing I did was it was for an insurance company, so all of the examples were about issuing policies and whatnot, and pivoting that to this machinery company. It, it, it crawled all the way through all the pages and output it.
[00:23:45] Mike Allton: Yeah, no, I, I, I could see Canva going in that direction. Mm-hmm. Great. What I've used the AI within Canva so far is I say, Hey, this is what my deck's about. This is who I'm talking to. Give me a great new, the original template for this deck. And it gives me some fantastic options. So I've used that a couple times and I know they're continually developing, so that's something I need to keep my eye on.
[00:24:10] But just a few weeks ago, I was at the Marketing AI Conference in Cleveland, Ohio. Yeah. And over and over. I heard from folks and they showed me and they talked my ear off about this tool called Wispr, and it's one of the tools that you talked about being essential. I'd never heard of it before this event, and then all of a sudden I'm like, wow, where have you been all my life?
[00:24:31] Walk us through how that fits into your daily workflow, and of course, what it is for the folks listening who like me had never heard of it before.
[00:24:38] Kevin Williams: Okay, so. I literally call this the smallest technological change in my life that has made the biggest difference in my entire life and what Wispr Flow is.
[00:24:51] And, uh, they have a number of different competitors out there. There's another one called Super Whisperer is it is an LLM layered dictation tool. So we all have had the experience with Mac dictation and it's, it's getting better. Don't get me wrong. Apple Intelligence isn't. Totally garbage. But it's not great and requires a lot of rework.
[00:25:14] The unlock for these guys was to add an LLM layer on top of it. So it's listening to everything you say in a se in a section, and it's doing away with the ums and the uhs and the spaces, and it's making just a little bit more sense out of it. It's not trying to rewrite it, although they, there are settings in these tools that you can use as sort of.
[00:25:36] Prompt central to output specific things based on whatever you're working on. But personally that drives me nuts. It just does a better me. Basically, and on a Mac, you just hit function and babble away, and, uh, you release function and wherever your cursor is, the text comes in all neatly cleaned up and ready to go.
[00:26:00] So the, the, the realization that I had is that I just looked at my data actually earlier today. I'm now dictating 26,000 words a week, which is a crazy amount of output that's coming out of me. And I realized that my brain probably works, works at a thousand words a minute. I'm a fast talker, so my mouth works at about 200 words a minute, and my fingers, my poor fingers, as a poor typist only can do about 50 words a minute.
[00:26:27] So you're creating these natural impediments in your communication in order to accommodate a stupid keyboard, which, you know, anybody who's particularly geeky in your audience will know that a QWERTY keyboard is actually designed to slow down your typing so it avoid type, type. Type jams in, in, uh, typewriters.
[00:26:48] So we use a QWERTY keyboard because that's what we were all taught to do. We type because it was the best scaled means of communication that humanity had available to speak with machines. We forgetting entirely that you and I would never sit at a table and type to each other. I mean, maybe 14 year olds would, but.
[00:27:07] In general, we're gonna talk to each other. Why the heck should we be putting artificial filters that, in my case, are limiting my output to 25% of what I could put out? It means that I'm not getting the material out of my brain, out of my mouth. Such that I'm populating an LLM or whatever I'm working with, with the fodder that it needs to have rich context and this might be the part that that, that people miss a little bit.
[00:27:35] That yes, it's cool to dictate, but all of us in this community talk about. The importance of added context, LLMs, like a lot of information that they can use to then chew on to get great results for you. Well, if you're artificially limiting the amount of information you're putting into the LLM by typing, you're not getting as much information into it as you could to get better outputs.
[00:28:00] So it's an unlock on multiple layers. Again, I, I basically don't type anymore. It's super annoying to my wife, but I basically talk like a crazy person to my various devices all day long. And if something breaks, because you know, the internet is held together with wine and duct tape somehow, like if it goes down, I don't even know what to.
[00:28:20] Do with myself because my, my, my fingers won't catch up to what my brain is doing. And I've so easily shifted into this output mode where I can get so much out. I, I actually recently, uh, received an email from a friend who was like, man, Kevin, your emails are so long and detailed. I almost never hear that from a male executive that I'm like, yeah, well, I have friends.
[00:28:46] Mike Allton: I find this so funny because it's a mental block for me. I prefer to type, even though I'm seeing the evidence, I know for a fact everybody can speak far more than they can type. All these folks that I ran into at Macon, they're showing me the Wispr app on their phone because they were so proud of their report card.
[00:29:03] That Wispr was saying, Hey. You've communicating at 150, 200 words per minute instead of typing at 40, 50, 60 words per minute. And the other thing that I'll add real quick that I saw with Wispr is sometimes we will correct ourselves. Wispr will catch onto that and only output the corrected version.
[00:29:24] You might be saying, Hey, let's say this to John. Oh wait, no, I mean this needs to go to Mary, whatever. It'll come out in the corrected version in your typed way, so that's not something you'd have to spend time back, back, back, back, back, retyping it in whatever the case might be. Folks, we're talking with Kevin Williams about the essential tools in his tech stack, and I've got several more questions that you're not gonna wanna miss.
[00:29:45] But first, a quick word from our sponsor, which for me, is also an essential tool.
[00:29:50] This episode of AI for Business Leaders, the AI Hat Podcast is brought to you by Magai. Tired of juggling multiple AI subscriptions and trying to figure out which tool to use and when? Magai simplifies AI for businesses by offering access to all the leading language models and image and video generators for one low price.
[00:30:11] With Magai, you can create custom instructions to personalize AI for every business need. And use them across all LLMs, including ChatGPT, Gemini, Claude, and even Deep Seek. No more building custom GPTs or gems that are limited to just one platform. Magai also lets you create customized workspaces for your entire team, giving everyone in your organization access to the same AI assistants and tools and prompts. Ready to streamline your AI workflow and empower your team.
[00:30:40] Tap the link in the show notes to visit Magai today to learn more and now.
[00:30:48] So Kevin, here's something I'm curious about. Do these tools actually work together? Do they integrate with each other? Do they handle, are they just, you know, completely separate functions within your workflow?
[00:31:00] Kevin Williams: Um, as much as possible, I, I do try and choose a few different tools and operate with them as we were, we were talking about a little while ago. I would say with, with Wispr is particularly interesting in that it is essentially communications glue for anything that you're using. Tabbing over into GPT or uh, going over to Gemini, or in my case, using an AI email client like shortwave.
[00:31:25] It's deeply integrated there. Um. The other integrations you need to think through and you need to decide how much the deep integration is really worth to you. I, I'm just not a fan of building productivity like hacks that save you fractions of a second, but took you all weekend to do. I think there are probably some people in the audience who would hate me for saying this, but, uh, Superhuman the, uh, the email clients has been around for a long time and people who love it absolutely love it because they need to hit hotkeys to start this or do whatever.
[00:32:00] Never once in my life have I really needed to do 400 hotkeys in a day. Um, I'm not saving myself other than maybe a couple of clicks with a mouse now and again. So, if I do wanna build a connection these days, it's very, very easy to use something like n8n or Zapier to connect them together. But it is still a little bit disconnected.
[00:32:23] So I do copy and paste between things as I'm working my way through some processes, particularly as we're still in the stage of learning about what the process is supposed to be.
[00:32:36] Mike Allton: Got it. Got it. Now, I'm wondering, 'cause you've built your whole company on this principle of. Adventure education automation.
[00:32:42] How does that philosophy, do you think, apply to your personal tool selection process?
[00:32:48] Kevin Williams: Oh, a hundred percent. So I'll, I'll elaborate just a little bit. On top of our internal documents, we have this, this tagline, it says, the first time we do something, it's an adventure. The second time. It's an education. The third time, it's an automation.
[00:33:02] And the idea is if you're going to do something three times, you're highly likely to do it, four or five, 400, whatever it might be. But if you try and automate that right outta the box, like okay, I want to create a new lead flow and we're gonna do all of these N eight end automations to make it happen.
[00:33:21] And you haven't actually had the experience of muscling your weight. Through the flow, you're going to automate in ways that. That misinterpret what the goals are and you need to have that adventure where you're totally flailing through it and cutting and pasting and trying to figure out what this is, and then have the focus time to figure out what's right and what's wrong with your approach to whatever.
[00:33:43] It's as opposed to developing some big hairy flow. That you're then sort of stuck with to a degree because you built the flow before you built the process. So the flow becomes the process as opposed to being able to better iterate.
[00:34:00] Mike Allton: I know that's probably one of the challenges I've run into when trying to create like an AI agent.
[00:34:08] I, I, I'm really obsessed with this idea of I need an AI agent. But I haven't been able to find a process where today's AI agent capability will actually work. Instead, I end up going down these rabbit holes where I'm gonna have to change the process and change the workflow to fit what the AI capability is today.
[00:34:28] And even then, it's, it's problematic. What do you think one of the biggest mistakes you've seen when it, when it comes to building, uh, their AI tool stack? Are there red flags solopreneurs should be watching out for.
[00:34:40] Kevin Williams: Well, I mean, trying to build independent agents that are sitting off to the side somewhere and not integrated into your workflow is definitely a mistake.
[00:34:48] At least right now, because the agent development curve hasn't really caught up to this idea of orchestration agents, which is the idea that you have essentially, uh, a frontier level
[00:35:00]
[00:35:00] Kevin Williams: agent that's controlling the other agents, because that would help with. Document and knowledge transfer between different platforms and that will come.
[00:35:09] But at the moment, that's not really here. And honestly, we don't, we, we, we literally have dozens and dozens of different tools because we are always experimenting with things. The purpose of an agent is to be able to accomplish a relatively broad based goal with limited human interaction over and over and over and over again.
[00:35:30] Right. So I'll default back to n8n that Yeah. It's a workflow. What is an agent other than a workflow that has intelligence that's built into it. And when you're using a flow builder, you have the ability to tie whatever model you want into it that's the best model for the case. So you're dealing with really large context.
[00:35:52] Well, cool. You're gonna tap in Gemini and do whatever that process is, and then you want to hand it over and have better writing. I mean, I'm still, uh, one of the people who loves Claude for writing, so I'm gonna pass it through Claude and a different node, and then if that model changes, I can super easily just snip it off and put whatever new model I want in there, and it just gives me so much flexibility to get to that end result as we're sort of waiting for that. And we have dozens of workflows that we use internally and what will happen over time is we'll be able to string those together such that workflow one, the lead collector then leads into workflow two, the lead enricher, which leads into workflow three, the, you know, whatever it is, the outbound or whatever piece pretty seamlessly.
[00:36:47] And that that being able to switch out pieces has value .Versus this idea that there's, there's one platform to rule 'em all, or, oh, this, this agent builder's gonna do all of the things. You wanna kind of choose the best of. So another platform we use extensively is Appify. And Appify has agents, they call 'em actors that do very specific things very well, but they don't purport to do multi chain things.
[00:37:14] So for instance, we use an appify actor for LinkedIn. That allows us to do certain things on LinkedIn, but that's all it does. And then we use our own logic to do other things after the fact. You don't, you no longer are obligated to buy this whole package deal. So it's sort of annoying when these agent companies want to try and sell you that when they're, A lot of them are jacks of all trades and masters of none.
[00:37:40] Mike Allton: Yeah. Uh, this, this chaining of tasks. I guess that's one of the things I love Magai so much for. I can start a conversation in Magai with Gemini and my AI chief of staff are very high level, uh, strategic thinking. But yeah, once I get to the point where we want to actually write some copy, I gotta bring in Claude. Gemini no matter how much I train, it just isn't going to write as well as Claude, at least by today's standards, you know, so I can have a conversation with Claude inside of Magai. Like preparing for this podcast episode, I'm bringing in my assistant who is trained on my podcast Run of Flow, and my showrunner does all the guest research and that sort of thing.
[00:38:22] Now, when we're done recording and I've edited it, now, I feed the transcript back into that same chat, and I change to a different AI assistant that's trained to help me prepare for the promotion and the blog post and all that kind of stuff. That's all chained together inside Magai. But one of the things I feel like, Kevin, we've been talking for a while on this show.
[00:38:42] We've, we've named, dropped a lot of different tools and I think some of the folks listening, they might have already been drowning in tools and we've just added at least a dozen other things. Oh, maybe you should think about this. What, how did they figure that out? What is your recommended process for kind of auditing or even streamlining their own tech stack?
[00:39:03] Kevin Williams: I, I think, you know, good process engineering comes from identifying processes that are frequently used and create friction in an organization. So, you know, any Six Sigma friends that we have that are listening to the pod that'll resonate with, and that's their job, is to figure out where friction sits.
[00:39:23] Well. Well, that's exactly what we're generally doing with Ai when we're trying to solve problems, we're feeling the friction, we're identifying the friction. Ideally we have some sort of ROI that's applied to it, and then we're resolving problems, ideally using tools. The way that you go about identifying those points of friction is through self-analysis and logging.
[00:39:44] So, how often do I do this? Am I doing this once a day, once a week, once an hour, whatever it might be similar with the tools. And we sort of have a sandbox where we put tools that we're testing out and then we get a sense for. How much we're actually using them for what they're accomplishing. And sometimes the, the juice isn't worth the squeeze and the, they don't make the cut.
[00:40:08] Other times we only need a tool occasionally because it does do something specific really, really well. Um. For those, we really try and find options that have per use costs. You do see that pretty frequently, um, in some models, uh, voice agents in particular, um, assembly AI or VAPI, uh, those sort of, platforms.
[00:40:30] There's no monthly fee, so you only use it. You only get charged when you use it, so that's great. We'll keep those around forever. But, um, I am relatively cheap and SaaS bloat is real. So I look at my P and L and I go into that AI tools thing and try and figure out when, uh, the last time somebody actually used it for something that wasn't just dinking around, but was actually an ROA based objective.
[00:40:55] And you purge 'em. So, I mean, I don't necessarily know that's, that's not even an AI answer. That's just a, a, a, a, a subscription hygiene approach. Yeah. To deciding if you're actually getting enough out of a tool to keep it around. That said, we, we, I do feel per our earlier conversation about, you know, Gemini, Claude, the Open AI.
[00:41:18] I have all three open all day, and I'm, I'm not using Magai, I should probably check it out, but because they have different strengths, what you're going to see is a coalescence or a convergence of that over time. And, in the SaaS community, when one of the frontier models comes up with a new feature it might just completely wreck your business model.
[00:41:40] They call it being featured, so you have to pay attention. I have no doubt that Gamma at some point, will be under threat from Claude in particular. Claude can already do, uh, slide outputs, but once Claude can do slide outputs close enough to Gamma, why on earth would I be paying for Gamma if I'm already paying for Claude at the end of the day?
[00:42:03] So you have to pay attention for the direction that's going and be a little bit of aware of your own habits. As in, is this just a habit that I'm using these tools? And I did admit that for me, gamma was a bit of a habit, but, or is there a better, cheaper, faster way to do this such that it's more integrated?
[00:42:21] And that's sort of the, the key that would turn that would really kick me over, um, with integrated tools is I'm already living in those platforms. If they can now do this thing that I have to log into, open another browser tab and do something different and then have to like move the results around, that's really inefficient.
[00:42:39] So, it's, it's obviously part of the business plan of the Frontier model is to do this selectively such that they get more and more workflow advantages built into their own platforms.
[00:42:52] Mike Allton: Oh, I couldn't agree more. And it's funny, I'm. Honestly reminded of Alton Brown. I grew up with Alton Brown and his Good Eats program going way back a little bit there, but one of the things that he ingrained in me and his show was how much he hated to buy kitchen utensils and tools that were only there for one single purpose and for like 360 days a year, it just clutters up your tool drawer and that one day you bring out this one utensil to do this one task. He hated that, and I've always had that same now, you know, philosophy about any tools that I use, I love how you talked about auditing your time, folks, I'll, I'll actually download or a link to a download called the Scalability Audit, where it's just A-A-P-D-F.
[00:43:33] You'll print it out and it'll help you keep track every single day for a week or two of how you're spending your time. And the other thing that I teach in my community is for folks to build a little notion database and start keeping track of all the tools that they're paying for. How much they're paying, how often they're paying, and when was the last time they got a meaningful output out of that that impacted their business.
[00:43:55] That's like literally what I title the property in Notion. 'cause if it's been more than a month or two, you might wanna rethink that tool. But Kevin, for someone listening right now. You know, maybe they're starting completely fresh and they've got a clean slate and they want to build a strategic AI tool stack the right way.
[00:44:14] So based on your experience with your essential Daily tools, what do you think? Like the very first question they should ask themselves before they download anything, and what's the one principle that should guide every tool decision that they make going forward?
[00:44:28] Kevin Williams: Well, it, it, it goes back to my previous answer of what is the, what is the output that you're trying to solve?
[00:44:34] Um, is it you're trying to automate proposals? Well, there are tools that can do that for you. Is it you wanna automate, uh, content creation? Well, there might be tools to do that for you as well. And look at sort of the, the scale of them. And to be honest, specialty for most people starting off, they may not need these specialty tools.
[00:44:56] They might be just better off getting better at using LLMs with more advanced prompting techniques, using GPT projects or gems to sort of test out whether or not this is something that they need to invest more time in. And if they find themselves investing more time into it, maybe then it's time to look a little bit farther afield and see if there's a solution that answers their question.
[00:45:21] But it is about ROI. So starting backwards where, uh, the question I like to ask, uh, ask people generally in larger companies is, what are the most expensive human mistakes that you make in your organization, and it often has accuracy or customer communication or something like that, and how can you correct those as rapidly as possible using automations?
[00:45:45] So similarly, where are the places where you spend. Let's just call it low value time because a lot of creators really, really enjoy what they're doing. And cutting out the part of your job that you really like, just because you can do it faster maybe isn't necessarily the answer, but there's a lot of stuff that sits around that job that you can probably address.
[00:46:07] So where are the low points in your day that you're just like, oh, I really don't wanna do that. And is this something that I'm going to do more than once? You know, you're having an as a, as a solopreneur who's getting going every day is an adventure, right? So keeping track of when that adventure turns into the education, and I'll admit it's not always that three steps.
[00:46:28] Sometimes it's like, oh man, I've already done this four times. So okay, next time I do this, I'm going to the education step such that I can actually bring this to be an automation in the next bit. So you, you have to be aware of that because our days are just. Busy, right? You're just full of stuff and finding what are the things that you can actually solve and what are the things that you can't, I can't answer that for you.
[00:46:54] It depends on your own workflows and your own focus and your own happy place where you wanna be spending your time, and then removing the things that are getting in the way of your happy place. From a business coaching perspective. Try and make your happy place something that actually makes you money.
[00:47:09] But other than that, like trying to fo to, to keep you in, um, what, what you might call your zone of genius as much as you possibly can be and do away with all the other stuff. Um, I sort of joke that, uh, a lot of organizations, solopreneurs and beyond do a lot of things that make them feel busy.
[00:47:30] And it's sort of like productivity pornography where they're just like, yay, I'm, I made this spreadsheet and I made this report and I did this thing. Or, heck, I spent all day doing a workflow that I'm never gonna use again. Like that's not really that valuable. You're just filled your time with stuff and you felt good at the end of the day.
[00:47:48] Okay, great. But how can you put yourself in this place where you're going to be more productive in alignment with the things that you're good at? And dare I say it, the things that you actually like to do.
[00:48:00] Mike Allton: Kevin, I literally could not agree more with everything that you just said. It's the kind of topics we talk about all the time.
[00:48:06] The admin tax that comes with being a bottle a bottleneck in your own business as a solopreneur. I know folks, their minds are buzzing already with this conversation. You've been amazing. I know folks have more questions and they're gonna wanna reach out to you. Where should they go?
[00:48:23] Kevin Williams: Uh, I'm easy to find on LinkedIn.
[00:48:24] Maybe we can put that in the notes. And also Ascend labs.ai is our site. Um, we do a whole bunch of different stuff and we're always happy to help people out.
[00:48:35] Mike Allton: Fantastic. Thank you, Kevin. Thank all of you for listening. Of course, we'll have all of Kevin's links in the show notes below. That is all the time that we've got for today, but I will put that Scalability Audit in the show notes.
[00:48:45] Check that out. I think that should be your first step folks listening to help you understand where you're spending your time and where potentially AI or an automation tool might be able to help you out. Until next time, welcome to the grid.
[00:48:57] That's a wrap on another episode of The AI Hat podcast. I hope you've gained some valuable insights and actionable strategies to boost your business with AI. If you enjoyed this show, please subscribe, leave a review, and share it with your network. And don't forget to visit the ai hat.com for more resources and to connect with me.
[00:49:15] Until next time, keep innovating and remember, the future of business is human and it's powered by AI.
