TechPeace CEO Allan Edmond breaks down how AI agents, agentic coding, and personal automation tools are changing what one person can build and run, safely.
GROWTH PILLAR: AI & Automation
WHO THIS IS FOR: Solopreneurs / SMB owners / Non-technical founders / Leaders building systems
WHAT THEY'LL GAIN: A clear picture of how AI agents work, when to use them safely, and how to start automating your business without breaking anything — or leaking data.
Most solo operators are doing the work of five people. Allan Edmond built a business around fixing that.
Allan is the CEO of TechPeace, based in Grimsby, Ontario. He helps non-technical founders turn AI tools and custom software into working business systems — without the security risks that come from building blind. With over a decade of experience across agencies in Singapore, Australia, France, and Canada, and as one of fewer than 50 Glide Certified Experts globally, Allan has shipped 100+ apps for founders who had big ideas but needed someone to make them real.
In this conversation, Allan walks through the full arc — from why he left traditional coding, to the no-code era, to what agentic coding actually means for a business owner today. He explains what a personal AI agent is in plain language, why the prototype you built in Lovable might be a security risk, and how to think about local versus cloud AI models depending on your risk tolerance.
Key topics covered:
Allan also writes a daily French-language newsletter — La Toolbox d'Allan on Substack — packed with practical AI tips for business owners who want to stay sharp without getting overwhelmed.
Free consultation for business owners exploring personal AI agent setup: calendly.com/allan-edmond/free-consultation
Connect with Allan:
A big thank you to Blair Patton from Benchmark Insurance for the introduction.
— PARTNERS ON THIS EPISODE —
@Hive = Real-world business networking in person and online. Starter membership $99 CAD/year.
Profit10™ = Take the free Profit Snapshot and find out what's really holding your business back. Ten questions. No cost.
Canada Growth Network = Business connections plus GoHighLevel CRM for $47/month Canadian. Start your trial today.
NEXT STEPs
Bernie (00:21)
question to the audience. What if you could run your business like a full team and you're still doing it solo?
That feeling of being stretched too thin, too many tools, not enough hands, it's real. But something is shifting and today's guest has been living it from the inside. Alan Edmond, the CEO of TechPeace based in Grimsby, Ontario, builds web and mobile apps using Glide and no code tools that are powerful but easy to use. Alan works with solopreneurs and non-technical founders who have big ideas but don't want to get lost in complicated technology.
What makes him different? Well, he went solo twice after a decade working inside big agencies. He knows both worlds. Now he's back to coding, but this time through AI agents. And he helps his clients build custom solutions that actually fit how they work.
He also writes a daily newsletter over at LA Toolbox Diana on Substack. It's written in French. His tagline says, it says it simply, peace of mind powered by technology. Alan, It is really awesome to have you here today. Folks can't...
Can't see you but you're smiling man and it's a good smile.
Allan (01:25)
Thanks for having me, happy to be here.
Bernie (01:27)
⁓ absolute pleasure. Do you have a favorite quote or saying?
Allan (01:29)
so the quote I start to use more and more is basically, AI lets the solo technologist operate like a small product team. And it's not something that you will be able to do someday. It's like happening right now.
Bernie (01:40)
That it is. That it is. So you went out solo twice. What made you go, ⁓ know, agency, you've been great, but chow. Why did you leave twice?
Allan (01:51)
So the first time around was actually just opportunity. We were living in Singapore with my now wife back then girlfriend. And so she had an opportunity to be transferred to Australia. And so I just followed her and so basically couldn't keep the same job. And so I just took this opportunity to start doing my own little agency for small businesses and we're building websites. So that was the first time around. So that one wasn't really...
a big move, was just because I was losing my job anyway and so I was like, huh, might as well just try start something on your own.
Bernie (02:25)
Wow, you're a traveler. That's even better. This is really great stuff. So over the years, I've met people that are developers that do coding. And right now we're talking about no code or agentic coding. What caused that shift? Why did you move away from sitting at the keyboard going, strange letter, strange letter equals strange letter?
Allan (02:45)
Exactly. Yeah. So I have a background in engineering and so I started as a developer back in the 2010s. And so at the beginning, during school, college, it was great to code because you can see, you can build stuff. But when you start being a developer by trade, you will start realize you spend more time troubleshooting and fixing than actually building. And so it was a huge pain for me. I started not liking to code anymore.
And that's one of the main reasons I switched to no code in the first place, because no codes is basically what you see is what you get, clicking around, dragging stuff around, and you end up building and seeing what you build. So that was great. And I went for no code for a good four or five years. But with the evolution of AI, I started playing around again with agent decoding since last Christmas.
You see everybody now is using code code and being able to do stuff. So I started on my own as well. And from there, I realized that I can go back to coding, but without the pain of the thing that made me leave in the first place, the troubleshooting and the fixing and even not the writing. So that's what basically helped me go back. And, I really enjoy it because I'm no longer the minion fixing the bugs.
Bernie (03:57)
I can appreciate not being the minion in that instance. I truly can. So it's a tech shift, but it's more than a tech shift. It's a business model shift too, right? So it's time spent, and again, just kind of going back, right? Being the minion means now you're laboring away as opposed to something else that's building it for you. Do you think this is going to get even easier as time goes on? Or is it still going to be the province of someone who goes,
Yeah, you can do your own, but if you want complicated, you'll need someone that has a little more under the hood, a little more knowledge base.
Allan (04:29)
Yeah, so that's something I'm going to be talking about more and more, especially on LinkedIn. Now that everybody has access to agentic coding and it started always lovable really, which was a website building and they can build web apps that are great. And you can quickly spin up a UI that functions more like a prototype. That's really convincing that has transactions and everything, but it doesn't have necessarily all the backend architecture and the security foundations.
that engineers spent the entire training, learning and applying and being corrected by professors. So I think, and also you hear a lot of stories. I was shared a post by somebody in the parenting space explaining how there was a tool she found where she managed as a user to see everybody else's emails and names and even kids' information. So that becomes the threat now because people can build.
tools and platforms and web apps and websites. But AI only does what you tell it. And also AI has an incentive to make you happy. So it will eventually cut corners to show you something that you ask so that you are satisfied and you keep using it and not go away, essentially. As opposed to us engineers and software developers and the technical guys who have
the background knowledge of knowing how stuff should be built and the security implication, the architecture, the importance of architecture, the data model, the security and everything. my say is basically, if you are doing a tool for you, for your kids, something that you only have access to and it's maybe just on your laptop, go for it, man. You no longer need the developer and that will unlock the power from so many people and empower so many people.
As soon as you want to have software that's going to be used by somebody else and other people, should at least, I don't say you need to hire an engineer in the first place. For sure, start with a lovable prototype, start playing around with the idea, get feedback, get people interested. As soon as you are serious about launching, you should definitely get someone technical with you to oversee and review what you've done and make sure you don't end up having a breach or security hack or something very bad that can happen.
Bernie (06:42)
I was at a presentation last night and there's a company, they're out of Toronto and they're using AI. The product is called Kitten Capsule. So they got a free plug here. And what it is, it's about getting the stories from the parents or you, anyhow, it's a family story and it's a place that's a repository. And what was happening was they reached out to Algonquin College and they had some students help.
build around like doing the next iteration of what's happening. And I listened to one of the students and you just kind of highlighted the engineering piece that you needed. The young lady was explaining, she says, yep, it took up memory on the computer as opposed to or in the processor as opposed to over there. And I guess if it's a light use, you really might not care. But if it's drawing on ⁓ heavy resources, it might slow your machine down.
When you're coming across, if someone gets an app, for example, how do they know it's, I'll use the word tight enough, but how do know it's good enough? In other words, it's not leaky and it's not going to make your machine wobble when you hit, oh, I'm going to hit this button. next thing you know, leave your computer alone for 10 minutes. It's sorting stuff out as opposed to using the resources in the cloud.
Allan (08:03)
So there are two things. There are web apps and local apps. So if you are building something on your machine, and especially with the tools we have today, they are very memory hungry. because they, AI, because we could forget this, I just like to repeat it. AI is only good at predicting the next word it's going to say. in order to do it, it's not thinking really. It's just consuming text and reproducing text, whether that text is.
a sentence is going to give you or code or whatever, it's always producing text. And so in order to predict the best next word, it has to consume as much data as it can. And so that's where the memory and it's caching some memory over there. So it's basically building its own brain as file text, text file, sorry, inside your machine. So that's what might take up a lot of space. To this,
There isn't really a big solution right now in the way it's done because that's what makes it powerful is because of the context it has. But it's basically just sweeping the data as much as you can and just give it instructions to kind of clean itself up. So that's for the software hungry memory on your machine. On the web part where it's about leaking data and being hacked, there is never a 100 % chance you will never get hacked because hackers are very skilled people.
And if somebody is determined to hack you, will find a way. And even if it's not your infrastructure necessarily, it could be phishing attacks. So there are so many ways to do malicious things that there is no 100 % chance you will not be affected. However, there are many foundational and very not basic, but foundational is the word we're actually, approaches to build software. And so...
Building software is not something that just appeared with AI coding. Engineers and developers have been doing it for decades. so along the way, again, to go back to this is our trade and our training, we've been building best practices and habits and rules and guidelines along the way that we know and we learn and we follow. So that's what you need to follow. And it's out there. You have documentation online, you have even the...
the backend services and the frameworks or whatever the tools you use would produce or have documentation on how to follow these guidelines. But if you're not a technical person, again, how do you don't know what you don't know, basically. And so again, you should know you should focus more on your idea, your vision and what you want to achieve and get somebody that's worried about, where is my database going to be and what are the rules to access this role? This you shouldn't.
try to learn every single thing and just get surrounded by the people. I mean, that's the same of the company, right? One CEO doesn't do everything. You hire people to do the different departments and skill sets that you need. So it's the same thing. It's not because code is cheap that you don't need nobody anymore.
Bernie (10:46)
So with that in mind, if I'm a business and I'm going, listen, I want to run an bot, and I'm going to call it a bot, or a Gentic. And I want to run it just in my local environment. In other words, I get an email, it's all bossed, there's an email, I've read it, and it should go here. Or some sort of transaction happens. Even when I'm doing the podcast processing, I have a series of steps, both for
bringing people into into the into my space and then you're doing doing the recording and then post production and you know it's a series of things so it would thread things I could see a bot threading things through for me so I'm not sitting there and then I just look for errors in case there's a typo or something like that is that is that something then less less disconcerting versus something that has to face outwards
And to go back to like, know, other people are using it so that it's not leaky or it's not doing something silly. So if a business wants to do something like a solopreneur or a small, you know, five to 10 men, five to 10 person shop. So that's something that they can do themselves and then maybe then reach out to you or reach out to another expert to help. Okay. If I got it right, can you take the screwdriver and tweak the knob so that I'm in the right spot in the right time?
Allan (12:01)
Yeah, so I read an article, I think it's a co-creator of Django, which is a Python framework. So a very skilled, very known guy, very smart. And so I think it's him. He wrote an article about the rise and the stress of having personal AI agents. what you mentioned is basically an AI agent. So you give it a task and it will be able to do things for you and not just giving you text and giving you more work in the end.
The issue is, the biggest threat is when you have three things. You have access to your information. I'm worried I'm going to give wrong things. But basically, the ability to communicate outside, access to your own information, and getting instructions. I think that's the three things. So if you have all three, you're in danger. And that's basically the promise of AI personal agents.
There is no way to circumvent that entirely, but you can minimize and mitigate as much as you can. So for the example that you gave of a solopreneur wanting to have stuff done, if you remove the ability to go outside of your company, then you have basically mitigated the entire risk. So it has access to your information, which are your emails, and is able to do a bunch of things. But because it's internal, there's no way of leaking outside. And so you could have your agent running on
a machine, ideally not yours, you never know, and just give it access to your email or rather what I do even for my own agent is forward email to it or give it an email and just forward specific emails to it so it doesn't have access to everything. So again, just to minimize as much as you can and work with it to build what now we call everybody calls it a skill.
which is a set of instructions on how to basically, so I'm going to send you the link to the podcast. The podcast source video is in this folder. These are the instructions that you need to follow to read the video, pass the text, transcript, build these, these, do that, do that. And at the end, create a package for me to do XYZ. So you can definitely do that and people are doing it. My coach has done that. He has a podcast that he's been running for a long time.
And he's having his agent to do all the extraction of the transcripts for him. And even did that for the backlog of years of transcripts he had. And he was saying basically, I'll do that someday, which he knew he would never do. But because AI can do it while he's sleeping, he managed to do that across the board for the last five years. So definitely something happening today.
Bernie (14:25)
Well, you perked my interest there going backwards in time. Sorry. It's like, hang on, hang on. That's a bit of mana from heaven. So if you have an AI agent and everything I'm understanding about large language models, you know, where everything's and I know there's other other elements. think RAG is one of the other ones. I might be using the wrong acronyms, right? Does that stuff?
Does the intelligence reside on your box? Does it have to reside in the cloud? Is it a hybrid? If I'm thinking from a security point of view, I want things in my four walls. But if I don't care because it's nebulous and the cloud's OK too because I'm not paying for storage, which is better? Where should a business store their smart component? I'm going to call it that for now.
You know, where do they put their brains?
Allan (15:14)
So about the brain, so the files and everything obviously on your machine, the brain you have a choice, but every choice is pros and cons. You do have today quite powerful models that you can run locally. So that would be something like a seven gigabyte file or whatever system you have to download because it's billions of parameters, whatever. I'm not an expert in AI and LLM, so I won't go down the rabbit hole, but basically you could download an entire model on your machine and have it run.
and ask it to do what any LLM does. The issue is somebody built that model with one training set, whatever, whatnot process they do, and created that artifact or that box that you can download. On the other hand, and that's something you can download, most of them are free, I believe, and you can just download it, run it, and it's completely isolated. It's your machine, no leaks possible whatsoever, no training on your data. Everything is local.
On the other end, you have OpenAI and Anthropic, who are AI providers. So their business is to make LLM models. And so they keep releasing new versions. so training more and making it smarter and releasing new versions. think yesterday, Cloud released another version of their most powerful model that's going to be even more powerful and even more expensive. And so...
It's really depending on what are your requirements. If you need the maximum reasoning, then the cloud models will be more powerful. But the trade-off is you are sending your data out. Most of them say they are not training their models on your data. You have no way to make sure that's not happening. But that's what they say, because you pay them.
Technically safe, you just have to take the award for it. But you have maximum reasoning. If you are a security absolutionist, then use a local model, but the power of the brain is going to be lesser. So it's basically a choice to make. so for my own personal, so I'm using OpenClaw for my personal agents, just one flavor of all the personal agents that we have today. And I do have a local model running on the host. But as I was...
asking the LLM itself which model should I use for which task. was like the downloaded model is great for writing the brief or reading x, But if you want to interact in the chat and have actual reasoning to figure out what to do next, this model might be too limited. So you might want to do the cloud model. So it's really depending on the task.
Bernie (17:39)
Something that came up in previous chats and I've experienced it in the 80s. It's a piece of software I was using and I really getting used to how to program it. It was called Dbase 3 Plus. I it was, just, you you can do kind of little bit of coding and I'm not a programmer. won't know, like an engineer would be able to say, you store it this way, you save space, you do it this way, it speeds things up. And then one day,
It disappeared. think Microsoft Access was the next level. Almost like word perfect to word. There are revolutions like that. And to me, all of a sudden, I realized that software is disposable. Is this something that's built, whether it's the back end, if you're using an Anthropic or a Chat GPT or a Lama or, you
take your pick of the poisons. Is it something that's, here today, it might morph an oar tomorrow, like Sora, gone.
Allan (18:38)
With the one, I think it's a one billion deal with Disney, like just next day. ⁓ by the way, we're not doing that anymore. Bye.
Bernie (18:44)
So is this, would this be considered disposable software or is it something like, you know, if I build it, right? If I build this house and I put, you know, build it out of concrete, it's not going to disappear too easily, right? Does it have this, if you build it in a Agentic AI, is that something permanent or is that something you have to have another revision over time, like every five years kind of go back and go, okay.
We've been running this now. Here's the upgrade or a different way of doing it.
Allan (19:13)
So for personal agents, think it's going to be, I mean, many people think it's going to be the future of the world actually. And something that piqued my interest and made me dive into OpenClaw so much was that the Chinese government is having a huge propaganda to have every single citizen get up to speed on OpenClaw and have their personal agent. If an entire government is pushing its population to use something, that means something, especially China.
which is kind of at the cutting edge of technology. So that kind of made me like, you know what, if they are saying this, you should maybe get inside and see what it's like. So the concept of having agents is basically having secretaries. Everybody in, we've seen movies, we know high, very top executive having secretaries and personal assistants. You need to be in a certain position in your business or in your life to have somebody working for you just to take care of your life.
now is AI, everybody will be able to have their own agent that helps them. So the concept will not disappear, but the mechanics of it will evolve. And I think the basis of OpenCLON, some people are doing their own version of it, and most of the big providers are doing their own version of it. So the concept itself, the implementation changes, but the concept doesn't change, which is having a brain, what we just touched, which is the LLM. So this keeps changing every...
every two weeks and pick your poison like you mentioned. And you have the three major ones we know, Google, Entropic and OpenAI, but you have the Chinese ones, you have other ones. have like, even in OpenCLOS, you have a list of I think 30 model providers you can pick. So this one just go wild, see which one you like and be okay to change next week. So that's the brain. Then you have the memory system. And this is just basically writing files to some...
the box because again the brain cannot remember everything same as us so for us we tend to forget for AI it writes files so it doesn't forget and then the next piece is being able to take action that's the skills we touched which are again a set of instructions and using the brain having access to memory and being able to do things with the skill you can have your personal hygiene being able to remember think and act
And that will not change, but the brain will change, your memory evolves and gets bigger essentially. And the tools that you are able to, the skills and the actions can evolve with you as a business owner or even an individual. Because today you need the help to do your taxes. Tomorrow you need to transcribe your podcasts.
And next week, you're starting a new venture and you need assistance to do X, Y, So basically, it keeps evolving because it keeps having needs that you can get AI to help you fulfill. So it will change in the implementation and in practice, but the overall concept will remain.
Bernie (21:58)
if I'm in business and already using certain software, like I've mentioned, like, you know, using the Microsoft Office Suite as an example, or it can be QuickBooks or, even the video editor I'm going to be using to edit this podcast, Riverside FM. Does the AI, the agentic or the AI agents, will they integrate well or is that
Is that something that I should have an expert kind of hand hold me through and find out how to make that pardon me or the expression application interface API, right? And, you know, I can say it, but you know, have I done it? No. So, you know, at a certain point time, I'm going off, I stick my finger in here. Is it going to break things? Is there, is there things that businesses find when they try to do the inject the AI into it? That's not part of the package.
where things go sideways.
Allan (22:45)
So the thing is, it's very early stage. so it's early adopter level. You have to get your hands dirty. It breaks every now and then. And so for business owners that need something reliable, I believe at least for now, you should get somebody that's been doing it for themselves and for other people that knows where it broke and knows how to fix especially because even if
Let's say I've worked on six different use cases for myself and my clients. Your new use case might be different, something I've never done. And so we might try and it's still going to fail. But because I've been failing so many times with my own personal news aggregation that keeps breaking every morning because some RSS feed stopped working, I know the steps to take to investigate and fix it.
without having to ask AI to tell me when sometimes the AI doesn't even know what happened because otherwise it would have done the mistake in the first place. So for now, I guess you still need help because the early adopters are the ones who can have the experience. It's not even experience because it's like a few months, but still experience because some people lose sleep against personal agents. But...
Again, the big providers are working relentlessly to release consumer grade, more secure version. So soon, everybody will be able to do some version of their personal agent or their company level agents that's secure enough. So you could wait or you can just play with a few very simple use cases. Just start getting used to it. The earlier you get in, the more...
I'm comfortable and confident you get to use it and the more ideas it gives you basically. yeah, everybody should start now. I just know that it's not going to be super easy as of today. Still very hands-on mechanics.
Bernie (24:31)
So don't pick your critical software, mission mission, critical software that goes sideways. Pick something in the periphery and learn from that and then go from there. But you know what it also sounds like then they should perhaps ⁓ reach out to you. So what type of clients do you engage? If I'm out there listening to this podcast, I'm going, know, hey, I agenda and I want to make it work well and I want to be there. Right. I want to be the top of the curve.
Who is that person for you?
Allan (24:58)
So most of my clients come from the development side currently. and I'm helping them. So usually it starts with a brief, I want to build an app, blah, blah. Before it was code, no code, now it's AI coding. And as soon as we release the app, most of the time people have extra needs. And so we start discussing during the project, there's also this, do you know somebody or this is something you can help with. And that's kind of where I expand my scope and kind of give a CTO level advisory.
companies. So always, you have to start somebody, I can just say, I'm going to help you do your agents, nobody really looking for somebody to help them build agents, not many people, some people, but not many. But software, many people know they need software. So usually starts from there. But it's, it's solopreneurs or small companies that, again, big staff is short. And so you need to get the most out of the few
employees or collaborators that you have. And nowadays, and that's something I've been realizing, like I start preaching, is that you actually won't need to hire so many people anymore, which is a good and bad thing. It's another debate, but for the basic stuff and for the repetitive stuff, the annoying stuff, you will not need to hire entry-level assistants anymore. It's going to be done by AI.
Bernie (26:17)
And if you have zero coding background, you know, just like, you have a bit of a venture spirit in you, where do you start? If you're looking to build your own AI agent.
Allan (26:27)
So it depends. If it's a website or web app, I would say go lovable because they have a very good design approach. And so that helps you visualize quickly. So that's for web apps. If it's for actual software, it still might be web in the end, but you are not really sure. But you have files you need to handle, or you are trying to figure out how to sort out your Stripe exports.
⁓ because you have to do taxes and you are using GHL and it absorbs the tax. So you need to sort this mess out. You can use Cloud Code or Codex, which are identical coding versions of, mean, software of Entropic and OpenAI. And it always starts with a prompt. You give it the files, start with a prompt, start always with what you're trying to achieve and ask it to come up with a plan on how it would...
help you address that. so review the plan, make amendments, correct the few assumptions that were wrong, and from there give it a shot, say start with X. And also most of the time when you ask for a plan, it's going to give you like a phased approach as well. And so you can always start with phase one, see the results, play with it, and move slowly. But even slowly is actually pretty fast because within 30 minutes, it's going to give you a pretty solid first version.
of what your dreams are.
Bernie (27:46)
30 minutes is slow? no!
speaker-3 (27:48)
Yeah.
Bernie (27:48)
So if I build my own AI agent or have someone build it for me, that's custom map. Is there off-the-shelf products? Which is better then? Should I do one that's tailored to me or one that's like, yeah, it's got most of the security, most of the architecture built in, but it's not quite custom?
Allan (28:06)
So currently you don't have... Clo... Anthropic is trying to... So OpenClo was the first technology piece that really figured out the personal agent application. It's really not for everybody. Very hands-on, very not stable, very frustrating.
but super rewarding and full capability, full flexibility. You have no limits. You do. You manage everything yourself. Anthropic is adding more more bricks to their cloud offering, the cowork piece and the code piece to kind of do this kind of delegation of tasks and working on your computer on fires and so kind of doing what OpenClo has opened. just bringing everything in.
And I'm sure OpenAI basically hired the guy who created OpenCross. So is they going to release their own version soon? Obviously. You have Perplexity that created their own version as well. That's called a Perplexity Computer. So depending on your VR level of how much hands-on you want to be, that's why you can pick your poison, keeping in mind the less hands-on you are, the more money you have to pay to the provider.
So you pick your poison, there's still a poison to swallow at the end of the day. But I mean, for me, technically, I'm a technical guy. So I actually enjoy going through the pain of fixing and figuring out and understanding. So that's just something I enjoy doing. But most people won't because it's just frustrating. So in that case, would say go to an off the shelf, see the capabilities, read, not necessarily the documentation that's a bit useless, but ask it, is this something you can do? Yes, no.
Okay, what do you recommend? I Cloud Cowork to do an assessment on the task. I forgot what it was, but I was like, I'm looking at doing, was trying to automate my bookkeeping because I hate going to the, it's not ⁓ QuickBooks based on the software. was like, is this something you could do? I give you, you have access to my email. You can just pick the receipts and go to the thing, go to transactions, update, categorize it. And it was like, ⁓ not really because you know, I'm more.
managing the files and those are, ⁓ that's a bummer. I thought you could do everything. So that's for example, just asking Claude, I want to do this. Is it something you can do? Yes, no, no. Okay. I still haven't done it even with my open Claude, but I have automated my insurance claims, which is fascinating. So it's basically literally going to the portal. So I still have to log in manually and there is a two factor authentication. So I get the code on my phone. So still security. I log in first.
Then I forwarded the emails with the claim statement, which is the month's worth, and it will pass the PDF, extract the five appointments I had with the price and the date and the blah blah blah, go to the portal, click around and submit the claims for me. Or even prepare and at the end I just go and click once. And so this is the kind of things you just experiment, start with one of them and see how it can do, what it can do and how you like it.
Bernie (30:51)
So five years from now, what's it gonna look like?
Allan (30:54)
Five years from now.
Bernie (30:54)
Yeah, it would be.
Allan (30:56)
I mean, two months ago, I had no idea that this would be today. So five years flying cars and we are on Mars. I don't know. And so, but definitely I'm seeing people having everybody having their own like butler for themselves. And so your own and some people are already doing this with OpenClaw. So you have an all personal assistant. You have an assistant for the family.
You have an assistant for work. again, you can create, as I watched a podcast where a lady said she had nine agents. That might be a bit much, but she was saying one for her, one for her family stuff, the kids and stuff. One because she's starting a new venture, she needed accountability. One for X, one for Y, and she's building one for her kids. And so that's going to be what we have. And on top of that, I think we're going to have the physical version of it.
which are what Tesla is working on, the Optimus robots and stuff. So you're to have actual robots that can impersonate these agents, the brains, and go and do things for you. So get your groceries physically. And I'm seeing, maybe not five years, but I am seeing a world where nobody goes to supermarkets anymore. It would be like the Amazon.
Not factories, the Amazon warehouses. It's all robots running around getting the items and getting the items. It's not organized for a human to be able to navigate, but the robots know it can navigate amazingly. And so all the items are for robots and other robots come and pick up the lists of items and you just wait home and get your stuff done. And that's it. Nobody has to walk down aisles and push carts to buy food.
Bernie (32:34)
So for yourself, what type of business owners or collaborators are you looking to connect with? If someone, again, is listening to this podcast and they are either people that could be the client and or whether they need the product or someone you might do a joint venture with, who would be the ideal people?
Allan (32:51)
Yeah, so I work mostly with founders that are not technical, because if you're technical, then you might as well, you can figure out by yourself. And I have a bunch of friends doing that. But my clients are all not necessarily solopreneurs, but founders that have huge ideas, that are aware that there are possibilities, but they don't necessarily have the time to spend on it. They can do a little bit and they
They try to poke at stuff and investigate, but they can't. They have a company to run, essentially. So you can't stop for a week trying to figure out how a lobster can work with your balance sheet. So this is the kind of people I help, non-technical founders that still have time to sit down and bounce off ideas and try
play around, but needs to be able to delegate at some point the overall execution of the work. And that could be any industry. So many people say you have to niche and focus on one vertical. I actually enjoy, I have clients in famtech. So famtech does something apparently now around household management and alignment between partners. So that's one of my longest clients.
I work with people in real estate, I was with people in finance. So I like to have the diversity of the industries. But the person is always the same. Founder, small team. I want to play with stuff. I don't have enough time. I need somebody to help me get things done.
Bernie (34:22)
Perfect. You're French and you lived in France, but that's where you live least. What's the backstory of that?
Allan (34:24)
Yes.
Yeah, I'm from, I'm French definitely. I was born in Martinique. Martinique is an island in the Caribbean. We were a former colony in our French department. So we are born French. So I was born there, grew up there up to high school. And I graduated and went to uni in France. So that was five years originally because masters in French are, in France are five years. But I left to Singapore in my, in formal last year.
So spent four years in France, lived in Singapore for eight, then moved to Australia for one, and then after that moved to Canada. And I've been in Canada since 2021, so that's now five years. So took over France also. yeah, 17 years in Martinique, four years in France, and then Singapore and Canada now.
Bernie (35:17)
Welcome, welcome to Canada. It's good to here Just to kind of wrap things up, Alan writes a daily newsletter called La Toolbox d'Alan. It's going to be en français, so if you might have to have a translator handy, maybe you'll have an AI agent to do that for you. But it's packed with practical AI tips for business owners who want to stay sharp about staying on top of things without getting overwhelmed. All right.
And you can find it on substack at LA or LA toolbox, Dallin. That's D apostrophe Allen dot substack dot com. Allen's also offering a free consultation for business owners who want to explore setting up a personal agent, but aren't sure where to start. So to book that, reach him at Callanly dot com slash Allen, A L L A N dash Edmund, E D N O D.
slash free consultation. And if today's got you thinking about, jeez, I might want to have AI help me out, get your own personal Jarvis, so to speak, if you're not in the Marvel Comics section, reach out to Alan. And it's not just personal, but also business, right? I would like to send out a big thank you to Blair Patton from Benchmark Insurance, who made the introduction between Alan and myself. So thank you so much. And you know what?
Alan is solving some of the quietest problems in business right now. The solo operator, you know, who's doing work of five people. I can attest to that, you you can't do it all. So he'll save you some time and he'll save you some grief as well. And he also structured the app so that you'll stay on top of things and you won't have a leaky sieve. Head over to techpeace.ca.
and take Alan up on that free consultation. You'll find the links in there.