# how to transition from ai automation to agentic workflows

## Метаданные

- **Канал:** Nick Saraev
- **YouTube:** https://www.youtube.com/watch?v=5rNu19PfgFg
- **Дата:** 05.02.2026
- **Длительность:** 10:27
- **Просмотры:** 38,359

## Описание

🔥 Join Maker School & get customer #1 guaranteed: https://skool.com/makerschool/about
📚 Watch my NEW 2026 Claude Code course: https://www.youtube.com/watch?v=QoQBzR1NIqI
📂 Download all project files here: https://nick-saraev.kit.com/aaa09595c6

📚 Free multi-hour courses
→ Claude Code (4hr full course): https://www.youtube.com/watch?v=QoQBzR1NIqI
→ Vibe Coding w/ Antigravity (6hr full course): https://www.youtube.com/watch?v=gcuR_-rzlDw
→ Agentic Workflows (6hr full course): https://www.youtube.com/watch?v=MxyRjL7NG18
→ N8N (6hr full course, 890K+ views): https://www.youtube.com/watch?v=2GZ2SNXWK-c

Summary ⤵️
simplest way to transition to agentic workflows from older school no-code platforms like n8n, make, zapier, etc.

My software, tools, & deals (some give me kickbacks—thank you!)
🚀 Instantly: https://link.nicksaraev.com/instantly-short
📧 Anymailfinder: https://link.nicksaraev.com/amf-short
🤖 Apify: https://console.apify.com/sign-up (30% off with code 30NICKSARAEV)
🧑🏽‍💻 n8n: https://n8n.partnerlinks.io/h372ujv8cw80
📈 Rize: https://link.nicksaraev.com/rize-short (25% off with promo code NICK)

Follow me on other platforms 😈
📸 Instagram: https://www.instagram.com/nick_saraev
🕊️ Twitter/X: https://twitter.com/nicksaraev
🤙 Blog: https://nicksaraev.com

Why watch?
If this is your first view—hi, I’m Nick! TLDR: I spent six years building automated businesses with Make.com (most notably 1SecondCopy, a content company that hit 7 figures). Today a lot of people talk about automation, but I’ve noticed that very few have practical, real world success making money with it. So this channel is me chiming in and showing you what *real* systems that make *real* revenue look like.

Hopefully I can help you improve your business, and in doing so, the rest of your life 🙏

Like, subscribe, and leave me a comment if you have a specific request! Thanks.

Chapters
0:00 Introduction to AI-Generated Workflows
1:18 Setting Up the N8N System
1:44 Building for Clients vs. Personal Projects
2:24 Simplifying Complex Systems
3:24 Summarizing Workflow Tasks
3:54 Testing the Email Reply System
5:18 Benefits of Automation in Workflows
6:07 Steps for Implementing Agentic Workflows
7:38 Overview of Modal and API Setup
9:18 Transitioning to Agentic Workflows
10:06 Join the Maker School Community

## Содержание

### [0:00](https://www.youtube.com/watch?v=5rNu19PfgFg) Introduction to AI-Generated Workflows

Here's how to set up and sell AI generated automations or Aentic workflows. In just a few minutes, I'm going to show you a method that uses all of the same approaches that you guys are currently using to sell simple noode automations in make. com, nadn, zapier, or other platforms. And you don't need to change more or less anything about your workflow except for one tiny little tweak that'll take you only a few seconds to implement. How do you do this? Well, the key is to wrap this agentic workflow inside of a pre-existing noode automation. The reasons why are one noode platforms like nadn, make. com, zapier, etc. have already solved deployment for us, which is a massive headache when you start selling custommade automations, python scripts, and custom servers. In addition, as you see here, it's highly visual. So you can take this super complex scary thing which is you know serverless systems and then package them in a simple way that clients understand. We have nice colorful blocks and arrows between them. And the third is that this is much more maintainable because of built-in error handling of these platforms, built-in retries, and then the ability to build in pretty cool flows. We can deal with what happens when one of these things goes wrong really easily with the method I'm going to show you. So as mentioned, you don't need to change anything about your systems or about how you're approaching the work. You just need a claw. md file, which I'm going to give you guys down below, and then an ID of your choice

### [1:18](https://www.youtube.com/watch?v=5rNu19PfgFg&t=78s) Setting Up the N8N System

along with a noode platform. So, just so we're all on the same page here, this is an NADN system that really, really broadly just replies to my emails. Um, I run many campaigns simultaneously on platforms like Instantly and so on and so forth. We're basically just cold emailing people to sell a service. In this specific instance, it's for a dental marketing company that I run called Dental Connect. We do about $2 million a year. And this system was not trivial for me to build. This probably took me somewhere between 3 to 5 hours.

### [1:44](https://www.youtube.com/watch?v=5rNu19PfgFg&t=104s) Building for Clients vs. Personal Projects

Now, this is my own company, so it's a little bit different. But pretend for a second I actually wanted to build something like this for a client. Okay? Somebody paid me money to do this. Now, I could go through the rigomeroll of putting together a very complicated system like this. I'd have to do my own error handling and bug fixes. I'd have to build in the processing logic and the AI agent and the prompts and so on and so forth. Or I could just turn to my AI agent and have him do it for me in a few minutes. The key is in combining the best of both worlds. Essentially what we do is we use nadn uh make. com or whatever for the trigger and then the action and then we abstract everything else to our AI. Okay, to make a long

### [2:24](https://www.youtube.com/watch?v=5rNu19PfgFg&t=144s) Simplifying Complex Systems

story short, what I have here is I have the same system but only the very beginning of it and then the very end of it. You know, at the very beginning I'm getting the response. I'm looking through all our conversation history and I'm just like aggregating it to present the data easier to AI. And at the end, what I'm doing is I'm just sending a reply. But if you guys think about it, if we just zoom out, I had like 10 or 12 nodes in between this and this. Where did all those go? Well, this is where our AI is going to oneshot that entire system, significantly increase the simplicity of it, and then allow us to build it in with, you know, approaches that we're already familiar with. How do I do it? I've opened up Visual Studio Code here and I've just given it a prompt saying, "Hey, I have a flown nad that responds to emails. I want you to build a system that works as follows. " What this complex system did back here is it did a bunch of data formatting as I was mentioning. It extracted campaign IDs, looked a bunch of reply templates up in a knowledge base like this, which is pretty clean. We then even got like a bunch of examples of real replies that we've given to other people, you know, our business name, how we do the

### [3:24](https://www.youtube.com/watch?v=5rNu19PfgFg&t=204s) Summarizing Workflow Tasks

qualification, and so on and so forth. Basically, we searched all this stuff up dynamically before passing it into an AI agent, right? like it's a fair amount of work. What I did is I basically just summarized that you're going to receive an array of all the conversation history between myself and the prospect. You must then consult a database to retrieve a knowledge base and then email templates. You then take all the conversation history, the knowledge base, and the email templates. And then you fill in and reply using this information. Stick closely to the email templates. You should return the email copy and plain text with Neolens Opus 4. 5. So I just sent this. Okay. And keep in mind that this is like 90% of the 3 to 5 hours that I was spending building

### [3:54](https://www.youtube.com/watch?v=5rNu19PfgFg&t=234s) Testing the Email Reply System

this workflow. Okay. About a minute later, we've now finished our email reply system. It's given us um down here a ready to use curl command. So if I just copy this over, I can then go into my system, select HTTP request, paste all this in, it'll automatically map everything for me. Okay. And then all I need to do is just test this myself. So if I click execute step, I've given it a bunch of information, a bunch of instructions, and a bunch of examples of what I want it to do. And it's now handled all of that intervening middle stuff for me. It's set up a knowledge base. bunch of reply templates. It's done a bunch of research on the person and their company. It's combined all this stuff together into an example reply. What I can then do is I can just connect that to my last message. And all I really need to do is just map the reply, right? I think we're doing plain text. So, I'm just going to do this here. And then I'm actually going to run the whole thing end to end. Okay. And then just giving you guys an example here. Um the first email was basically, "Hey, how's it going? You know, I love your stuff. Big fan. Here's what I do. What makes me different? " And then, you know, do you want to jump on a quick call? My hypothetical response was, "Hey, this sounds interesting. Could you tell me more? I might be into this. We tried some similar stuff in the past. Hasn't worked out yet. " And then here's the reply that the automated system put together in just a few seconds. Hey there, thank you very much for getting back to me. We're called DMC, working out of Calgary, Alberta. There's no front charge. Work with a variety of clinics. I also run a YouTube channel at around 200k subs, blah blah. So, I mean, like this, the reason why this is valuable is because this is the same quality of the workflow that I

### [5:18](https://www.youtube.com/watch?v=5rNu19PfgFg&t=318s) Benefits of Automation in Workflows

put together procedurally. It's just I can rebuild this now in like two or three minutes. What we did is we took the logic that previously would have taken us significantly more time and energy to put together. Okay, all of these nodes over here, uh, the tooling, you know, the knowledge base, the database schema, the architecture, everything like that. And then we just gave all that to the agent and we've turned it from one that might have been 15 to 20 nodes into one that realistically is just like four or five. If it's not super clear, the key is just abstracting away all of the heavy lifting to the agent. You then just get a URL. And the benefit to having a URL like this is again, it's still within the same flow. you still get credit for building it. It's still visual in so far that you know there's boxes and arrows and so on and so forth that go between them which I know doesn't sound that important but if you've worked with small businesses for any reasonable amount of time you know stuff like that's essential to getting any sort of like congruence buyin understanding of

### [6:07](https://www.youtube.com/watch?v=5rNu19PfgFg&t=367s) Steps for Implementing Agentic Workflows

what's going on in terms of steps for both this workflow and any other ones all we really do is we start by triggering something in NAD maker etc. So, this is going to be like a web hook call. It's going to be an inbound event coming from your CRM, whatever. It's just stuff that you you're all probably used to in the no code building. So, it's not that big of a deal. Once we're done with that, we just send the data to an agentic flow. Um, I have a bunch of templates set up that you guys can just copy into your own workspace with like a pre-unctioning claw. md and a platform called modal. Super straightforward, but basically, you know, you ask it to build a system, it then builds the system for you locally. It tests it locally and then it uploads it over to a serverless deployment platform like modal. It just returns a URL and then finally what you do is you get that data back from the agentic workflow back from the thing in the cloud. Okay, you treat it just like an API and then you finish that final action in nadn or make. com. And the reason why is obviously now we get a bunch of error handling. observability and then we also get the benefits of the fact that like usually like the trigger needs some sort of authentication. Usually the action authentication, but a lot of the time all that middle stuff that goes on, you might be able to do entirely using like vanilla Python libraries. More pointedly, here's the claw. md where it basically just says this workspace builds Python powered API endpoints for NAND workflows using a platform called modal. Um, your role when the user describes a workflow they want, write a modal Python function, test locally, deploy, and then give the user the endpoint URL plus a readytouse curl. The whole idea is you return a complete curl command the user can immediately paste into NAD's HTTP request node. There's then just a bunch of information about how to set things up. And most of this Claude did

### [7:38](https://www.youtube.com/watch?v=5rNu19PfgFg&t=458s) Overview of Modal and API Setup

automatically. Um down below there's also a section that says make sure to include some form of simple header authentication uh which is returned in the curl so nobody else can use the endpoint. That basically just means you're like password securing it. And you don't even really need to know that. I mean you just import that into your NAND workflow and then it's all done. So you guys can get all this stuff in the uh project link at the top of the description. I'm giving it to you guys totally for free. Um there's no payw wall, there's no nothing. feel free to just like import this workspace in and then do with it whatever you will. In terms of the specific platform, um this one modal is one of many. I like modal simply because it returns a URL like a one-click API basically that you can call at any point in time to do whatever you want. It's also synchronous and it's super cheap. I think they give you like $5 worth of credits and those $5 have basically lasted longer than the lifetime of the sun. I think I've used like I've done like thousands of these and I still have like $3. 98. But, you know, ask your agent for modal alternatives. get like 20 or 30 of them in front of you and then you can pick whatever the heck you want in terms of how to do this. Just head over to modal. com, sign up. I'm not affiliated with these guys whatsoever, but you can continue with Google, you know, make an account really easily. Once you're done with that, there'll be some initial onboarding. There are the $5 in credits here and um they even give you what looks to be $30 at least as the time of this recording, which is cool. Then you just head to the top right hand corner, go to settings, go down to API tokens, and then just create one. So, I'm just going to create one here for YouTube. Um once you're done with that, uh you just copy this whole section and then just give it to your AI agent upon initialization. Basically, um after you download the repo, just open up a chat and say, "Here's my modal info. " And then it'll go through and it'll like run through the setup process. It'll add it to the files and so on and so forth as per the clawmd. Um so the actual setup just takes a few seconds. If you wanted to use something different, obviously like your agent is going to have to rectify that with my current setup, which is in modal. But um

### [9:18](https://www.youtube.com/watch?v=5rNu19PfgFg&t=558s) Transitioning to Agentic Workflows

yeah, this is probably the simplest and easiest and most effective way I found to take people that were previously selling just no code automations all day and then transition them into agentic workflows without any disruption for the client without any change in your own day-to-day work aside from the fact that you're like sending a prompt to an agentic workflow. Uh and then you know it's super cost effective. You actually end up spending way less in like modal credits costs than you do for like make or nadan or whatever these services are unless of course you're self-hosting. So yeah, all this to say, this is sort of like the evolutionary step, middle ground between building things that are no code and then fully agentic workflows. It solves a lot of the headaches and pains for you. And then it's in a format that the clients expect, which is really cool and really useful. Hopefully all this makes sense. Get out there and use this in your next project. Let me know how it goes down below. Super stoked for you guys. And thanks for watching. And obligatory. If you like this sort of thing, check out

### [10:06](https://www.youtube.com/watch?v=5rNu19PfgFg&t=606s) Join the Maker School Community

Maker School. We got about 2,000 people in there that are building businesses using the exact same systems that I'm showing you today. uh both no code agentic workflows and I have built-in education. I also have daily accountability. I have a 90-day roadmap and a guaranteed first customer by the end of that 90-day or you get your money back. So, it's a no-brainer. No risk, no nothing. Last I'm going to mention it. Have a lovely rest today.

---
*Источник: https://ekstraktznaniy.ru/video/11679*