n8n + Claude + Apify: Automate Contracts, Resumes & Job Search with AI | Real Use Cases 2025

本站添加: 2026-06-01

[00:01:58]So, hello everybody. Good evening. We're going to wait a few minutes for everybody to join. We will start in in in a few minutes, okay?

[00:02:09]In the meantime, please share where you're joining us from in the chat.

[00:04:14]>> Cool.

[00:04:15]So, I think it's it's time to to start.

[00:04:18]It's already going to be 6:00 p.m.

[00:04:21]Um so, let's let's at least intro- introduce this session.

[00:04:26]Um so, welcome everybody. Um I really appreciate you joining this this evening.

[00:04:33]Um I see that you're you're joining from different parts of Europe. Um I'm I'm in Spain, okay? So, I'm located in Spain.

[00:04:41]Um and why are we doing these sessions?

[00:04:45]So, we see we're we're meeting every week with several companies uh because we help them automate their processes, okay?

[00:04:52]And we see a lot of demand uh for, you know, automation um automation and AI agent knowledge.

[00:05:02]We're also seeing how companies are demanding more and more uh decent level of AI literacy, but also an ability to understand systems and processes, okay? So, we decided to host these workshops um in order to, you know, to show different cases that we're working with companies, but also to show cases that might be useful for people that might be looking for a new job or a career change, okay?

[00:05:28]So, we really appreciate uh everybody joining today. Um at the So, the the dynamic is going to be simple. At the beginning, we're going to um explain the agenda, okay? And then we will see practical cases, and you can uh put your questions in the chat. And, uh, uh I will explain, there are several ways in which we can stay, uh, in touch, okay? Uh, before starting, would appreciate if you subscribe to the channel, but it's totally optional. And let's start, uh, with a few ground rules, okay? Um, so first of all, um, super interesting to know where you're coming from. So please share in the chat where are you located? Um, I see people from Edinburgh, Australia, Bristol, uh, Poland, so nice. Brussels, um, UK, uh, Oxford, Turkey, UK, nice, okay. I'm in Spain, okay, so I'm in Canary Islands.

[00:06:32]Um, and so it's it's not sunny as people might think, but it the weather is decent, okay? So hopefully this weekend it will be sunny enough.

[00:06:41]Um, second, I kindly ask everybody to be nice and respectful to each other. Um, I also encourage people to ask any questions that they might have through the channel. I am super happy to answer them as soon as, you know, I'm able to to make a stop, I'll try to answer them, okay?

[00:06:59]Uh, also just wanted to to provide to give a heads-up. Uh, sometimes Claude or N10 or even myself can make mistakes, so I please ask you to be patient and to be flexible, okay?

[00:07:12]We will do our best to deliver a quality presentation.

[00:07:15]Um, also after the session, uh, the video will be available, only the videos, so the chat, uh, history won't be available for people to check it afterwards, okay?

[00:07:27]So, uh, if anybody wants to wants to see the video afterwards, you can just subscribe the channel and the video is, uh, saved in the, uh, reproduction list, in the playlist of meetup events, Okay?

[00:07:43]And finally, before starting, I would kindly ask you to like and subscribe to the channel, but again, totally optional. Do whatever you please. Okay?

[00:07:53]So, let's see the agenda for today.

[00:07:55]Um so, first we're going to explain a little bit the automation and AI agent space, what we're seeing in the automation and AI agent space, what are companies looking for. Um then we're going to explain a little bit why companies are choosing N A 10. Okay? Uh what are the reasons behind companies going with N A 10, even though there are several platforms that are worthwhile mentioning. Okay?

[00:08:19]Then we're going to discuss how companies are using it, but that's going to be very quick because the idea is to actually see real examples so that we can uh comment on them. Okay?

[00:08:30]Um finally, I want to see also a little bit about the cloud ecosystem, okay, because they're going to cloud and N A 10 are going to live together. So, it's not a matter of what to choose. We're going to see how they're going to live together. And companies are, you know, making sure that they combine the correct one uh when needed.

[00:08:49]Okay?

[00:08:50]Before starting the session, I just wanted to mention that um the session is hosted by by Rest Flow, which is my company. We do uh automation as a service for companies, uh especially in the US, UK, but also but in the European Union also.

[00:09:07]And we focus on deployment on of automations, but also we do um corporate trainings and and boot camps, okay, for individuals. Okay?

[00:09:17]Cool. So, let's start. Um First of all, what is the automation and AI agent space looking like right now?

[00:09:25]Okay?

[00:09:26]So, in general, we're seeing how companies are wanting to implement AI agents, um even though what we're feeling is that sometimes what they need might not be an agent. Many times what a company actually needs it's an automation, but I think we're very much focused on buzzwords, we're very much focused on keywords, but in reality most of the things that companies want to do are automations, okay? It's true that some agents are are worth the investment, but I think agents still have a little bit to to go um a little bit of of let's say project you we need some project management approach for for agents, and you need a lot of monitoring after implementation, okay?

[00:10:13]When we're talking about workflow automations, the main tools that we're seeing uh succeed are Make and n8n, even though there are more. So, you have Microsoft Power Automate, you have also um Zapier automations.

[00:10:27]Uh you might have UiPath, which is quite interesting also in the EU.

[00:10:32]Um but specially I would say that the the the one that is taking the lead is n8n, okay? And we will see and we will explain why.

[00:10:42]When are companies implementing these automations? When are companies implementing n8n? When it it's related to a repetitive uh sequenced and deterministic workflow, okay? So, if it's a process that is repetitive, if it's a process that is sequenced, meaning step one, step two, step three, and step four cannot each one cannot execute before the previous one has finished, um and the being them deterministic, meaning for the similar types of inputs you would expect similar types of outputs, so n8n has been the go-to choice for these automations, okay?

[00:11:19]Um Make is doing really well, it's actually picking up in my opinion. It's implementing uh tools that will allow it to compete with n8n, okay?

[00:11:30]Uh but again, I think n8n is taking the lead. When it comes to AI agents, companies are implementing them, especially uh, for, you know, document management, chatbots, um, let's say, agents that would handle reconciliations.

[00:11:46]And usually, they're focusing on implementing these agents when it's a matter of non-repetitive, non-sequence, and non-deterministic tasks, which means that the agent will be able to make a decision, and therefore, can actually make mistakes.

[00:12:00]So, therefore, when you're implementing these agents, you really need to focus on monitoring and telemetry, and there needs to be a real project afterwards to make sure that it's delivering the right results. Okay?

[00:12:13]Um, in terms of So, what what what is our typical discussion, right? So, companies are saying, "Hey, what should I choose, Anybrain or Cloud?

[00:12:22]Or which which model is the best model?"

[00:12:25]And we always say the same. So, depending on the task, you're going to be using either Anybrain or you're going to be using Cloud. It depends on many variables. It depends on how many times are you going to run the task. If the task is going to run 1,000 times per day, you shouldn't be using Cloud because you will hit limits very quickly. And we're seeing all the issues that, you know, in the market, where companies are announcing that they're stopping uh, usage of Cloud. That's I think we've gone to an extreme of of using Cloud for everything.

[00:12:56]That doesn't make sense from a practical perspective. If I want a simple answer, I won't go to Cloud. If I want to extract information from a document, I can use OpenAI very easily. Okay? So, for some reason, it's there's been like a FOMO effect on using Cloud for absolutely everything. That doesn't make sense. Okay? So, you have to choose the right battles in which you need to use Cloud, and you cannot use Cloud for everything. So, I I always say, um, have you ever written to Claude uh, something that just needs a simple answer and instead you will get you get a whole report, right? So, I think Claude can can be an overkill sometimes.

[00:13:38]So, imagine we want to automate the extraction of invoices.

[00:13:42]Do you really need Claude for that?

[00:13:44]Doesn't a simple Open AI model work well? Okay. So, this is a typical debate that we have with companies to make sure that they're efficient and that they're effective when implementing the different technologies. Okay?

[00:13:57]So, um why are companies choosing that doesn't mean that Claude is not good. Claude is great, incredible and we're going to see it today also. It's just you need to choose the right bottle the right bottles for each tool. Okay?

[00:14:11]Um cool. So, why are companies choosing N8N when it comes to automation? Okay?

[00:14:17]Um so, first of all Well, first of all, I I I'm not getting paid by N8N. Okay? So, I'm just explaining what we're seeing in the market and you know, why we have companies with both tools.

[00:14:31]Um so, one of the reasons that companies are choosing N8N is due to privacy.

[00:14:35]Okay? So, the ability of N8N to be hosted locally or privately will help you in having a lot of flexibility in data protection.

[00:14:44]Um also due to its flexibility when integrating with different tools and ecosystems. You can integrate it with your Microsoft, with your Google, with you know, whatever tool you please. You have a lot of flexibility in terms of integrations.

[00:14:58]Um Zapier also has a lot of pre-built connections, but I would say N8N is it's it's just more flexible. Um maybe less pre-built connections, but more ability to create integrations.

[00:15:10]Um also due to its integration with AI.

[00:15:13]So, you're able to do a lot of things with AI. You can use different nodes.

[00:15:17]You can use it with Ollama, with Claude, with Gemini, with Open AI. Okay, you can connect it with Azure. So, you have a lot of flexibility on integrating it with AI. Okay? Also, due to its ability to become as customized as complex as you wish or as simple as you might be needing when you're starting with. Okay?

[00:15:38]And finally, due to the stability of the cost. Okay? When you're hosting anything locally uh, or privately or you're setting it up on a server of the company, the price doesn't scale, which is a typical issue on why companies are leaving all these SaaS companies, right? So, one of the main, uh, complaints that we see when companies are leaving all these software as a service companies like Salesforce, HubSpot, whatever, is usually the the escalation of pricing. Uh, besides may- maybe other similar other smaller factors, uh, but one of them is just that the the price scales too too fast and too quickly. Okay?

[00:16:16]Cool. So, let's see a few examples. I'm not going to stop in this one. I think we're going to be able to see them uh, realistically. So, the first workflow I want to see, okay? So, this is the N8N interface, is this automation interface.

[00:16:31]And we're going to start with a document reviewer. Okay?

[00:16:35]This document reviewer, this sorry, this document reviewer is a workflow that's going to basically get contracts from a folder and it's going to extract the relevant information from that contract and transpose that contract into an Excel file.

[00:16:50]The same way it's connected to an Excel file, it could do it with a CRM, with a database. So, you would be able to get a contract, extract the relevant details of the contract, and transpose this information into a Google Sheet file. Okay?

[00:17:03]>> [snorts] >> In this case, this contract, the contracts that we're going to work with in this in this example, these are lease agreements. Okay? So, these are contracts where a tenant and a landlord reach an a lease agreement. And we're going to have several of these contracts, okay?

[00:17:21]Once we place the contracts into this folder, what's going to happen is that NA10 is going to start running, okay? So, what I'm just going to publish it so it starts running, okay? So, now it's enabled and it's going to run every 30 seconds. We see it here, okay?

[00:17:38]So, what's going to happen is that if I go here to executions, we see it hasn't run for a while and now what I'm going to do is I'm going to drop a few contracts in this folder, contracts to process, okay?

[00:17:51]So, here we've dropped these contracts and what's going to happen is that hopefully, in less than 30 seconds, we will see how it starts running again, okay?

[00:18:01]And when it starts running again, it will fetch start fetching these contracts and start transposing the details of the contracts in this Excel file.

[00:18:12]Once it has moved, you know, transposed these uh these details, it will uh reallocate the so, it's running now and we're going to see how um the data is being transposed to the Google Sheet. And once it's done, it's going to move the contract into the uh processed uh contracts folder, okay?

[00:18:35]So, let's see how it runs and now we're going to double-check these these uh details to see if it if it's if they're coherent with the contracts, okay?

[00:18:49]So, let's go to process contracts. We see how now we have the contracts here.

[00:18:55]And so, we're going to open this one, San Fran- Sacramento H Street and we're going to double-check, okay?

[00:19:02]So, here we have Sacramento H Street, so the contract date is the 10th of July of 2024.

[00:19:11]10th of July of 2024.

[00:19:14]Okay?

[00:19:15]The landlord is Thomas Bradley. The tenant is Erica Dawson.

[00:19:20]Landlord is Thomas Bradley. Tenant is Erica Dawson. Okay?

[00:19:25]Uh property address H Street.

[00:19:29]Property address H Street. Okay? Lease agreement will start 1st of September 2024.

[00:19:37]So, if I go here, lease agreement starts at September 2024 and finishes 31st of August of 2025. Okay? So, for now it makes sense.

[00:19:50]Imagine I would change the contract of the format.

[00:19:53]This happens, no?

[00:19:54]But if it's the same contract, the AI will understand it. And here we have another example, which is Olivia Bennett.

[00:20:02]Olivia Bennett with Daniel Moreno. So, here we have Olivia Bennett with Daniel Moreno, the 15th of January of 2025.

[00:20:12]Okay?

[00:20:13]And the the the property is in 2338 Maple Grove.

[00:20:20]Okay?

[00:20:22]Let's see the monthly rent 3150.

[00:20:25]So, rental payments, yeah, monthly rent 3150. Okay, makes sense.

[00:20:31]So, let's see step by step how this workflow worked. Okay? So, I'm going to go back to the editor. I'm going to unpublish the workflow.

[00:20:41]And let's We're going to run it manually. So, I'm going to drop these two files in this folder, contracts to process.

[00:20:48]Wait 1 second because I think it was running. Let's see if it stops.

[00:20:52]It stopped already. Okay, so it's stopped.

[00:20:55]So, what I'm going to do is I'm going to drop here these two files and let's do this manually, step by step. Okay?

[00:21:05]Cool.

[00:21:06]So, the first step is going to be uh the schedule trigger, okay? So, this will allow the workflow to initiate every 30 seconds, but in this case we have let's say stop the auto running to running manually. This node basically what's going to do is it's going to be connected to my Google Drive and it's going to search in the folder contracts to process.

[00:21:31]So, it's going to get this contract, okay? It's going to show the two contracts that this locate that are located in the folder.

[00:21:39]San Jose Alameda Avenue and San Francisco Mission Street.

[00:21:44]So, what's San Jose Alameda Avenue, San Francisco Mission Street, okay?

[00:21:50]Then it's going to loop over items, so it's going to start processing one by one. So, first it's going to fetch the first one San Jose Alameda Avenue and then it's going to download the document, okay?

[00:22:03]After downloading the document, it's going to extract the information from the document. So, here we have the content of the of the lease agreement.

[00:22:14]And then it's going to send all the information to an Open AI model, which is a GPT 4.1 Nano from Open AI, which is a very fast and efficient and cost-friendly model with a prompt that will explain how we need to extract the information from this contract.

[00:22:32]Then with this data of the contract, okay? So, Open AI is going to respond with the information of the contract.

[00:22:40]This is the response from Open AI.

[00:22:43]Then we're going to use a node that will help me transpose all the details to an Excel file, okay? To this Google Sheet. So, if I run it, what's going to happen is that we're going to see a new line appear. Yeah, here we have it, okay?

[00:22:58]And finally, once the the new line has appeared in the Google Sheet.

[00:23:03]We're going to move the the contract to the folder process contracts instead of contracts to process.

[00:23:11]So, what's going to happen is that we're going to run it.

[00:23:14]Here, we're going to see how the contract disappears.

[00:23:17]And now, this same contract is here, okay?

[00:23:22]So, this is a very simple example on using an A10, okay? So, here we're we're we're seeing how it's used. You can use it to extract invoices. You can use it to extract contracts, okay?

[00:23:36]Um so, somebody's asking, "What do you use for OCR that OCR that piece of to produce a JSON file?" No, we're not using OCR technology. We're using AI.

[00:23:48]So, we're sending to Open and Open's AI API all the information about the contract. And Open AI Open AI's API is extracting the details for this. What does this allow me? That if the contract has misspellings or issues or is handwritten, we won't have a risk of error. So, it reduces you know, we're not we're not we're not using OCR technology.

[00:24:15]Um here, what's happening, this node is not using OCR technology. It's extracting from a PDF file, which is text, okay? If you were to have an image like like a JPG, okay? You could use another node, which is the Open AI send a file node, okay?

[00:24:33]You can use upload a file, and what happen is that if you have an image, for instance, imagine I have a ticket, right? A a receipt a receipt. And I want to make sure that this is also automated to my to do my taxes.

[00:24:46]So, I could use this upload a file.

[00:24:50]Okay? Um upload a file to send the the picture to Open AI, and it will get the information and return me the details of that, okay?

[00:24:59]So, OpenAI is very flexible in sending it, you know, unstructured documents and extracting information from them.

[00:25:07]Somebody's asking, um, do you use the PDF metadata? You could use it, but here we don't use it necessarily, okay? So, here I don't use it for anything. I just use the content, okay?

[00:25:26]Um, so somebody's asking, I assume you used an OpenAI API key for this loop. Do you have tips on how to combine anything and AI desktops like Cloud Desktop?

[00:25:38]So, it depends, right? So, here, um, you're you cannot So, with anything you cannot use your chat subscription, your $20, $18 chat subscription. You're using either Anthropic's API or OpenAI's API, which is the tool that integrates with anything, okay? So, I'm not connected to my typical ChatGPT platform. I'm connected to the OpenAI API platform, and I can control very well the usage that I'm doing of of this platform, okay?

[00:26:09]So, um, and I use it a lot, and if you choose the right model, what's important here is to choose the right model. We We I I still see how a lot of companies want to use the shiniest and the brightest model, and this task is very simple to use the shiniest and the brightest model. You want to be efficient. You want to be effective. You want to be cost-friendly. If we were doing a market research, I might be using the shiniest and the brightest model, but if not, I might use this one. And this one, these models, we you need to pay per use. But if you use the right one, it can be very efficient.

[00:26:42]For instance, I could run this 100 times a day and not spend more than 10 cents of a dollar, okay?

[00:26:49]Uh, Uh, somebody's asking, what about security since it's contract information? What about confidential contract?" This is a These are great questions, guys.

[00:26:57]Uh, so um, in general, um, uh, you could use so OpenAI is secure, okay?

[00:27:06]So, if if it's located in if you're in the UK and you're using OpenAI with UK servers, that's going to be secure.

[00:27:13]But, even if you are very very careful with the data, you could use local models to do these tasks. Okay? I would say that maybe local models are still getting better over time. So, now we have recently Gemma 4, which is great, okay? Uh, but these are you know, it require they're very power hungry. They will they won't run very well in normal computers. So, you need you kind of need to be careful, but in a for a company, this could easily work, okay?

[00:27:42]Um, cool. So, let's uh, let's let's see another example.

[00:27:51]So, somebody's saying you can use your subscription if you root the AI operations to a server running Codex with a subscription enabled.

[00:27:59]That could be an option, but I I I I don't know if um, so for for for us, we run this, we won't be even spending $2 per month, okay? And so, I'm not that worried for specific task about getting a subscription when sometimes, you know, if you choose the right model, especially with data extraction. With data extraction, you can choose very efficient models and run them a lot and not be too much worried about the price, okay?

[00:28:29]Um, cool. So, let's let's continue with another example that I want to see.

[00:28:35]Um, another another way we're seeing that might be interesting, okay?

[00:28:42]is for data extraction from product mar- marketplaces, okay?

[00:28:47]So, in this case, what happens is we had Sorry, I'm going to open the website, okay?

[00:28:54]So, we have here a company that is an e-commerce.

[00:28:58]And so, maybe I'm going to go here to the UK website. And maybe you have similar products and you want to improve the way your your, you know, your product descriptions are are stated, okay? Or the way you handle product descriptions.

[00:29:13]So, this workflow, this automation, basically what's going to do, it's going to go to the website, it's going to start It's going to understand what type of behavior it can do as a robot, okay? And it's going to start extracting information from the products, okay?

[00:29:30]So, if we run it, what's going to happen, I have limited because this website has more than 6,000 products, but I limited to 20, okay? And once I have run it, let's see what's going to happen. So, if I go here, what's going to happen is I'm going to start getting all the information of the product, okay?

[00:29:49]We're going to have the features, we're going to have the name, the price, the shipping policy, okay?

[00:29:54]Um let's see how this runs. Now, the second product, okay? So, I have I have limited to 20 items, uh because if not, we could be here all day long, but I'm going to explain to you how this one works, okay?

[00:30:07]So, let's allow it to run and we can see how it's extracting information from these products. Okay, this is useful to use with, you know, other other marketplaces or any other types of products. I'm going to stop it right here, and let's see how this works.

[00:30:23]So, first of all, I need to do an evaluation of what the site allows me to do as a robot, okay?

[00:30:30]So, a way to do this is you can go to the website, you can go and say and you can write here robots.txt.

[00:30:38]This will give me information on how uh the the website reacts to a automa- automations, okay? On how it works with uh with robot behavior, okay? If then, so I will see a little bit about what I can Apparently, here they're not blocking seeing their products. If then I go to sitemap XML, I'm going to get a list of the available sites of this website.

[00:31:07]Ideally, we want to choose the one that is giving the information about the products. If I go here, what's going to happen is I'm able to see all the sites, all the products sites, okay?

[00:31:21]So there's like a It's like an index of all the pages of the products that they offer. So here, if I go here, I will get a product.

[00:31:32]Okay?

[00:31:35]Cool. So basically, what we're doing here with this HTTP request is replicating that behavior against this latest website that I have entered, okay? So basically, what this is doing is exactly the same that I've done here, but automatically, okay? If I go here and I click on show data, you're going to see how well it's less structured, but you're going to see this XML structure, okay?

[00:32:01]Then, what I'm doing is I'm cleansing a little bit, I'm cleaning a little bit that structure, so that we get an easy-to-read item, okay? Which is the URL and the product ID, okay?

[00:32:15]After that, here I put a limit because I don't want to run the 6,000 for this demo, but you could easily remove this and have it run 6,000 times, okay?

[00:32:25]Then, what's what I'm doing is I'm saying, "Okay, so now I'm going to So now I'm just going to run by one by one, okay?"

[00:32:34]And so I'm going to start with the first URL.

[00:32:37]Now, what it's going to do is going to pass the URL as a variable and it's going to enter this website. So, it's basically what this is doing is basically entering here and fetching all this data, okay?

[00:32:54]All this information.

[00:32:56]Then, I'm going to use an open AI model, again a 4.1 nano, super efficient, to extract the relevant details of the product.

[00:33:08]Okay?

[00:33:09]And finally, I'm just going to transpose these details, the title of the product, description of the product, price, and shipping policy into my Excel file, okay?

[00:33:22]Here we have it.

[00:33:24]This one, which seems a very simple workflow, is super useful for anybody that wants to um improve their product pages, that want to have an idea about products, okay? And you can make it even more complex. You can connect it to Apify and connect it to Amazon products and extract the information from Amazon, okay? So, that's also another option, okay?

[00:33:45]Cool.

[00:33:46]So, let's see uh another example that we're we're we're using. So, basically, I collaborate with several talent networks and what we see is that there's a little bit of a turmoil in the job market and a lot of people are either thinking on a career change or they're thinking on switching jobs or they're think they're thinking to make a move, basically. And so, the demand for, you know, job-related workflows is is quite high, okay? And what I did is basically um so, we have here Let's see if this one should work. One second.

[00:34:23]Um I think this one.

[00:34:27]Yes.

[00:34:28]So, here basically what I'm doing this is this a our Again, I want to explain a little bit of this workflow. This is a workflow that's connecting to several job sources.

[00:34:36]One is Himalayas, one is Remotive, one is We Work Remotely, one is Remote OK, one is Jobicy, one is Job Espresso.

[00:34:45]This one is uh Remote Jobs, and this one is Active Jobs, okay?

[00:34:51]What this workflow is basically doing is every time it runs, okay? It's going to go to the website, it's going to scrap all the details of the jobs, then it's going to filter the jobs that are more recent, and finally, it's going to transpose the details of the job in this job tracker, okay? So, what's going to happen ideally is that I'm going to see hopefully two jobs. Okay, here we have it, Himalayas with the date of today, okay?

[00:35:23]Um cool.

[00:35:27]Uh so, here, if we run this one, it's doing a similar thing, the same similar structure, but with Remotive. So, maybe there's no data. Let's see if there's data.

[00:35:38]Apparently, there's data, five jobs, and OpenAI will extract the information of these jobs, and it's going to allow me to structure this information in a way that I can use it to apply, okay? In this case, for product-related roles.

[00:35:54]So, the structure of these nodes are very similar, right? So, first, I'm going to the website, I'm extracting the information of the jobs, then I'm cleaning the data and filtering by date, I'm filtering by type of jobs, I'm using keywords. Here, I'm putting a filter related to let's not get dates before this date, okay? In this case, I think I'm just filtering with uh less than 1 day, so in the last 2 days, okay? And then, I'm using OpenAI to clean that data and extract the relevant details that I actually care about, which is the link of the job, the the title, the company, etc. Okay?

[00:36:32]So, this workflow is very useful to gather information from different sources for jobs, filter by keyword, and then consolidate all these jobs in my Excel, so that I can start focusing and filtering the one remote, do I want this company, and this will make my job search a little bit more efficient.

[00:36:52]Okay?

[00:36:54]The next workflow that I want to see, okay, is um this one, which will help me generate a resume for my job, okay? For these jobs.

[00:37:06]So, basically, this workflow uh it's done in a way that Sorry, I'm going to show you quickly how is this connected. So, basically, here we go to job applications.

[00:37:19]Here we have any.10.

[00:37:22]Okay, so what's going to happen is that any.10 it's connected to this job tracker and to this job description automation, okay? Job description sheet, sorry. And so, for instance, if I do this one, which is Workato, which is that chief client architect, and I just copy the job description, and I go here and I just open the chat and put the job description, what's going to happen is that it's going to get the job description, it's going to get information from the companies that I have worked, that I have in my Google Docs, my experience, my education, my achievements, and then it's going to create a resume in the Harvard format, okay?

[00:38:08]And then it's going to place that resume in this folder. So, it's going to create a new folder with the name of the company that I'm applying.

[00:38:17]I have so many things open that I'm going crazy. Workato, okay? So, what's going to happen is that we're going to see how it creates a new uh folder, ideally with the name Wakaito.

[00:38:27]Depends if it if the name appears in the job description or not, okay?

[00:38:32]And uh let's see how this is doing. So, yeah, now it's going to create it.

[00:38:39]Okay, Wakaito, nice. And what's going to do, it's going to put the job description that I have pasted, so that I have it when the come times to prepare the interview, because sometimes I were I was creating the resume and I was forgetting about the job description and I couldn't couldn't find it.

[00:38:56]I mean, sometimes they spend so so long to call you back, okay? And so, here you're going to have the job description that you passed.

[00:39:03]And here you're going to have a resume that is adapted with the Harvard format to apply to this job, okay?

[00:39:11]This is going to be super useful to adapt uh each resume to the keywords of the job.

[00:39:19]Okay?

[00:39:20]Another [snorts] way of doing this, instead of putting it on a chat, okay, is actually joining the work the work the whole workflow, okay? So, here for instance, I have a a little bit more of an advanced version of this workflow, where basically I'm just fetching the information from the Google Sheet directly. I'm not using the chat. I'm just running it from this Google Sheet.

[00:39:44]So, it will go to this Google Sheet and it will fetch any any job any job description whose process column is set to no, okay? So, if I put here yes, okay, and these are no, what's going to happen is we're going to start creating the resumes from here onwards, okay? So, if I run it, let's see if this is going to fetch it correctly. Ideally, we're going to see here Adobe, nice. So, let's run it.

[00:40:13]And this is going to do exact the exact same process without me having to put it on the chat.

[00:40:19]And what's going to happen is that here, ideally, we're going to see how now we have a an Adobe folder, and we're going to see a CV, a resume created with this format, okay? With the with this format, I mean with this adapted to this to this job description, okay? Let's wait a few seconds for it to run.

[00:40:50]Uh oops, taking too long.

[00:40:57]Uh now, we'll see it now, hopefully.

[00:41:01]Adobe, nice.

[00:41:03]Now it's filling in the details, okay?

[00:41:05]So, it's going to fill in the details.

[00:41:07]Just 1 second, okay?

[00:41:09]And hopefully, if I go in, I will see my resume adapted to this job description, okay?

[00:41:19]So, the resume is being generated with this model and this this file, which adapts the resume to an HTML that is rendered to a Google Docs, okay?

[00:41:31]How much can it cost me to run this workflow? I would say that every time I run this workflow, uh since I'm using a mini model, which is a little bit more uh it's a little bit more costly than a nano model, but less than a pro model or a normal model.

[00:41:46]I would say that maybe each run it's like 3 cents of a dollar, okay?

[00:41:50]Um but you know, I I I you know, I've been 1 day sending resumes the whole day. Uh last year, this year I'm focused on my company, but last year, I could run it and run it, and I wouldn't spend more than $1. I mean, you got you have to run it to a lot of times to start consuming 2, 3, 4, 5 dollars, okay? I say that's in dollars because OpenAI uh OpenAI I'm using OpenAI and OpenAI is in dollars. Okay? Even if I'm in Europe, I just see dollars.

[00:42:21]Okay?

[00:42:25]Um this is not of only this. Okay? I've been running automations for the whole day. I've been doing a lot of training, so this is not only for this execution.

[00:42:34]Okay? Actually, since I've been using mini maybe I can see because mini this is the model that I have used for the resume and we've used 4 cents in the whole day.

[00:42:50]And I have used mini for several times.

[00:42:52]Okay?

[00:42:53]So, I think this is this is this is feasible. This is something that anybody can can can do. Okay?

[00:43:00]Um cool.

[00:43:03]So, we've seen a little bit of about N10.

[00:43:07]I would like to see a little bit about Cloud Co-work.

[00:43:10]Cloud Co-work is getting better and better and I'm actually being using it.

[00:43:14]I'm super happy to actually use it. I've been doing stuff. Okay?

[00:43:19]And so, Cloud Co-work now, you know, it's so I was trying to run uh Cloud Co-work to for instance I'm so 1 second.

[00:43:33]I thought there was a question. Okay.

[00:43:36]So, um something that I'm liking about Cloud Co-work is it's allowing me to go to my folders and folders that I might have uh very messy. Okay? Folders that I really have out of control and I'm being able, you know, to see the different invoices that I have. Also, see a table of the invoices and organize my invoices. Okay?

[00:44:00]Another way that I'm using Cloud Co-work this is taking a lot, but Um, let's see if I have another example here.

[00:44:10]So, So, another way which I'm using Cloud Co-work is to go through my emails to get all the subscriptions that I have uh registered for and basically provide me with the details of the subscription, like any newsletter or any, you know, any email that's probably uh advertising or random stuff that might not be super important for me.

[00:44:38]So, Cloud Co-work can go into your email account, into your Gmail account, fetch those details, okay?

[00:44:45]Uh so, I'm going to stop this because it's been running. Sometimes it's too slow.

[00:44:49]I'm going to just start again.

[00:44:52]All the G Let's see if I can use Hypo for this because only this for some reason today Sunny hasn't been working with for me. Um, hopefully Hypo will work. Here basically we're using Cloud Co-work with a connector to Gmail.

[00:45:57]And for me it has been super useful to fetch I did it with my Hotmail account with my with my Hotmail account. And what I did is I got all the email subscriptions that I have been accumulating for years, okay?

[00:46:09]And I got a table and I started unsubscribing with all this garbage that I have been accumulating. For me, this is a this is a very interesting use.

[00:46:18]Um and with Gmail it's even easier, okay? So, the connector with Gmail works really well.

[00:46:22]And so, what's happening is um um it's going to go to my Gmail account, filter, and then put me a list, ideally, if it doesn't take too much, uh with this uh newsletter, you know, the the sender, the topic, and a link to unsubscribe, okay?

[00:46:42]Let's see if this gets done quickly.

[00:47:04]Um It's it's taking way longer than what this was taking yesterday, to be honest.

[00:47:14]I don't know what's happening today with Cloud and Tropic status page.

[00:47:22]So, I think Cloud has been failing today a little bit.

[00:47:26]There's some outages, okay?

[00:47:29]But let's see if this works.

[00:47:34]>> [snorts] >> So, somebody is asking if I place myself on a market as an AI automated agency, where do I get information on how much to charge for a specific automated service?

[00:48:04]I mean, that's a great question. Um I would say it depends, right? It depends on the type of client. For instance, bigger companies might be willing to work with retainers and with uh higher fees than in Europe, for instance, $100 per hour, no, €100 per hour. A bigger company will require you to have a uh infrastructure, okay? So you cannot go as a freelance. Um but then, once you have their trust, you might be able Oh, so here we have it.

[00:48:34]Um So yeah, here we have a list of subscriptions of new let- newsletters.

[00:48:43]Uh it's taking too much, in my opinion.

[00:48:45]Usually it's faster, but I think there's some outage with Anthropic API. And here I just fetches I just fetch the last 30 days. This is super useful. Imagine getting all your subscriptions or your unused newsletters and studying unsubscribe to all of them. Super useful, super effective, and great a great way to reduce noise, okay? Um so back to the question that somebody was making, um so it depends a lot of things. I would say you need to understand uh how much a freelance uh you know, a simple freelance is making because it's going to be competing. Most of the companies are are be you know, are usually debating internally whether if they prefer an agency or a freelance.

[00:49:26]You also need to see the market, okay?

[00:49:28]It's not the same Spain than Germany. In Spain, anything above than €70 per hour is already too much, whereas in Germany, you could easily reach to €90 per hour.

[00:49:38]Um I think a medium-size company is different from a big company. A big company is going to require a lot from you, but if you deliver, you're going to get a lot of stability, so that's going to be great because you're going to be able to put some uh let's say like a like a package of hours every month and that will give you a lot of stability. Whereas a medium company will require you less.

[00:50:03]They won't be on top of you, okay? They won't be just pushing you so hard.

[00:50:07]But you're going to be having to identify new opportunities with those type of types of companies, okay? And when it comes [snorts] to medium companies and small companies, ideally you want to make sure that you are creating something that's scalable. You cannot do automations a la carte. Uh you know, you cannot do body shopping automations.

[00:50:28]At least when you're starting because that's going to make it really hard for you to scale.

[00:50:32]Okay?

[00:50:33]Um I hope that I have answered your question. It's complex. We're all finding our way, even myself.

[00:50:39]So if you know, if you want to have a chat, I'm super happy to you know, to share information, of course without uh names of companies, but I'm super happy to share information for you to make better decisions than what I made at the beginning, okay?

[00:50:55]Um cool. So we've seen one example, which is using Cloud Cover to fetch my these newsletters that might not get a lot of use for. Let's see another example, which is using Apify. So I don't know if you've heard about Apify, but Apify is a lovely tool. Um I don't know has anybody heard about Apify?

[00:51:17]So I'm just going to put it on the chat so that you can uh say if you have heard about it, Apify.

[00:51:24]Has anybody heard about Apify?

[00:51:29]If anybody has, please let let me know through the chat.

[00:51:33]Or if you haven't, please just say it clearly just to know.

[00:51:42]No?

[00:51:43]Um okay, so Apify is a very interesting tool that allows you to get data from LinkedIn, Amazon, and all these anti-robot websites, okay?

[00:51:55]So, what's going to happen is sometimes you're you want to get information from LinkedIn or you want to get information from from um Amazon and all these websites are profoundly anti-robot, which means that you're If you do it with your account, you are putting in jeopardy your account, okay? So, Apify is an um um incredible incredible uh tool, which actually has a free plan, uh that allows you to use $5 per month and you could actually use it, okay? And it's quite useful. And for instance, I have an automation with N10 that scraps information from LinkedIn and I'm going to show you this one, 1 second, this one.

[00:52:37]So, here basically, we can get jobs from LinkedIn with Apify. And this is super useful also to extract information related to um to filters that I have set with my my LinkedIn account, okay? It will use basically the filters that you So, you can set Sorry, I'm going to I'm going to re-explain myself.

[00:53:00]I think I expressed myself unclearly.

[00:53:02]So, you basically can search normally in LinkedIn and decide if you maybe want the jobs in the last 24 hours, the past week, and maybe if you're always looking for the same job description or job position, you maybe put here one product manager, okay? So, what's going to happen then is you're going to fetch the search and you can now run it automatically every week or every day, okay? With this actor that will go to Apify to to LinkedIn, fetch the information, and provide it to you in a structured manner.

[00:53:36]After that, I'm going to be able to get the the job description and the job details in my Excel, okay? But the good thing is that actually Apify works really well with Cloud. So, Cloud connects really well with Apify. So, if I do Hey, connect to Apify.

[00:53:58]Apify and search for the uh 20 latest jobs in London related to uh AI product management in tech industries or fintechs, okay?

[00:54:19]Um and so, what's going to happen is Cloud is going to connect with the Apify connector.

[00:54:27]Um and it's going to start running the actors. It's going to go here So, if I go here to my Apify account, what I will see is if I go to runs, we're going to start seeing how uh it's going to start running. Maybe it already did.

[00:54:47]Let's see.

[00:54:48]So, yes, it hasn't started. I think it will start now.

[00:54:53]Let's see.

[00:54:57]Call actor.

[00:54:59]Now, it's running, right? So, it's getting from in this case from Google. I I would maybe state that I wanted to use LinkedIn.

[00:55:07]Um but it's getting from Google. Uh AI product management jobs, but I can actually do the same and ask it to to use this actor, okay? I can actually go here and say you're going to use this one, okay?

[00:55:23]So, 1 second.

[00:55:26]So, Cloud Co-work connects itself really well with Apify, and it helps a lot in getting information from Amazon, LinkedIn. Uh you can get post rankings.

[00:55:37]You can you know, you can see which posts related to a topic are ranking really well in terms of topic in in terms of likes, comments, reposts, etc. Okay?

[00:55:49]So, let's see.

[00:55:57]Something happened here.

[00:55:59]Yeah, get me a list of the same from LinkedIn with this actor.

[00:56:07]Okay, I'm just going to put this because this was inefficient.

[00:56:11]Let's see if now it does it well. So, now I'm telling the actor that it should use.

[00:56:16]Okay?

[00:56:17]And hopefully apparently cloud is should be working well.

[00:56:29]>> [snorts] >> Sorry.

[00:56:31]Um, and if I go to runs we will see hopefully how it starts running.

[00:56:40]Okay? Now it's running LinkedIn job scraper.

[00:56:47]And let's see.

[00:57:20]It's taking too long also. I don't know why.

[00:57:23]Usually it's faster. So, it usually takes 20 to 12 seconds. I don't know why it's taking so long.

[00:57:30]One result, too little.

[00:57:33]It's London.

[00:57:34]Okay, this.

[00:57:36]Why?

[00:57:44]Something is malfunctioning here. I don't know what's going on.

[00:57:48]>> [snorts] >> For some reason, this this one is is failing.

[00:58:09]Uh when usually, you know, Appify works really well with cloth.

[00:58:14]So, let's see if now we get it done.

[00:58:29]Um let's see.

[00:58:36]So, something's going wrong, I think, because usually it takes way less.

[00:58:49]So, yes, somebody's asking how to find uh okay, apparently it got it well.

[00:58:58]>> [snorts] >> You [snorts] see, cloth, for instance, it says the information is too large. Actually, it isn't, okay? So, now we have the list, but it isn't. I think cloth needs to do a good job in being more efficient.

[00:59:17]You know, when I'm asking for a job, don't tell me all the story, just give me the title, the you know, the details.

[00:59:23]So, I think that's why we always we're we're still recommending companies to choose you know, to choose each task.

[00:59:30]So, to decide per task what do they need to use? Now this a huge over killing many ways.

[00:59:37]It consumes a lot of tokens. It's great, super powerful, but sometimes you don't need that much information. Sometimes you just need something to be efficient.

[00:59:45]Okay?

[00:59:46]Um so somebody's asking where to find and install Apify connector on Flow. Super easy.

[00:59:53]So, you basically need to go to here in Apify, you just need to go to settings and copy the API key, which I'm not going to show.

[01:00:00]And here in Flow, you go have you're going to go to customize, connectors, you're going to go here, you're going to go explore connectors.

[01:00:08]And you're going to do Apify. Okay?

[01:00:12]And here you can basically configure your API key that you have copied from Apify.

[01:00:19]But then what's tricky when using Apify is that you need to choose the right actors for your day-to-day.

[01:00:25]Not only due to effectivity, because there's a lot of actors for LinkedIn and not all are effective.

[01:00:31]But also due to price. Okay?

[01:00:33]So, here we have actors in this case that could consume a lot of a lot of my my available budget and so you need to be very efficient in pricing. To be honest.

[01:00:46]Okay?

[01:00:48]Um kind of a niche question. Have you heard of ICM?

[01:00:54]And where is there any value to the combination of an A10 and ICM?

[01:00:59]I I think I have heard about ICM, but I'm not very sure of what it is. So, I I don't want to talk about something that I don't actually know. And at the end of the day, I'm focused on what the market needs, what our clients need, and what we know that is robust enough to implement. So, we cannot be doing experiments. I love to try, I love to be curious. But when it comes to clients, I need to make sure that what I implement is effective and is stable and is scalable.

[01:01:27]Okay?

[01:01:29]Uh so, we're running out of time. I really appreciate everybody joining today. I don't know if there's any more questions.

[01:01:39]Um so, first of all, really appreciate you joining. Super appreciate it. I hope that you have a great evening.

[01:01:45]Um second of all, we're, you know, moving forward with our Meetup channel and with our YouTube. We're going fast.

[01:01:51]Um I hope that this has this has been helpful. I hope that we can stay in touch.

[01:01:55]The recording of this session uh without the details of the attendees will be available as a as a video in my YouTube channel. Uh I would really appreciate if you like and subscribe. That's good for me, but you know, I I really take an effort on making these sessions practical and market-oriented, okay?

[01:02:13]Um so, somebody's asking, "Can you please share the link of the tool?" Of Apify, you mean?

[01:02:18]So, Apify is this link. I can actually share um So, Apify is this the link of Apify.

[01:02:28]Okay? Um and yeah, we will do more sessions. We will do more sessions. I hope that we can see each other in the next in in the in the next session, okay?

[01:02:36]Um yeah, I would appreciate a good review on Meetup. I would appreciate uh the subscription, the like on the channel. Um thank you very much, everybody, and have a have a great evening.