WEBVTT 00:00.280 --> 00:05.000 Now we are going to take on the board and build an AI resume analyzing system. 00:05.000 --> 00:07.320 That's going to be a game changer for recruiters. 00:07.440 --> 00:11.320 It will speed up your hiring process and again, it will be so powerful. 00:11.320 --> 00:12.880 So watch this video to the end. 00:12.920 --> 00:19.200 Because what it does, it retrieves all of the data actually from the resume and also evaluates the 00:19.200 --> 00:24.000 candidate based on its experience and previous jobs. 00:24.560 --> 00:28.000 But now, without being too much, let me show you how it works. 00:28.000 --> 00:31.400 So I sent an email to myself with the resume of John Smith. 00:31.440 --> 00:35.000 This is the sample, and inside I've got the summary experience. 00:35.160 --> 00:37.760 Um, you know, I've got the education certifications. 00:37.760 --> 00:41.240 So all of the information that proper resume should have. 00:41.680 --> 00:45.760 Um, and then what I can do, I can actually execute the workflow. 00:45.920 --> 00:52.360 So simply, uh, what it does, it retrieves for us this specific resume based on the filters we filter. 00:52.360 --> 00:54.360 So we check exactly if it's the resume. 00:54.400 --> 00:56.000 We upload the file to Google Drive. 00:56.000 --> 00:56.960 We download it. 00:57.000 --> 00:59.440 We use this node to extract the information. 00:59.680 --> 01:07.200 And also we use AI to also extract their information and also a agent to evaluate our candidate. 01:07.240 --> 01:09.680 After we do this, we get everything. 01:09.680 --> 01:17.880 So the name, email number, skill set, qualifications even to additional fields from our agent, um, 01:17.920 --> 01:19.960 which are the score and justification. 01:19.960 --> 01:22.520 So we've got a score which is nine and justification. 01:22.520 --> 01:28.400 So John Smith's extensive experience in full stack development, particularly with JavaScript frameworks 01:28.400 --> 01:34.200 like react and backend systems in node JS, aligns closely with typical requirements for a full stack 01:34.200 --> 01:35.440 developer role. 01:35.440 --> 01:36.040 So that's great. 01:36.040 --> 01:37.760 Actually, we've got all of the information. 01:37.800 --> 01:44.120 And again it may be a huge time saver for recruiters if you manage a lot of resumes. 01:44.440 --> 01:50.680 Um, also what's very important you upload the resumes inside a specific folder so you store them actually 01:50.720 --> 01:52.640 also use AI to evaluate them. 01:52.640 --> 01:53.360 That's great. 01:53.840 --> 02:00.080 Um, additionally, I need to mention, remember you can change the trigger, so maybe you've got another 02:00.080 --> 02:02.040 place where you store the resumes. 02:02.040 --> 02:03.160 It may be the slack. 02:03.200 --> 02:08.160 I don't know, maybe notion, um, maybe other specific place. 02:08.160 --> 02:10.240 And the only thing you need to change the trigger. 02:10.280 --> 02:11.040 That's great. 02:11.080 --> 02:13.320 Um, for example, you've got here the alternatives. 02:13.600 --> 02:19.680 But for now, I recommend you to watch this video to the end, because actually, I will guide you through 02:19.680 --> 02:21.920 each step in the specific automation. 02:21.960 --> 02:29.280 Also will have some, I can say, special and unique ways of how we set up this, um, specific notes. 02:29.600 --> 02:34.080 Um, and for sure watch it to the end because it will be so powerful. 02:35.200 --> 02:38.640 So let's go to the point and set up this entire project. 02:38.640 --> 02:42.760 And the first thing we want to add, as usually is the trigger. 02:42.760 --> 02:45.120 And for that we are using Gmail. 02:45.560 --> 02:47.400 So we've got on message received. 02:47.640 --> 02:55.000 And here I can say we have something special and some terms because I will be applying the filters and 02:55.000 --> 02:56.240 also the prefix. 02:56.480 --> 02:58.750 So for that case what we have we've got the moat. 02:58.790 --> 03:04.990 We can leave every minute or just you can actually no, every hour check every day for new emails. 03:05.350 --> 03:09.190 However, the problem is we don't want to check for every email. 03:09.190 --> 03:12.270 We don't want to get like every email we receive. 03:12.590 --> 03:20.550 We want to indicate, okay, I want to have an email that contains the specific attachment that should 03:20.550 --> 03:26.230 be actually the PDF, because this is the format that, um, you know, um, most people use for their 03:26.230 --> 03:26.790 CVS. 03:27.990 --> 03:33.470 Um, and also it should be the resume because obviously we may have, I can say the like the PDFs, 03:33.790 --> 03:36.110 um, about many different things. 03:36.110 --> 03:42.910 So we need to actually somehow filter it, filter all of the emails so it works the proper way we can 03:42.910 --> 03:45.710 add the filter, um, which is the search. 03:45.950 --> 03:46.630 That's great. 03:46.870 --> 03:53.910 And now the first thing that comes, um, to our mind and also we see it, um, is this specific comment 03:53.910 --> 04:00.590 which is has um, we've got the colon attachment If we go there to this link. 04:00.590 --> 04:05.070 So more info, you will see how this specific field works. 04:05.270 --> 04:08.230 So we've got different operators like from. 04:08.230 --> 04:11.230 So find emails sent from a specific person. 04:11.270 --> 04:14.790 We've got a tool so find emails sent to a specific person. 04:15.150 --> 04:20.710 And here when we scroll down we've got for an example file name we'll be using file name PDF. 04:21.110 --> 04:24.990 Uh you've got different operators you can use. 04:25.390 --> 04:31.750 And that's also actually the tip for you in the future if you're using Gmail note which is for sure. 04:31.790 --> 04:39.070 Like I guess, uh, like a lot of people, um, also me are using it every day for every automation, 04:39.070 --> 04:40.310 almost every automation. 04:41.230 --> 04:42.710 Um, you can apply these filters. 04:42.710 --> 04:43.910 They are very useful. 04:44.390 --> 04:51.190 So the first thing I want to actually provide in the filter so in the search option is has colon attachment. 04:51.990 --> 04:55.150 What it does it checks if the email has attachment. 04:55.150 --> 05:03.030 So we are one step further in this automation, so we don't take all of the emails that we receive in 05:03.030 --> 05:03.830 our inbox. 05:03.830 --> 05:06.950 However, only these that have, um, attachment. 05:07.510 --> 05:17.870 The next thing, um, what we can provide is the file name, uh, pdf, and every file that is in the 05:17.870 --> 05:20.910 PDF format comes with the PDF extension. 05:21.030 --> 05:22.350 So we say all right. 05:22.390 --> 05:22.870 Okay. 05:23.150 --> 05:26.350 Um, whenever we've got the file. 05:26.350 --> 05:29.990 So new attachment because remember that attachment can be an image. 05:29.990 --> 05:33.230 It can be the XML like the file. 05:33.430 --> 05:35.150 So there we check for the attachment. 05:35.190 --> 05:39.310 However here we specify okay we want to have only the PDFs only the PDF. 05:39.550 --> 05:41.510 This is the format we are interested in. 05:41.670 --> 05:50.030 Otherwise just yeah you know like it won't run so it won't run on any image on any other format. 05:50.190 --> 05:54.190 Um, so we've got the resumes, um, in our PDF format. 05:54.390 --> 05:55.830 So I hope you understand. 05:56.510 --> 05:58.870 So we've got for now two search options. 05:58.870 --> 06:02.110 And the last one is simply the word resume. 06:02.230 --> 06:06.230 So we check for the word resume inside our email. 06:06.510 --> 06:08.990 So it covers 99% of the cases. 06:09.150 --> 06:10.390 So I can say that's great. 06:10.390 --> 06:12.470 We've got our search options set up. 06:12.670 --> 06:17.870 The next thing what we want to do is to disable the option, which is simplify because we want to have 06:17.870 --> 06:18.710 some options. 06:18.990 --> 06:22.110 And the first option is actually attachment prefix. 06:22.390 --> 06:26.790 So whenever we've got an attachment we want to name it as CV or even resume. 06:26.830 --> 06:32.590 However for now I would like to have the CV and additional option is download attachment. 06:32.590 --> 06:37.710 So we want to download this attachment right away after our Gmail uh trigger. 06:37.750 --> 06:38.190 Okay. 06:38.630 --> 06:41.270 For now we can fetch test events. 06:41.750 --> 06:42.950 Um actually event. 06:42.950 --> 06:47.230 And we can see we've got our file which is CV one and this is the resume. 06:47.230 --> 06:48.350 So our John Smith. 06:48.390 --> 06:49.070 That's great. 06:49.710 --> 06:52.910 Um, the next step inside this automation is Google Drive. 06:53.030 --> 06:54.350 So we simply want to. 06:54.350 --> 06:58.070 arm and upload the file to our folder. 06:58.510 --> 07:01.030 And now for the input data field name. 07:01.030 --> 07:03.630 We want to specify this specific name. 07:03.630 --> 07:05.150 So this is the CV zero. 07:06.390 --> 07:09.070 For the file name we can go ahead to schema. 07:09.630 --> 07:13.510 And now search for uh what we have here. 07:14.150 --> 07:17.790 We'll search for the specific text. 07:18.070 --> 07:23.710 So for now we can actually take the file name which will be the name of the person. 07:23.710 --> 07:26.310 So from you know we've got here the variable. 07:26.310 --> 07:32.310 So from and this is the email address from, from which we, we received the email and the resume. 07:32.310 --> 07:33.750 And this is the name of the person. 07:33.750 --> 07:35.830 So this is Chris and Vitorovic obviously. 07:35.830 --> 07:38.270 So we can take this variable and put it here. 07:38.750 --> 07:39.470 Um yeah. 07:39.470 --> 07:40.230 So that's great. 07:40.670 --> 07:48.270 And for the Google Drive um, I've taken my drive and there I've got the um, the folder which is already 07:48.270 --> 07:51.830 created, which is actually, uh, where do we have it? 07:51.990 --> 07:53.110 Job resumes. 07:54.030 --> 07:55.550 Let me open it very quickly. 07:56.510 --> 07:58.750 And inside I've got the resumes. 07:58.990 --> 08:03.590 As you can see, I've tested like this automation many times, so I've got a lot of resumes. 08:03.630 --> 08:06.630 However, you need to create your folder here to upload the file. 08:06.910 --> 08:08.070 Um, actually files. 08:08.590 --> 08:11.150 Um, then what we do, we execute the step. 08:11.470 --> 08:12.950 That's pretty straightforward. 08:13.310 --> 08:17.390 And then, um, we've got the web content link. 08:17.790 --> 08:19.150 Um, so that's great. 08:19.870 --> 08:26.150 And the next thing we want to actually download this file from Google Drive because it's needed for 08:26.150 --> 08:27.030 further steps. 08:27.350 --> 08:32.790 So we pick drive Google Drive and there download file. 08:33.790 --> 08:39.390 Um for the file we specify the ID uh and that's pretty straightforward. 08:39.390 --> 08:41.590 We just take this variable from upload file. 08:41.590 --> 08:42.750 So this is the ID. 08:43.190 --> 08:44.910 And then we execute this step. 08:45.230 --> 08:48.750 We should get the file downloaded inside our automation. 08:48.750 --> 08:49.430 That's great. 08:49.630 --> 08:50.230 Perfect. 08:50.910 --> 08:54.460 Um so we've got our three specific notes. 08:54.460 --> 09:02.740 And the fourth one is to take our PDF and extract from it the information, all of the information. 09:02.740 --> 09:04.100 That would be the best case. 09:04.660 --> 09:07.780 And now there are different approaches you can use. 09:07.820 --> 09:10.980 Maybe the tool on the internet which is PDF. 09:11.020 --> 09:12.260 I don't know maybe IO. 09:12.300 --> 09:18.460 I'm not sure about it, but you've got different applications that can extract for you the information 09:18.460 --> 09:19.260 from PDF. 09:19.860 --> 09:25.700 Um, in case you've got the image, you can use open AI Image Analyzer. 09:25.740 --> 09:31.460 However, it's not for that specific circumstance because we've got a PDF or even we've got the best 09:31.460 --> 09:35.180 approach which is actually using the extract node. 09:35.380 --> 09:37.100 So extract from file. 09:37.980 --> 09:41.860 And here you have extract from PDF. 09:43.180 --> 09:46.180 Um, simply you don't change anything here. 09:46.420 --> 09:51.100 You just click on execute step and then um that's that's very cool. 09:51.100 --> 09:55.420 You've got a text, entire text extracted by the specific node. 09:55.620 --> 09:57.580 So pretty long data. 09:57.620 --> 09:59.460 Yeah, we've got everything here. 09:59.780 --> 10:00.820 That's correct. 10:01.940 --> 10:03.420 So we have our first steps. 10:03.580 --> 10:07.940 And the next thing after we have the information we want to use. 10:08.220 --> 10:10.660 Um information extractor. 10:11.300 --> 10:17.260 This is the node that uses AI to extract specific information. 10:17.540 --> 10:21.700 And now there is the specific case we need to have the text. 10:22.260 --> 10:26.420 Um so we need to specify the text in order for it to work. 10:26.660 --> 10:30.860 We can use the variable from the previous node which is extract from file. 10:31.020 --> 10:32.900 And the variable will be text. 10:33.180 --> 10:34.860 So let's drag and drop it here. 10:34.900 --> 10:36.300 That's pretty straightforward. 10:36.780 --> 10:43.620 However actually we need to specify the schema type which is defined using JSON schema. 10:44.140 --> 10:46.140 And there provide a specific code. 10:46.980 --> 10:56.660 And what we want to receive um by This and by using this specific note and also OpenAI model three things. 10:56.820 --> 10:59.860 Firstly, the name, phone number and email address. 11:00.620 --> 11:04.900 Um, actually we need to use the specific JSON code. 11:05.020 --> 11:10.940 And for this circumstance, I prepared for you this code right away without you doing anything like, 11:10.980 --> 11:16.180 you know, you can just copy and paste it from the file that will be in the resources of this material. 11:16.420 --> 11:19.060 Also, we will find here the JSON template as usual. 11:19.700 --> 11:23.820 Uh, but what we have here we've got specific information. 11:24.540 --> 11:27.860 So the first variable we want to retrieve is the name. 11:27.860 --> 11:29.340 So the type is string. 11:29.380 --> 11:31.380 Over here we've got a name type string. 11:31.580 --> 11:34.940 The second the second one is overall the email. 11:35.220 --> 11:36.980 Again we've got the type string. 11:36.980 --> 11:38.780 Also the format which is email. 11:39.100 --> 11:41.340 And the third one is phone number. 11:41.340 --> 11:42.540 And there the pattern. 11:42.540 --> 11:44.140 We've got the specific pattern. 11:44.420 --> 11:47.100 Um I've created that with ChatGPT. 11:47.380 --> 11:49.300 And that's also great indicator. 11:49.300 --> 11:54.380 Door so you can do this stuff with your specific model. 11:54.620 --> 11:59.220 So create a code that returns for you specific information you want. 11:59.740 --> 12:03.620 Um, let's use the model which is OpenAI. 12:04.340 --> 12:08.780 Um, and we'll be using for for this particular case. 12:10.300 --> 12:13.100 So let's pick for oh that's great. 12:13.300 --> 12:17.020 And now in general let's execute the step. 12:17.420 --> 12:21.220 And we should get again the name um email and phone number. 12:21.260 --> 12:21.980 Let's preview it. 12:22.020 --> 12:22.540 Okay. 12:22.580 --> 12:24.380 We've got the name email and phone number. 12:24.420 --> 12:25.260 That's correct. 12:25.540 --> 12:26.100 Nice. 12:26.780 --> 12:33.660 Um, the next note will be our agent that will be responsible for more things. 12:33.940 --> 12:36.300 So let's let's preview it. 12:36.340 --> 12:36.780 Okay. 12:37.060 --> 12:43.220 Um, it will retrieve for us the skill set, qualifications, job history, score and justification. 12:43.220 --> 12:45.300 So five different variables. 12:45.940 --> 12:47.220 And that's really cool. 12:47.220 --> 12:50.620 So we've got three variables that are static from our resume. 12:50.660 --> 12:54.780 However, these two are given by our agent. 12:55.180 --> 13:02.540 Um, that's cool because not only you can store your resumes, you can track them via multiple platforms. 13:02.540 --> 13:06.220 So as I said at the beginning, you may have different triggers. 13:06.260 --> 13:11.180 Let's say slack, let's say notion, let's say I don't know, maybe outlook. 13:11.340 --> 13:16.860 Um, however, also um, you can actually evaluate the candidate. 13:17.100 --> 13:23.540 So not only again save the information however also like give it some score or justification. 13:23.540 --> 13:24.420 Why why why. 13:24.460 --> 13:27.980 It's for an example nine not 5 or 4. 13:28.340 --> 13:30.660 Um, so I hope you know what I mean. 13:31.780 --> 13:34.300 Um, and it's pretty useful for recruiters, obviously. 13:34.500 --> 13:36.620 But let's go back to setting it up. 13:37.140 --> 13:41.220 Uh, for the model, we can connect it also with OpenAI model. 13:41.220 --> 13:49.740 So use one model and provide right here as typically our prompt and system the message for the prompt. 13:49.820 --> 13:54.340 Um, we've got the resume, so let me pick it. 13:54.660 --> 13:55.620 We've got the resume. 13:55.620 --> 14:03.580 And what we do, we take the text from extract from file note and for the system message it's more sophisticated, 14:03.580 --> 14:04.820 more more advanced. 14:04.820 --> 14:06.140 But let's preview what we have. 14:07.020 --> 14:08.460 So we've got the purpose. 14:08.460 --> 14:11.340 So we are a CV evaluation and summarization agent. 14:11.340 --> 14:17.660 Your task is to extract key information from a CV and output it in a structured format. 14:17.700 --> 14:25.140 You must also evaluate how well the candidate, um, fits a specified job role and assign a role score. 14:25.140 --> 14:25.940 Relevance. 14:26.100 --> 14:32.500 Um, we can even, um, put it as score 1 to 10, followed by short justification. 14:33.180 --> 14:34.460 We've got how it works. 14:34.460 --> 14:39.260 So when provided with a job with a CV and job description, follow this process. 14:39.260 --> 14:42.260 So firstly extract summarize information from the CV. 14:42.580 --> 14:45.540 Um so we've got three specific variables explained. 14:45.860 --> 14:49.970 And secondly we've got to evaluate the candidates for the job role. 14:50.250 --> 14:54.370 So compare the candidate qualifications and experience with the job requirements. 14:54.410 --> 14:57.810 Assign a score from 1 to 10 based on relevance. 14:58.170 --> 15:02.090 And we've got the explanation for the rating for the score system. 15:02.290 --> 15:09.250 So 1 to 3 is weak match 4 to 6 moderate strong which is seven eight and excellent which is 910. 15:09.770 --> 15:12.010 Provide a brief justification for the score. 15:12.050 --> 15:15.810 Mentions strengths and gaps directly related to the to the role. 15:15.850 --> 15:18.210 No assumptions or interpretations. 15:18.690 --> 15:22.210 This is pretty everything that we need, only like the prompt. 15:22.210 --> 15:25.610 So the text uh from our extractor. 15:25.770 --> 15:30.010 And this is the message and it should give us, um, the data. 15:30.570 --> 15:36.330 Uh, what's very important also, uh, we should get actually, uh, one field. 15:37.250 --> 15:42.610 So it will be the output our model thinks. 15:43.530 --> 15:44.410 And here. 15:44.450 --> 15:45.170 That's great. 15:45.250 --> 15:47.250 Um, I guess we've got everything right there. 15:48.610 --> 15:49.090 Okay. 15:49.290 --> 15:53.610 Later we'll check it, because sometimes, um, it may have some problems. 15:53.610 --> 15:55.610 I had some problems with this automation. 15:55.610 --> 16:02.210 Like, um, sometimes it didn't return for an example like skill set in a specific form. 16:02.250 --> 16:05.010 I had the problems later, so I will show you what I mean. 16:05.050 --> 16:08.810 However, I've got the way, um, to to prevent the problem. 16:09.130 --> 16:11.290 Um, but let's move on. 16:12.530 --> 16:18.610 The next thing what we want to do is to add the edit field to clean up a little bit the data. 16:18.970 --> 16:20.530 Um, I mean, not clean up. 16:20.530 --> 16:22.450 However, change the name of the variable. 16:22.650 --> 16:26.610 So to the content, uh, it's pretty straightforward. 16:26.610 --> 16:27.970 We can execute this step. 16:28.210 --> 16:31.450 Um, and here we've got a content and this exact data. 16:31.930 --> 16:34.770 Um, next what we have, we've got the code node. 16:35.730 --> 16:41.930 So what we want to do, uh, we've got the information from our agent. 16:42.170 --> 16:49.570 However, for now, we want to take all of this information and structure it in the specific variables. 16:50.810 --> 16:53.250 We are using the code node, which is really powerful. 16:53.290 --> 16:56.690 Like it's one of the most powerful nodes inside notion. 16:57.170 --> 16:58.890 And so we want to specify the code. 16:59.930 --> 17:04.010 I've got a code for you which is this JavaScript entire code. 17:04.370 --> 17:09.530 Um, that the only role is to extract for us all of the information. 17:10.050 --> 17:11.210 I won't explain that. 17:11.210 --> 17:14.970 I mean, like, you know, um, it's really sophisticated. 17:15.250 --> 17:19.690 Um, I use ChatGPT to create it, um, to customize it for me. 17:19.810 --> 17:23.330 I change also some stuff, um, because I had some problems. 17:23.330 --> 17:24.970 However, it works for now. 17:24.970 --> 17:29.690 So I came up with this entire code and let me show you what it does so we can execute this step. 17:30.330 --> 17:33.210 And it gives us actually the information. 17:33.210 --> 17:34.530 It gives us the information. 17:34.530 --> 17:38.850 So education qualifications, job history, skill set score justification. 17:39.290 --> 17:41.530 And the problem is we've got the null values. 17:41.570 --> 17:48.010 And I said I've got I had the problems with this specific this specific automation. 17:48.730 --> 17:50.730 So let me show you what we can do with that. 17:51.290 --> 17:56.370 Our agent misinterpreted the instructions and provided the form. 17:56.410 --> 18:02.410 Actually the output that doesn't align with the code note and the exact code we have inside. 18:02.810 --> 18:17.450 And the walkthrough is to go here and type up, provide the form of the answer that aligns this specific 18:17.450 --> 18:31.210 code to retrieve, that retrieves the information that, uh, retrieves the information, make it this 18:31.210 --> 18:40.730 way so it works, and what we can do, it's a great, actually trick we can paste the entire code we 18:40.730 --> 18:43.810 used in the further step of this automation. 18:44.210 --> 18:49.850 So we give our agents some background, some information, how the code looks like. 18:49.890 --> 18:56.490 And then it takes into account, um, that data when creating the response. 18:56.970 --> 19:00.490 We can again execute the step and see if that solved the problem. 19:01.130 --> 19:02.610 It's running, running and running. 19:02.650 --> 19:05.250 And after a few seconds we've got our content. 19:05.250 --> 19:10.050 So now what we can do we can go back to canvas and test out again the code. 19:10.090 --> 19:10.530 Note. 19:10.530 --> 19:11.330 So let's see. 19:11.810 --> 19:12.250 All right. 19:12.250 --> 19:13.690 So still we've got a problem. 19:13.890 --> 19:18.170 However I guess maybe it's the problem with uh with the sentence. 19:18.610 --> 19:25.090 So, um, provide a form of the answer that aligns the specific code that retrieves the information 19:25.090 --> 19:33.010 so I can type, um, this is the JSON code um, we are using. 19:36.250 --> 19:37.970 Um, to retrieve. 19:40.290 --> 19:41.130 Retrieve? 19:41.370 --> 19:46.360 Um, data provide output. 19:46.560 --> 19:51.280 Um, so it will work. 19:52.120 --> 19:54.320 So it works with it. 19:55.200 --> 19:55.920 Perfect. 19:56.120 --> 19:58.520 Um, let's again execute this step. 19:59.640 --> 20:01.640 Um, I'm very curious. 20:01.800 --> 20:03.040 So we've got the output. 20:03.080 --> 20:03.680 It thinks. 20:03.720 --> 20:07.080 It thinks we've got again, the content. 20:07.760 --> 20:08.960 Let's run the code. 20:09.800 --> 20:11.400 It's running, running and running. 20:11.400 --> 20:13.720 And after a few seconds, we've got our output. 20:13.840 --> 20:16.080 So now let's test out the code. 20:16.120 --> 20:19.480 Note once again, um, hopefully it works. 20:19.640 --> 20:19.920 Yeah. 20:19.920 --> 20:20.760 So we've got it. 20:20.800 --> 20:25.080 Educational qualifications, job history, skill set, score and justification. 20:25.400 --> 20:27.000 That's really cool. 20:27.760 --> 20:30.000 Um, and I can say this is really everything. 20:30.160 --> 20:35.840 So now with all that information, we can save the data inside, um, Google Sheets. 20:37.480 --> 20:40.480 So we can append the row and sheet. 20:41.320 --> 20:44.600 And when we are here, we can pick the specific document. 20:44.920 --> 20:47.080 So list of candidates. 20:47.080 --> 20:48.520 Um, let me show you. 20:48.760 --> 20:53.040 So in general in the resources also we will have this Google Sheets to copy. 20:53.080 --> 20:56.440 You can open it and then make a copy and use it. 20:56.480 --> 21:02.640 So again inside we've got name email number, skill set qualifications, job history score and justification. 21:03.200 --> 21:06.360 Uh and we need to map out the specific fields. 21:07.280 --> 21:08.560 So we have the name. 21:09.120 --> 21:11.600 Um, actually let's go here. 21:11.800 --> 21:20.960 Um, let's pick the name from information extractor, email, phone number and Rest information from 21:20.960 --> 21:22.160 our agent. 21:22.280 --> 21:24.200 I'm not actually our code node. 21:24.480 --> 21:35.360 Okay, so we have the skill set, um, qualifications, job history, score and justification. 21:35.720 --> 21:36.480 That's correct. 21:36.480 --> 21:37.920 We can execute the step. 21:39.000 --> 21:40.920 So it should appear here. 21:41.160 --> 21:41.720 Great. 21:42.400 --> 21:44.080 And now what we can do. 21:44.360 --> 21:49.640 Um, we just created our automation, so let's test it out once again. 21:49.880 --> 21:51.880 And for this specific case. 21:51.880 --> 21:59.040 So for testing and creating this project we had the John Smith, which is really, I can say experienced 21:59.040 --> 22:01.960 person in front end and back end. 22:02.000 --> 22:08.520 So like this person is really relevant for the job description and for the job, uh, for this specific 22:08.520 --> 22:08.800 role. 22:08.840 --> 22:09.280 Okay. 22:09.520 --> 22:16.680 But now I've got other resume, um, that I can use for the person that is not that, um, experienced 22:16.680 --> 22:18.040 for this specific role. 22:18.040 --> 22:23.600 So we can see if the entire automation works and if the score is different, justification and so on. 22:23.600 --> 22:25.480 So we can test out the entire system. 22:25.840 --> 22:26.560 For now. 22:26.600 --> 22:30.840 Let me send uh, my email actually email to myself. 22:32.080 --> 22:34.920 So now I've got the document for Emily Turner. 22:35.360 --> 22:37.800 And there again the the similar action form. 22:37.800 --> 22:40.800 So summary experience education Certifications. 22:40.800 --> 22:41.440 Projects. 22:41.800 --> 22:44.800 Um, so we can test out this specific workflow. 22:44.960 --> 22:45.960 Execute workflow. 22:45.960 --> 22:47.880 And we can preview if everything works. 22:47.880 --> 22:49.680 So yeah, we've got Emily Turner. 22:49.840 --> 22:50.080 Um. 22:50.080 --> 22:50.800 That's great. 22:51.200 --> 22:54.000 And now it it's extracting for us. 22:54.000 --> 22:57.320 The data uses our agent to do all of this stuff. 22:58.080 --> 23:01.080 And now we can see if we have the information. 23:01.640 --> 23:02.800 So let's wait. 23:02.800 --> 23:04.040 Yes we've got it. 23:04.040 --> 23:12.240 So name email number, skill set and the score which is seven because Emily Turner is a strong match 23:12.240 --> 23:18.120 for a junior developer role due to her completion of a full stack web development bootcamp and active 23:18.160 --> 23:24.720 involvement in practical projects, which align well with the technical requirements for a junior developer. 23:24.760 --> 23:31.320 Her skills and HTML, CSS, and JavaScript and react are crucial for many entry level positions. 23:31.360 --> 23:37.960 However, this um, actually this experience is worse than the John Smith. 23:38.240 --> 23:42.360 So in that case we've got the lowest score, which is seven. 23:42.760 --> 23:43.560 And that's great. 23:43.920 --> 23:46.480 We created the entire automation from scratch. 23:46.840 --> 23:50.120 Also remember you can use different triggers. 23:50.120 --> 23:57.560 So let's say you can use the trigger which is slack whenever you've got the attachment in um in your 23:57.600 --> 23:58.280 channel. 23:59.160 --> 24:03.920 So let's say on new message on new message posted to channel. 24:05.160 --> 24:10.920 And there you can choose the specific, you know, um, circumstance like how it should behave. 24:11.560 --> 24:17.440 Um, also, um, you've got the other trigger, which is it can be the on form submission. 24:17.440 --> 24:23.760 So you can specify that you want to have um, the file and then you require the file and someone uploads 24:23.760 --> 24:25.680 for you, um, the PDF. 24:26.400 --> 24:33.000 Um, so again you can experiment with this automation and change it for your specific application you 24:33.000 --> 24:34.600 use for hiring process. 24:35.320 --> 24:39.080 Um, additionally, we can test out another thing. 24:39.440 --> 24:46.880 Let's say someone sends us an email that contains, um, let's say image. 24:47.040 --> 24:51.880 Not not actually, um, you know, not actually the resume and PDF format. 24:51.920 --> 24:54.160 However, the image, maybe it's the resume. 24:54.160 --> 24:54.320 Okay. 24:54.320 --> 24:55.440 So let's do the following stuff. 24:55.480 --> 24:58.840 I can send the email to myself and now let's type resume. 24:59.440 --> 25:03.080 It's my resume okay. 25:03.600 --> 25:06.000 Um, now I can attach some image. 25:06.000 --> 25:07.760 Let's say it will be some random image. 25:07.920 --> 25:12.560 So we'll just, you know, like test the system because like, the title is resume. 25:12.560 --> 25:13.560 This is my resume. 25:13.840 --> 25:18.160 So, um, let's let's test out our filters for the image. 25:18.160 --> 25:20.040 It will be whatever it can be. 25:20.040 --> 25:20.760 Like this. 25:21.040 --> 25:27.360 Uh, what we have, um, this place card, even, uh, let it let it upload. 25:28.240 --> 25:30.560 Okay, so I just uploaded some other image. 25:30.560 --> 25:31.200 Whatever. 25:31.240 --> 25:34.880 Let's send the email and let's even wait a few minutes. 25:35.000 --> 25:38.230 Um, so we can make sure actually, you know, it works. 25:38.230 --> 25:41.590 And then we can run this specific workflow once again. 25:42.590 --> 25:44.390 Let's execute the workflow. 25:44.390 --> 25:46.710 And there is the moment of truth. 25:46.750 --> 25:46.990 Okay. 25:47.030 --> 25:49.470 So it took the Emily Turner resume. 25:49.470 --> 25:51.550 So it didn't take the last attachment. 25:51.590 --> 25:52.070 Okay. 25:52.430 --> 25:54.070 It only took the resume. 25:54.070 --> 25:55.030 So that works. 25:55.310 --> 26:00.550 Another thing we can send the resume however don't provide the word which is resume. 26:00.790 --> 26:04.430 So in that case we can see if everything is correct. 26:04.830 --> 26:10.590 Um, for that case, what we can do, we can actually again send an email. 26:11.790 --> 26:14.110 So I have the title of the email which is hey. 26:14.150 --> 26:16.510 And in the description I've got a job application. 26:16.550 --> 26:21.070 However for the attachment I've got the PDF and it's also an attachment. 26:21.270 --> 26:24.190 However it's like you know like different attachment. 26:24.190 --> 26:27.510 This is for creating clone agent whatever. 26:28.030 --> 26:33.470 Um, however in that specific case and we've got two requirements, um, actually fulfilled. 26:33.470 --> 26:35.870 So attachment file name. 26:36.070 --> 26:42.710 However, we don't have the resume, so it shouldn't take this specific email as a case for our automation. 26:43.110 --> 26:45.470 Um, let's send this specific message. 26:45.470 --> 26:48.430 And again, wait one two minutes to make sure. 26:48.750 --> 26:49.790 Uh, yeah. 26:49.910 --> 26:53.710 It actually, um, detected our email. 26:54.190 --> 26:56.030 Let's execute the workflow. 26:56.030 --> 26:59.110 And again, we've got Emily Turner. 26:59.310 --> 27:01.270 So perfect okay. 27:01.310 --> 27:04.270 It didn't take any other information. 27:04.270 --> 27:11.710 So every time, you know, it checks for the latest, uh, resume, um, that you have inside your, 27:12.110 --> 27:14.230 um, inside your Gmail. 27:14.750 --> 27:18.590 And for this automation, for this video, I can say that's everything. 27:18.590 --> 27:19.430 Thank you for watching. 27:19.430 --> 27:20.670 I hope you enjoyed it. 27:21.070 --> 27:27.550 Uh, and, yeah, I think it's a great automation that may help you, um, in many different cases, 27:27.590 --> 27:31.990 can save you a lot of time and definitely can speed up your hiring process. 27:32.230 --> 27:35.750 Thank you for watching again and I will see you in the next material.