
WebMakers Talks: How do AI micro-workshops with WebMakers work?
Topics covered:
We invite you to listen to the latest episode of our podcast, WM Talks, where we discuss technology and business topics related to the IT industry.
Damian: Hi, this is Damian, and today I’m joined by Szymon and Łukasz. We’d like to present how our micro-workshops on artificial intelligence work. To do this, we’ll create a small presentation where I’ll play the role of a business owner, a prospective client of our company, who will talk to the team here. Together, we’ll conduct a micro-workshop during which we’ll identify the needs of my company, explore the possibilities of implementing AI in my business, and determine the next steps for introducing artificial intelligence into my enterprise.
Łukasz: Hi, Łukasz here.
Szymon: Hi, Szymon here.
Damian, you came to us with a specific problem, which you had already described in the questionnaire. Since this is a micro-workshop, we’d like you to explain the problem in your own words. We’ll then try to extract what’s most important and valuable for you.
Damian: As I mentioned in the questionnaire, we’re a company managing other companies involved in fleet operations. In short, we manage driver databases, handle their records, process their work schedules, plan their salaries, and also recruit drivers.
The biggest issue we face, and the one that takes up the most time, is gathering information from the employees of all these companies. There are multiple companies involved, each operating by its own rules and processes. We cannot enforce the use of a centralized system. Each company sends us documents in its own way. Sometimes they come via email, other times as physical documents that we then have to scan or manually enter into our system. Sometimes we get information entered into various CRM systems. There’s no consistency, and we end up with a huge mess of data. We can’t make external companies input everything into a unified system, so we must handle it ourselves. This takes an enormous amount of time and human resources. My question for you is whether it’s possible to implement automation or artificial intelligence to make this process faster and more efficient, relying on software instead of manual labor.
Szymon: All the issues you’ve mentioned are solvable. However, there are several challenges your company is dealing with. To start, we’d like to focus on one specific issue. You mentioned data management, particularly gathering data from multiple sources. This is a fascinating yet broad topic. Could you tell us which of these processes is currently the most costly or urgent to improve? That’s the one we can address first, then propose possible solutions.
Damian: Recently in my clients' network there is a terrible demand for new employees, for new drivers, and due to the fact that our company deals with this comprehensively, well, we are also responsible for hiring these people, i.e. we are the ones who conduct the recruitment, candidates apply to us by sending their CVs, and we are the ones who actually classify them and only send the ready candidates to our clients. Well, because there is a lot of this going on lately, and also the candidates who apply to us send their CVs or their applications in various ways. It takes us a great deal of time to properly collect these CVs, to extract from them all the information that is necessary, which is actually used later to classify whether a candidate is suitable or not. This is, one could say, like the biggest pain in recent times. It consumes a lot of our time, especially since these resumes a day are counted in the hundreds. Each processing of such CVs simply takes a lot of time.

Łukasz: Damian, I’d like to ask if you’re currently using any IT tools, applications, or systems to support this process. Do you have any tools for CV classification or processing? Is it handled by a team, or do you have a central repository for this data? Is the process structured or more ad hoc?
Damian: Yes, we’ve looked into software solutions for classifying CVs. However, the diversity of CV formats is a major issue. Some CVs come as Word documents, others as PDFs, and sometimes even as photos. It’s hard to expect candidates to conform to a standard format. The tools we tested didn’t fully solve the problem because the documents aren’t always consistent. Ideally, we’d like candidates to follow a specific format, but it’s unrealistic in our market. We ended up using multiple tools to get a partial solution, which was far from ideal. We couldn’t achieve the uniformity and centralization we wanted.
Łukasz: So, part of your team standardizes the data from the CVs, organizes it, and then passes it on to another team that analyzes the information to identify the most promising candidates. Is that correct?
Damian: Here we have the first people who check CVs, which they classify to the appropriate companies, because companies are also looking for different drivers. After such a preliminary classification, from each resume we have to extract the relevant information, enter it into our CRM system, where the most crucial, interesting information is already included. Then, once we have all this entered in this central system, we also do not have here any such, so to speak, automaton, which would advise us whether this driver is more suitable for this company or that one. We simply have to do it more by feel, based on the experience of our employees, and we would like it to be done in a slightly more, one might say, such an intelligent way already at the system level, and that a person would really only check whether it was classified well.
Szymon: That is, I understand that in fact the solution you are looking for includes several things. First of all, collecting data from various places, that is, just from Word, from PDFs, from photos, or even from documents delivered physically. Then storing that in some place digitally and cataloging those people and getting some kind of score against the drivers that you might be able to work with in the future, yes? Is that how you imagine it?
Damian: Yes, exactly in this way. That's how I would imagine it would look like in the end, so that the whole process you've just outlined here would eliminate as much as possible this human factor, because here I haven't said it before, that it often happens so that when transcribing some information from a resume, there are some mistakes, which also later on affects our final decision as to whether we have classified a given driver correctly and whether he should really be hired or not.
Łukasz: I don't know if Simon will agree with me, while in this process that Damian briefly presented to us, the biggest pain, the most cumbersome stage is the downloading and organizing of data. And I think this is the stage that we should address and try to automate it in the first place.
Damian: And this is where the question to you, because as I mentioned, here are various sources you can say these CVs. Well, and what do you suggest? Is it at all possible to make it so that even a physical or email document we will be able to just somehow automatically upload into one place, into some central database?
Szymon: As much as possible. We are able to prepare simply a soft, into which will land all the data, CVs that you collect in a specific period. Of course, for those that are delivered in physical form, you would have to take at least a photo of them, or scan them, and for them to go directly into the soft. For the others, on the other hand, there is no problem at all. At this point we simply throw all these files into the software. It, with the help of just AI and with the help of tools such as OCR, will at least scan all these documents, extract specific information from them, then catalog them, organize them according to what you actually expect from these resumes, because I assume that each of these resumes also does not necessarily always look the same.That is, education from someone can be entered where they have certificates, or vice versa. Also, we are able to take care of this information chaos that appears on resumes, just by extracting such information directly from the context. That is, if someone actually titled something as experience, but put there, for example, certifications, then with the help of AI, with the help of artificial intelligence, we are able to classify it on the basis of the date, on the basis of some specific information, that this is, say, a driver's license, or some language certificate there, that these are certificates, however, and not experience. Well, and in this case simply such data will go into the appropriate table with certificates, instead of as the candidate entered earlier in the experience. Also, we are both able and able to filter this type of data, catalog it and collect it, first of all, from different sources. This is as much as possible to prepare, and in this way we would see the kind of soft you want to actually get.
Create your AI-based solution with us.
Łukasz: We have reached the place where we should raise one more important issue. It is about data security. We mean both the security of the data you process and the security of the business processes you secure in your own company and in the company of your customers. When it comes to artificial intelligence, we can use external solutions like tools provided by giants like Chat GPT, or Microsoft, or Amazon. On the other hand, if we care much more strongly about security and data protection, we can use internal tools that do not transmit any relevant information outside the organization. When it comes to resumes, personal data of candidates, we can already see here that there is information we should protect. Are there still any points in your process or your organization that we need to pay special attention to, so that no sensitive information is passed on to the outside?
Damian: This is a very good point. All in all, this is also what I forgot to mention earlier, that because of the fact that here are personal data, this is sensitive data, because it also speaks not only about the name or personal data, but also speaks about someone's career in fact, and here we must really take care of this security, so it is good that you raised this, it is extremely important. When it comes to such additional security issues, we are also strongly concerned that also a given candidate, who was not intended for a specific company, so that also here there is no mistake, that we will send the personal data of a candidate who in fact was a candidate for another company, and so that simply by accident this data does not leak, not somehow we can say publicly, but simply within our network of contacts.
Łukasz: I think that here, too, we will be able to apply tools that will be able to anonymize this data, or hide the photo, name, surname, before the final analysis of this data contained in the resume forms, so that the data that will finally reach the recipients will be anonymized.

Damian: This is also a great idea, because in fact honestly the company that works with us does not need to have full personalities, in fact it judges a person by his experience, by his earned certificates, how long he has been driving. It is not really important to him at this stage the name, the surname. It's only during such a final recruitment process already, so I like this idea as much as possible and it's a great solution.
Łukasz: I am glad, then we reach one more important point. You are a company that has been working or operating in the market for a long time, and you have certainly developed your processes and standards. When implementing a new solution, we have to adapt to these standards that you have already developed. Are you able to prepare for us, not necessarily now, but after our meeting, a list of the IT tools you use, so that we can also get an idea of what tools we will have to work with and how to prepare the whole solution so that it seamlessly complements your offerings, your tools, makes an evolution, helps, but does not make a revolution?
Damian: Yes, sure. No problem. We can provide the whole list. We can also more or less try to dissect our processes that govern right here the whole course of collecting this resume and processing it. The question is, do you have any guidelines on how this process should be dissected, what it should look like?
Szymon: Yes, as much as possible. We actually have a whole manual on how you should go through all the business processes, break them down step by step. It takes about, let's say, two hours of your time to put together most of the company's processes, and in fact, based on this information, we are already able to start working, we are already able to start talking about implementing new solutions to your company.
Damian: Okay, so I understand that this is simply such a, you could say, a download in quotation marks that will instruct me step by step on how to write out this process?
Szymon: That's exactly right. It's actually Excel, where you'll have all the information on how to break down a particular process, what it looks like, what's the next step that should be in your company at a particular stage. It will guide you by the hand to know how to provide us with the specific information that we're just going to work on.
Damian: Ok, that sounds very good.
Łukasz: Dissecting such processes is a value in itself for you and your company, because already abstracting from the IT solutions you can apply, you and your team will review the steps for yourself, see where you are wasting the most time, which parts of the process can be optimized, even without using any IT systems at all. This is generally the big picture of your company and a good point for optimizing the whole thing.
Szymon: And above all, too, you'll notice places where you'll still need automation, and some things will just fall into place for you. The next time we simply meet, you'll have concrete information about the fact that not only, for example, you'll need to automate the recruitment process, but let's say some areas in your company that you don't realize at the moment are still to be improved, and we'll be happy to help.
Damian: Okay, that sounds very good.
Łukasz: If after our workshop, because this is a micro-workshop, also we can't take up much of your time. You have given us a great deal of information and we have a good basis to work on and analyze your problems. If any questions arise with you, or you want to detail something, or work through a different process than the one we've discussed now, feel free to contact us.
Damian: Okay, all in all, I have one more question at the very end, maybe more of the nature of the possibilities of artificial intelligence itself, because you hear a lot about it. As we discussed earlier, here for me the important thing is just the importing of this data, then the subsequent storage, but also we often have something like this, that our clients ask us if we are able to filter certain candidates, who, because this is the transport sector here, has it that there is a high tendency of driver turnover. Would this artificial intelligence be able to further verify whether someone has changed jobs frequently? If so, it would also simply classify such resumes accordingly.
Szymon: As much as possible, we are able to search for all those people who frequently change jobs based simply on their experience that they have on their resume. Well, and let's say we set ourselves such a filter for changing jobs every year. We can even approach it this way, or changing jobs every month and that we reject such candidates right away. As much as possible we can set this kind of filters. And at this point we simply get only a list of candidates who meet specific conditions. This doesn't only apply to experience alone, but we can actually focus on that, for example, because this is a transportation company. Some people need specific certifications, for example, for the transportation of hazardous materials or that kind of thing. We can also prepare filters that specifically indicate all people with specific certifications. At this point it will definitely simplify your process and suggest the right people for the right job.

Damian: Okay, that sounds very good, and in general when we are dissecting the process we will also add such a function to be implemented as well, because as I said, it is quite common, it happens that our clients come forward to just filter out the right type of people with whom they simply would not want to work.
Łukasz: As icing on the cake, we can add to you and your organization, your employees a completely new interface and way of communicating with the information system. Your employee will be able to ask a question in human language, i.e., for example, search for 30 candidates who have resigned more than three times in the last year from their current employer. No complicated database access queries or filters are needed here. All an employee has to do is ask a question in his or her own words, and today's artificial intelligence is capable of converting this into real and competitive data, which it will receive back.
Damian: Oo, so I understand that it would work like ChatGPT, that I just ask him something and he answers me?
Łukasz: Here, of course, with the assumption that we have to send specific information to ChatGPT for it to process. In contrast, in the solution we are discussing, no confidential data will be transferred to the outside world.
Damian: This sounds very interesting and we will certainly talk about this topic already on the next steps of our cooperation.
Szymon: Sure. Then thank you very much for today's workshop. It is an amazing amount of knowledge that you offered us. Well, and all in all, we are waiting for the rest of the data, and in the meantime we will prepare the basic information that we talked about today, that is, the functionalities that we can propose for your team and the basic estimation of the working time that is needed to implement such a solution already in your company. Also, I think we have everything.
Damian: Well, I also thank you very much for today's meeting. A lot has been outlined here. Well, and we will be in touch.
Szymon: Sure. Thanks a lot.
Łukasz: Thank you. The micro-workshops, as the name suggests, take very little time, and I think they set a good example of how we can already turn the little information we get from you into valuable knowledge, which we will then translate into solutions that help you in your daily work.
Szymon: Micro-workshops, of course, we also conduct for all companies, for any type of business. It's not just shipping and logistics, as was the case with Damian as our client today. In fact, any industry is welcome here.
Damian: As the guys have presented here, we have actually shown you today only a fragment of how the micro-workshops look like with us, to which you are cordially invited.