Podcast: Resume Matching App

Listen to the podcast

Transcription

Introduction: [00:00:00] HR at Flexso for People.

Anne Vancoillie: [00:00:02] Hi, everyone, I'm Anne Vancoillie, your host today, and this episode is about the Resume Matching app. So today we have two guests: Andreas De Neve from TechWolf and Tom Mentens from Flexso for people. Together, they created and implemented a brilliant new application on top of SuccessFactors recruitment called The Resume Matching App. This app analyses the resume of an applicant not only based on the available words, but also on related skills. Andreas De Neve is CEO of TechWolf, a machine learning startup and Tom Mentens was already a guest in our podcast before. And is Competence Lead and HR Transformation consultant at Flexso for people. Welcome to both of you.

Andreas De Neve: [00:00:52] Thank you. Thank you for the invitation.

Anne Vancoillie: [00:00:55] So Andreas, maybe first some background for our listeners, what kind of business is TechWolf?

Andreas De Neve: [00:01:02] So TechWolf is an HR tech startup and we were founded in 2018 and HR tech... I think it kind of means that we operate at the intersection of AI and HR. And what we do is we help large organisations map out and track the skills of their employees. Once you know the skills of your organisations, skills that are present in the company, we can help with AI-based strategic workforce planning and we can predict the impact of attrition, ageism, um, et cetera, on the skills of your organisation. And compare that to where you should be five years from now to really identify, like, the future skill gap that a company or an organisation still needs to bridge to really get where they have to be five years from now. Then, on the other hand, and we have this skill framework, um, that we want to leverage and use across all HR operations and for example, in recruitment. And this is, uh, one of the applications that we tackle with Flexso and with the Recruiter Cockpit, but also in learning and development and pretty much all of talent mobility so that every time, um, an action is taken; a person is hired, someone followed a course, someone finished a project that the results of that action in terms of skills are measured within the same framework. And they all attribute to kind of a continuous monitoring state of how did this action contribute to the closing of our future skill gap as a company? And how did this action contribute to us getting future proof, uh, in terms of skills? So that's what TechWolf does in a nutshell.

Anne Vancoillie: [00:02:49] A mouthful. Okay. Sounds nice. And what does the Resume Matching App then do in-depth? How does the Skill Engine work?

Andreas De Neve: [00:03:00] So we use advanced AI to interpret both resumes and vacancies in terms of your background and the focus of your skills and then using our skills, we calculate the match between both on a bias-free, fair and explainable way based on relevant experience. And all of this information comes straight out Successfactors

Anne Vancoillie: [00:03:20] And why is it a new skill engine on the market? How does it relate to other products on the market? And what is it not maybe?

Andreas De Neve: [00:03:29] Just to set the scene, for example, with our Skill Engine on the... On an organisational level, you could start off by taking a look at the market- see which are the skills we need. How will the skills within my organisation evolve. When you've identified that future skill gap and the skills that you will need within five years from now b but you don't have right now, then we can go a level deeper and we can take a look at the skill gap and say, OK, maybe these and these skills, they can be trained. And we have people who have reskilling potential, upskilling potential to acquire these skills. And then there will be some subset of skills that will be very hard to reskill or upskill. And maybe those should be the people that we hire or acquire by acquisition of another company. And then within that small subset, the skill engine is, in case of the Resume Matching App with Flexso, able to link the right candidates to these skills. But as you can see, the scope of the solution is much bigger and just the matching is just one, one very small part of a larger solution, whereas most of the other solutions on the market and matching solutions - semantic technologies - they just link the candidates to the job, but they lose the bigger picture.

Anne Vancoillie: [00:04:51] Maybe a subquestion. Is it also available in Dutch? Because that's a lot of the times a problem in the Belgian market.

Andreas De Neve: [00:05:00] Yes. So are model is available in English, Dutch and French. And we have we have German, Spanish, Italian coming soon. But for Belgian customers and most European customers, we're fine because we really target large organisations and multinationals and most of them operate in English.

Anne Vancoillie: [00:05:23] So, yeah, maybe Tom. A question for you. What is the link with SuccessFactors then? Because SuccessFactors is already a full HR, um, implementation?

Tom Mentens: [00:05:35] I think, as you already know, SuccessFactors is a cloud solution, which already brings the best practices in the world of HR - so bi-yearly via their release schedule, the solution evolves with new and better features. However, sometimes customers want to innovate a bit faster or more disruptive, and they want a new solution, such as the Resume Matching App. And why do they want to do it? They want to integrate with SuccessFactors from a process perspective. That means that it's actually from a user perspective, it actually works in SuccessFactors, not as a separate app. Or from the perspective of the user experience to have one user experience for all your apps that looks and feels the same like a SuccessFactors. But a third one, which is very important, as well as from a technical perspective - to avoid a very complex landscape with interfaces, to actually integrate this type of apps in SuccessFactors. That means that we can also benefit from SuccessFactors. And in this case, it's very important one, the fact that we already have quite a lot of data in SuccessFactors, such as job descriptions, vacancies learning data and so on. So that's something this app can leverage on.

Anne Vancoillie: [00:06:57] You spoke about one view for the end user, is this vote for the employee, as for the admin user?

Tom Mentens: [00:07:03] Exactly, exactly. And this is a good example, is that we built the app from a backend perspective. That means that a recruiter can already work with it. And so the job matching or the resume matching is done for the recruiter. So that means that the recruiter actually can use the solution to pre-filter already and which I find strong in the solution is that the fact that you have full transparency of the skills that were actually matched or the skills that are lacking for the candidate. So that's a very important one. And also for the candidate as such, that actually the candidate can upload their CV via chatbot, for instance, and then they can find the best suiting vacancy without losing his or her way in a complex job site. So that's an important one as well.

Anne Vancoillie: [00:07:54] And how did the collaboration with TechWolf start?

Tom Mentens: [00:08:00] Good one. Andreas, it's been quite a while now. I think we had mutual contact at SAP and at SAP they think innovation, they think Flexso. So they brought us together and actually I think we quite fast found a client that was interested in using the technology. And that's a very important one as well. If you have a case or a client that wants to co-create together, then it's more... Then it's easier to actually start the journey together. So that actually was the case here.

Andreas De Neve: [00:08:36] Yeah, I think it was a little over a year ago that someone from SAP directly linked us with Tom. Uh, and then we kind of took off really quick. (Anne: It seems that way.) Yeah. Yeah. Under a year, uh, since from getting to know each other till go live and the first client so... It's quite remarkable. Yeah.

Anne Vancoillie: [00:08:59] And what was the business value of this piece of technology? You already talked about this a bit, but what was the business value you had in mind when you started the project?

Tom Mentens: [00:09:10] Yes, we started the project at a client, which actually provides HR services on a global level. And the client has a clear strategic objective of being innovative and digital in everything they do, actually. That means also from a backend perspective or from their internal work, how they work internally, but also if they go to clients. Digital and innovative, that's the way to go. And they actually were looking to support this type of activities. They were looking for a lot of digital profiles to carry out this mission, and they put quite a lot of employer branding efforts into that. And of course, if you're looking for innovative people, your selection process should be innovative and efficient as well. That's how you attract the right people and that's why they actually got us on board to first to build the Resume Matching tool. First from a recruiter perspective, as I said, that means that a recruiter is supported in pre filtering candidates. That was the first thing. And the second thing is also to build it, in this case, a chatbot to interact with the candidate. And that CV matching is part of that. Concretely, as I said, it's the candidate can upload his/her CV and they get the best matching vacancies and they're quite a big organisation internationally. So it's quite interesting to do the matching without actually finding your way in a complex organisation on the job site. So that was actually the business case.

Anne Vancoillie: [00:10:49] And what was the biggest value of working with TechWolf?

Tom Mentens: [00:10:53] Well, I think one of the... Apart from working on and in an efficient way together. And I think it was quite fun to work together. Apart from that, it's also the capabilities that TechWolf has. If you build a skill engine, it's based on a lot, a lot, a lot of data. So that means that they have been analysing resumes, job descriptions for quite a long time now. I think millions of data sets...

Andreas De Neve: [00:11:23] We have 300 million vacancies and we analysed to train a model.

Tom Mentens: [00:11:28] And that's something that we as a technology implementer can't do. That's something that's very specific for TechWolf. And I think, if we take a look at innovation and I think this type of co-partnership is really the best thing to do. We share... They have capabilities that we don't have. And so we complement each other.

Anne Vancoillie: [00:11:49] When you say 300 million vacancies - is it a bot that does the analysis or is there someone scrolling through every...?

Andreas De Neve: [00:12:03] It's an algorithm that based on the sentences in those vacancies learns on a deep level how skills relate to each other. It through these vacancies a couple of times. It really knows, um, how Adobe in Design relates to Adobe Photoshop and how Java and C are two programming languages, and that if you have marketing skills and you have recruitment skills, then you would be a perfect digital recruiter. That's the kind of knowledge that it infuses. So, yeah.

Anne Vancoillie: [00:12:36] Okay. So it's a bit training the bot...

Andreas De Neve: [00:12:40] Yes, yes. But I think if you would call it a bot then Jeroen, our technical lead would be quite mad when he hears the podcast.

Anne Vancoillie: [00:12:51] The app. Yeah. So maybe for our listeners and for me then, why Can't we call the bot. We, we already touched these in the previous podcast but it's a long time ago. So maybe um because in my head it's a computer doing those things.

Tom Mentens: [00:13:12] But I think, if I can speak for myself here Andreas. I think what we see is that a bot, if you're referring to a bot or RPA, then it's actually more rule-based. There's no real intelligence behind it and it means you have to tell the bot what to do and the bot will do it. I think there's a bit more behind the skill engine than that.

Anne Vancoillie: [00:13:30] The Computer brain.

Andreas De Neve: [00:13:31] Yeah, I think. Bot, in my understanding is like just something that that does a repetitive task over and over again or does something superficial. Where the algorithm is like humans and computers, they both have their skills. For example, uh, computers they can solve crazy mathematical equations that would take humans years to process, but the same computer isn't able to have a swift conversation like the conversation that we're having, um, on the level of of a six year old child. And there are things that the computer is better at. There are things that the human is better at. And one of the things that that AI or algorithms are better at is understanding massive large volumes of data. So it's not bots. It's more of an advisory, an oracle kind of thing. Some mysterious creature that's like made sense of the data. And that can help you or get you going instead of some bot.

Anne Vancoillie: [00:14:39] I think it's the link with the chatbot that made me confused.

Tom Mentens: [00:14:47] But actually the chatbot in this case is actually just to the means of conversation, is just a channel to have a conversation with the candidate. There's not more than that actually. Did the skill engine behind it is something different. So I think we talked about chatbot before as well. It's just the way to make a conversation with a candidate and to do the necessary upload your CV and so on.

Anne Vancoillie: [00:15:13] But then the real intelligence comes after in the backend.

Tom Mentens: [00:15:17] Exactly. And that's something we can't - we can't do that. And actually that we built together with TechWolf, we built the frontend in SuccessFactors so that you actually have the look and feel of SuccessFactors. It's actually connected to the SuccessFactors processes behind it. So that's where we find each other.

Andreas De Neve: [00:15:40] Mm hmm. I think that summarises it quite well. We are the AI experts and Flexso are the SuccessFactors experts. So it's a very natural partnership that makes a lot of sense. I think it's a way of bringing the latest AI innovation technology to the market. But through one of the companies trusted partners instead of having to work with, with a lesser known AI. So I think that Flexso mitigates a lot of the risk that's associated with working with startups while still bringing today's newest technology to their clients.

Anne Vancoillie: [00:16:10] Yeah, my next question for you was, what was the biggest added value of Flexso but you already answered?

Andreas De Neve: [00:16:15] Yeah. I think looking at Flexso their deep expertise in SuccessFactors and their longstanding relationship with many customers, um, it makes sure that they know precisely what the customer needs and that they are able to articulate it very well and then they can use our, our API to craft a solution for their customers.

Anne Vancoillie: [00:16:36] Ok, so maybe Tom or Andreas. I am a customer of SuccessFactors and I want the same application because I heard about it in the podcast and I thought 'I want the same thing'. I'm also all about innovation. Which steps do I need to take? Can I contact SAP or directly Flexso or do I first need to contact TechWolf, I'm lost.

Tom Mentens: [00:17:03] So what's important is that we actually built the app in that way that we can actually plug it in into the setup of another customer. We see that the setup for recruitment it's probably for 80 or 90 percent the same for every customer. That means that we can actually duplicate the setup of the app for other clients as well. And that's what we planned on doing as from the beginning. And we want to implement as much as possible in an efficient way. So in that in that sense, the app will be available on the SAP App center very soon. And that means that we can... Clients can find it there and actually opt to use it in their current setup. Or of course, it's also possible to contact Flexso for people directly - will take a look at it together with possible clients.

Anne Vancoillie: [00:18:00] So as I understand correctly, you can implement it with every kind of setup of SuccessFactors, because you when you do a setup of SuccessFactors, you always take certain implementation decisions, but it's plug and play.

Tom Mentens: [00:18:16] Well, it's not that easy. I think it's not like in your app store from your iPhone, for instance, and you just choose to install it and it works. There's always some configuration to be done, but of course the setup is being done that way that we can easily, easily replicate it to other clients. That means that the implementation effort is quite small.

Anne Vancoillie: [00:18:40] Ok, interesting, and let's say I'm now another customer and I have a new innovative idea, who do I contact for a new co-creation?

Andreas De Neve: [00:18:56] I would say TechWolf. I think if you're in HR or it doesn't really matter in which field your in, but, um, you want to bring digital transformation to HR and you want to start working with skills, I would say, uh, drop us a line.

Tom Mentens: [00:19:13] Yeah. Same here for Flexso. If you already have a SuccessFactors setup in place then we're the way to go then and we can help you. Of course, and again, I think that the co-creation part of it is very important. If we have a business case which can easily be duplicated afterwards, then it's something we can talk about to co-create together and take part of the investment costs together.

Anne Vancoillie: [00:19:38] And Andreas, what are other HR technologies that you are working on?

[00:19:44] Yes. So, I said, the recruiter cockpit is just the tip of the iceberg. I think that we're right now we're looking at, uh, organisations already working with SuccessFactors, who really want to leverage this skill framework throughout multiple, uh, if not all of their HR operations. So this is a point solution for recruitment. Um, but we want all of the HR operations to work within this skill framework and that they share a common language, being skills. And then we can also combine these insights with, for example, uh, SAP analytics cloud, uh, um, people analytics is a field within HR - that's definitely on the rise. It's up and coming. And we have solutions for that, to really allow people analysts and business intelligence analysts to really have this new found dimension of skills, uh, that they can link to performance, that they can link to other internal data sources to really build some actionable insights.

Anne Vancoillie: [00:20:47] So it starts for now with recruitment within this app. But the skills, the skill theme has also links of course in all other HR operations and people analytics, I suppose.

Andreas De Neve: [00:21:02] Yeah. Um, just to give a short example, if we take a look at the S&P 500 - the 500 large tech companies, um, in 1985 and only 17 percent of the assets of those companies were intangible and so tangible assets made up, uh, 83 percent of total valuation of those companies. If you take a look at today, we totally switched that around. So instead of 17 percent intangible assets, 84 percent of the assets of the S&P 500 are intangible. And we see that HR directors, CHROs, more often they're getting seats on board level - there in the board of directors. And we see HR is transitioning from a purely operational role towards a more strategic role. And because they manage the human capital of the company and human capital of the company is getting more and more important over time. So people analytics and everything related to HR analytics, business intelligence is, or will become much more important as time progresses, because it is a way for HR to really become that strategic partner that companies need HR to be.

Anne Vancoillie: [00:22:13] Maybe a question there. If a company has their own skill framework, um, can it be implemented in your application or do you use a market broad skillset?

Andreas De Neve: [00:22:27] It's a very good question. Yes. We can use the skill framework of the organisation. And it's usually a good thing, um, because people in HR, they spend quite some time developing - It's a work in progress. They developed their own skill or competency framework over time. And we can link their skills and competencies, that are very tailored to the specific use case, to the specific organisation at hand. Um, we can link those to a set of skills beneath the surface. You can consider skills to be the atoms, and the competencies and the roles that HR. defines to be the molecules. So different skills in our framework, build up to one competence or skill in the company's framework. Our solution is totally beneath the surface. And so we can reason with skills, we can act on skills we will track skills in our framework. But when we communicate with the company's HR, we are able to talk their language. And it's very important. I think, that on a very basic level, what we do today is instead of HR, instead of their spreadsheets and them trying to catch up and fill out what everyone's skills are, it's automating that process. And we fill in their spreadsheets for them, so they can stop wasting their time, uh, trying to track the skills or asking everyone to fill in your skills. No one really does that. People rather file a tax return then fill in their skills on a monthly basis. We do that part for them and allow them to take a look at the future again.

Anne Vancoillie: [00:24:06] So it's actually making the same language that companies sometimes already have or are building up throughout their HR - implementing without the heavy administrative task that are linked with a deep competence model. Because you you have a lot of organisation that starts with really role specific skills and then implement it and want also to link their training and development and stuff, to that skill set. But then after a year or two, it fades away because it's too much work.

Andreas De Neve: [00:24:56] Exactly. And that's the work that we try to automate. Um, as you said, it's very important that these companies are able to use their own custom, very deep competency model, because we have a client which, what we call a database administrator, at TechWolf the database administrator is an engineer. Um. In this case for the customer, database administrator is not a technical rule, so people have - the companies, they have their own language. A business analyst at the bank, might have an entirely different skill set than a business analyst at a telco company (of course). So it's important that we enable our model to really empower their competency model instead of trying to replace it.

Anne Vancoillie: [00:25:48] Yeah. And maybe, Tom, what is the link with strategic workforce planning or other skill management in general - seen in SuccessFactors or in general in HR transformation projects?

Tom Mentens: [00:26:04] Well, I think a lot of HR directors are talking about strategic workforce planning, but it's very difficult to actually go a bit further than the concept of a strategic workforce plan. And I think, the skill engine can really support strategic workforce planning. And I think you often read or hear about organisations that are preparing for their digital transformation - which actually means that they probably are looking for skills of the future and they need to shift towards those skills of the future. And I think via the skill engine, you can get a view on the skills of the employees. I think we have quite a lot of data sources. Some are ... one of them is SuccessFactors, for instance. So there's quite a lot of data in a SuccessFactors. You have their resume, you know which projects people are working on, you know what they post on Yammer, for instance, or the internal social media. So in that way, we got quite a lot of data about employees. And I think via the skill engine, you can get quite a good view on the skills that people have now. If you then defined there are also jobs of the future, you can easily define the skill gaps on the organisational level, to actually guide or to support your training plans and so on. But also on an individual level. And then you actually go to some type of digital job coach for any individual where the skill engine can actually guide you to the best solution to bridge that skill gap, via training or via networking to find people with similar skills and so on. So I think workforce planning is one of the better practical uses of all of the skill engine.

Anne Vancoillie: [00:27:52] When you talk about the skills of the future and roles of the future, I imagine that's about the strategy of a company or is it something general?

Tom Mentens: [00:28:04] Good question. I think it's of course, as an organisation, you have to define your skills of the future. But I think if you take a look at the data that TechWolf has, they can already see in the big data sets they have, they can already see which evolutions there are in rules. And that could be used as a basis as well. But still, you as an organisation, you have to define what skills you will need, how those skills will look like - That's something that TechWolf can help you with.

Andreas De Neve: [00:28:34] Mm hmm. Yeah, I second that. I think it's obviously up to the organisation to know the strategic direction they're headed. But the skill engine can definitely provide some insights. And, for example, um, the energy sector in Southeast Asia and Japan is, I would say, three or four years ahead of the energy sector in Belgium. So if you want to do workforce planning for a Belgian energy company, then it just makes a lot of sense to take a look at the skills that the energy workers have in Asia. As it this is a very good predictor of the skills that Belgian energy workers will have three or four years from now. So if we can get those people to learn those skills faster than they would've, then it's a very good application of using the skill engine and its market knowledge, um, to make certain decisions.

Anne Vancoillie: [00:29:35] So it's not only the technology, but also the market knowledge and the in-depth knowledge of skills in general within the market.

Andreas De Neve: [00:29:45] Yes, I think that's a very important part of strategic workforce planning. Um, is comparing your own future roadmap as a company with market data and where is the market heading to? Are we ahead of the market? Are we lagging behind? Are we closing the gap? Is the gap getting bigger? Um, if you want to assess which skills you should acquire through recruitment, through hiring, which skills you can acquire through re- and upskilling, it's also very important to know what will the supply and demand look like in the market for certain skills and certain trending skills. Um, there might be shortage two or three years from now, so we should try to develop those skills in-house instead of waiting two years and then try to fish in an empty pond. So the market is obviously very important to take into account.

Anne Vancoillie: [00:30:38] And now you took an example of the energy sector. But, um, do you have knowledge of a lot of different sectors or is it mainly focused on specific sectors where you already had clients in the past or something?

Andreas De Neve: [00:30:56] And we work in a lot of sectors. I took the example of the energy sector because we have a large project, um, with, InnoEnergy as a partner. It's also funded by, uh, partly by the European Commission to reskill workers in the energy sector in Europe, uh, towards the renewables, towards batteries, etc.. So that's why I took the example. Um, we mainly focus on sectors with a lot of white and green collar employees, um, and for example, telco, banking, financial, financial sector, uh, fast moving consumer goods. Um, so there are a multitude of sectors that we that we target. Uh, so it's not not purely energy.

Anne Vancoillie: [00:31:38] Okay, so if I'm a customer, there's a big chance, uh, you have some data about my market and otherwise it's always there's always a possibility to look into this.

Andreas De Neve: [00:31:49] Exactly.

Anne Vancoillie: [00:31:50] Okay, perfect. I asked all my questions to you guys, uh, is there something extra you want to, uh, say on this podcast, a last tip?

Tom Mentens: [00:32:04] Again, I think we are teaming, uh, well together. And I think, uh, we co-created well together and I think we are looking forward to do that a bit more. So, uh, always looking forward to a good business case. And then we are very, very curious to dive into that.

Andreas De Neve: [00:32:21] Yeah. Yeah, exactly.

Anne Vancoillie: [00:32:23] Yeah. You said in the beginning it was a lot of fun working together. So, um, I think if people having fun doing their job, it's always, uh, leading to a fantastic result.

Tom Mentens: [00:32:36] But I think in the market of technology or HR technology, that's the way to go - to innovate, uh, and co-create together and think that in such partnerships that it goes faster and goes better. Yeah.

Anne Vancoillie: [00:32:49] Ok, thank you very much for being in this podcast and hope to see you again.