Disclaimer: The Futurum Tech Webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.

Transcript:

Keith Kirkpatrick: Hello, and welcome to another episode of the Futurum Tech Webcast. I’m your host, Keith Kirkpatrick, Research Director for Enterprise Applications with The Futurum Group. Today we’re going to be addressing a really interesting topic focused on how analytics and AI can be applied to human resources data to improve the work environment. We’re also going to dive into a discussion on how AI is changing the nature of work, and what that really means for employees and businesses. Of course, no discussion about AI and analytics would be complete without addressing the need for transparency around its use, particularly when AI is used to aid in decision-making.

So to help illuminate those issues, I’m pleased to welcome two leaders from Ahgora, Thiago Quadros, CTO, and Angelo Frigeri, Data Team Lead. Ahgora is a Brazil based SaaS provider and an HR technology pioneer in cloud computing. The company is a leader in time and attendance tracking, leveraging facial recognition that is validated by AI. Ahgora also offers access control and monitoring, task management, rostering management, vacation management, and corporate education. Welcome, Thiago and Angelo.

Angelo Frigeri: Thank you.

Thiago Quadros: Thank you. It’s a pleasure to be here.

Keith Kirkpatrick: Great. Well, why don’t we just get right into it? So maybe we can start out by just addressing what kind of data is actually available within the human resources department, and how can that actually be accessed and leveraged to make HR more strategic?

Angelo Frigeri: Well, Keith, There is a number of data that we can work with, from bunch data to over time, absence, signal infractions, the list goes on. And one relevant data that presents a real challenge to HR today is attrition, because losing man power is time costly, it’s financially expensive. So how could HR personnel and managers work together to mitigate that challenge? Well, data can help them. And how do they get this data? They do so by having a solid portfolio of products that can help. So data work with us in this way. So for example, one very effective methodology the managers can keep track of their team’s health and so on, is by one-on-one meetings. And another thing that is very important, a product that we proudly released this year is Mood. Where employees they can out to evaluate their mood or their happiness in a scale so to say. So in one side, for example, in these one-on-one meetings, managers, they can register when those meetings happened, who participated in the meeting, or if that was a highlight meeting, something very important happened. And by Mood, the managers, they can see how their employees are doing regarding their happiness. So employees, they can say for example, oh, I’m happy because of my work, or I’m neutral because of my financial situation. Or they can say that they are sad because of their family. Something happened in all those cases. Having this working together, managers from one side, they can see what is happening through Mood. They can act, they can intervene through those meetings and they can follow up by following how Mood is happening after the action that they took. And AI can play an important role in that because AI make this data abstract. So I’m not only presented raw data, I’m presenting rich data, which is relevant for HR. It’s easy for them to understand, easy for them to take actions so they spend less and less time working in spreadsheets because we deliver just what they need in the format that they can understand. So they become more efficient, more strategic, they have time for what’s more important.

Keith Kirkpatrick: Right. That makes a lot of sense. But I guess one thing I’m wondering about is if you could talk a little bit more about the elements and the processes that are really required to build robust and useful data products within HR. What’s the data that we need to be really working with and what can you achieve with that? And then again, I know you alluded to this before, how does AI enable that?

Angelo Frigeri: We don’t believe that a data product should only present data. That’s not what we believe. We believe that more is involved. We strongly believe that product that is a data product is effective when users, they can easily understand data. And also reliability is important. The data they should trust in what they’re seeing. So saying this, one thing that is very important for user is the data they are seeing needs to in one side answer important questions. And to the other side, it needs to be easily translated into an action. If nothing of both of these is accomplished, then maybe it shouldn’t be that in the first place. And we also understand that different users can take different course of actions.

For example, a C level user will take a more strategic one. HR users might have more organizational actions toward their managers and the levels downwards. While managers, which are also users, they can make a more closer contact to the person, to the employee himself since they’re closers, as I said. So when we deliver value through data, we need to think who’s going to see it and who’s being helped with. And I need to deliver data that is compatible to the bound of actions that user can take. And AI fits into that in at least two different ways, by creating user-oriented AI models. So different levels of users require different levels of abstraction as well. So adapting AI to generate just what they need is crucial. And also by studying how we can have our product improved by tracking user data information from our product. This ensures that our users have the best experience possible. So in conclusion, building data products require two visions. Who am I presenting data to, and also the quality of my data, and it must be translated into action. One way AI ensures that they have a good experience on both ends, delivering and measuring.

Keith Kirkpatrick: Thiago, I’m wondering if you have any thoughts on this, particularly if you’re looking at data for let’s say employee behavior.

Thiago Quadros: Thank you. Thank you for including me in this topic because I really have some thoughts about that. I strongly believe that HR data, people data, employee data is really, really rich. Angelo started to grasp a few ideas related to punches, to vacations, to one-on-one records and Mood, which is something that we have here. But think of the vast complex portfolio of HR products that are out there. There’s so much things that you can look into that in search of data from people, behavior data from people like appraisals, like performance reviews, like OKRs and feedbacks and results. And imagine how much you can investigate this data trying to find correlations and to predict or evaluate future behavior and their outcomes. And that is data is all about. You’re trying to make reality a little bit more objective since people relations are so subjective. So when you are able to do that by inspecting and investigating vast amount of data of many HR products, you can come with really, really, really interesting solutions and to be able to predict behavior, attrition as we do here, and many other outcomes that we can assess through the correct use of data and AI algorithms.

Keith Kirkpatrick: Right. Well, what’s really interesting is we hear a lot about AI and obviously there is a lot of discussion about AI potentially changing the way people will work. A couple of big picture questions. First of all, is AI going to replace jobs or is it going to be used to help augment people in their workday? And then of course, how do we make sure that employees don’t use AI simply to take shortcuts? We’ve all seen the case study of people going out to Chat GPT, typing in something, and then copying and pasting whatever comes out of there and claiming it as their own. Maybe you can talk a little bit about that.

Thiago Quadros: Very good question, Keith. I really think that AI is going to change work, not necessarily replace jobs. As internet was a few decades ago, AI is going to change a lot of things, especially in office work. Because you will interact with computers and devices in a different way. You will have an intermediary in many situations, which is a companion, which can be a great companion to automatize tasks, to augment your decision making, to enhance your productivity since it can do many things for you. So in terms of professionals, I really think that people should be really interested in helping how to use AI tools effectively since it can really, really make a great boost for their productivity and outcomes and results. So there are also other ways that AI can help people to enhance their creativity. Since you can ask AI to generate many content for you and you can now judge instead of producing those contents, you can judge and interfere with pre-made work done by AI. But I really think that the human factor is going to be decisive in how companies are going to generate the results and the outcomes based on what the AI is pre-made for them. You always, it’s hard to say always in a search chart, in a search speed.

Keith Kirkpatrick: Generally speaking.

Thiago Quadros: Yeah, generally speaking. But you still have people judging what is being constructed and being revealed by other people and by AI. It’s really hard to see AI leading other people and to see the ramifications of that. So I really don’t think that… AI is going to replace a few jobs, I think. Especially jobs that are completely automatized are going to disappear from time to time, like internet did with many jobs years ago. And so I really think that this is going to happen, but many new positions will arise from this change. So this is going to be really interesting to see in the near future. But you made another question that was really interesting. How can we prevent that employees does not use AI to, I don’t know, to, I believe they used the term shortcut? So to just serve the work and asking the AI for do its job for you. As I said, it is going to be apparent that someone is doing that. Now today you can do that. You can evaluate someone’s work and see this is not this exactly a person’s job. And you will be able to judge the quality of their work. I really believe that the human is going to have a role in judging what is being produced by employees. And I especially believe, too, that AI is going to be a companion to evaluate what the work has been done is produced by a human or another AI. So we are going to have tools to evaluate that too. And in terms of especially creativity in roles that we need creativity to perform well, I believe that this is particularly important to have someone who is evaluating the results, the work that we’ve been produced to check if it is good or not.

Keith Kirkpatrick: Right. Well, yeah, I think you’re right there. There is a lot of… There is certainly potential for people to abuse, but there are also tools to be used to make sure that people aren’t abusing it. And of course, as time goes on, I think people will realize that you are able to ascertain when a human is applying their own thought and their own reasoning to something instead of just copying and paste. So certainly some interesting comments there. I want to quickly just shift a little bit and talk a little bit about how AI can actually be used in the HR discipline. Particularly I think one really interesting use case is perhaps looking at how can AI algorithm be used to predict when someone might be at risk for resigning their job? I was wondering if somebody could talk about that.

Angelo Frigeri: Yeah. We recently released an AI model that predicts attrition. What happened is since we released this, and we released this in-house in first place. And we changed even how our own HR personnel and managers they approach those potential risks of resigning. Before this AI model, our HR personnel had to intervene after resignation really happened. But now with the model, so they already have the answer to the potential attrition risks. So it’s not only corrective work anymore, now it’s preventive. So they’re going to work preventing problems, rather them only fixing them afterwards. And the actual now is towards retention. And if a resignation really happens, some of the cases we cannot prevent, they have further time to plan ahead a replacement. This improves the health of HR personnel. Talking about how AI change how people work because they deal with a lot of work every day. This improves the company numbers such as turnover, operational efficiency, improves their financial numbers. Managers have more reliable and stable teams. Well, benefits are huge.

Keith Kirkpatrick: Absolutely. I think, just to make sure I understand, it seems like what the algorithm might be doing is intaking various signals. Let’s say an employee is suddenly starting to come in late every couple of days, or perhaps the quality of their work is not up to what it be. Or maybe they’re even saying things in one-on-ones saying that they’re unhappy. I would assume that all of those can be fed into a model which then can say there’s propensity to perhaps disengage. And then as you get further along in that disengagement scale, that’s when you’re likely to see some attrition. Is that pretty much what’s going on with the model?

Angelo Frigeri: Exactly. So we have this behavioral data. We have also demographic data, thermographic data. We have exogenous data. So all of these when are crossed together and they’re analyzed, they provide for us two outputs. So we have one output with where employees that effectively resigned and employees that didn’t. So with those two examples, distinct examples, we can teach our AI to, based on a behavior that is happening right now, which of those outcomes is more likely to happen?

Keith Kirkpatrick: Right, right. Okay. Now one of the things though that’s a little bit, I don’t want to say controversial, but certainly something that organizations need to keep in mind is the issue of how do we make sure that this AI is transparent in terms of what is actually going on? What’s actually flowing into these algorithms? Because certainly someone, if you’re an employee, all of a sudden you get called in to your manager’s office and they say, “Hey, it looks like you’re going to quit.” And they go, “Well, how did you know that?” Maybe you could talk me through this whole idea of transparency and making sure that AI is in fact used for good.

Thiago Quadros: I think that you need to consider three pillars regarding this answer. You have to approach this problem by three different ways. One is a technical way, the other one is an organizational way, and the third one is regulatory way. And governments have a measurable in this area. And the concern about this is really big. You have Elon Musk a few months ago saying that we should stop developing new AI until we figure this out because the risks are so high. I think that if we can put ethical considerations into the design of the AI tools in algorithms, since the first phase, since its conception, I think we can improve that and we can make clear, transparently clear what is being done with each piece of data that is fitted into these algorithms.

The user and the companies that are using these products need to know exactly what is being done. And you must have a rigid regulation systems that punishes companies that doesn’t comply or doesn’t use their user’s data explicitly how they said it’s going to be used. So I think these three aspects can make a really, really huge difference. But transparency I think is the most important of them. But we need to consider too, that many, many people, how many times have we, the four of us here, the three of us here have to have already read a terms of use regarding to user data. We need to be more concerned about this and make users more concerned about this because this is what is valued today. We know that AI can only be built using data, and data is a rich thing. It’s a really rich thing. And if we make people more conscious about what they give permissions for a company to use, and as a society as a whole, we need to make products that raise this level of consciousness so people can make better choices and start to read in those terms. And agreeing or disagreeing in policy companies about the use of data. I think this is really important.

Keith Kirkpatrick: Obviously by adhering to those principles, you’re really making sure that you’re using that for good rather than harm. But for my last question I want to ask, so how do you actually bring AI into the human resources discipline in terms of getting buy-in? And I don’t mean just buy-in in terms of the people working in HR, but of course the employees that are going to be impacted by its use. What are some of the strategies or tactics that should be used to make sure that everybody is on the same page?

Thiago Quadros: It is really hard to build tools knowing exactly what is going to be made of them. So we build things mainly based on principles here. And we really believe that our product should help people grow. So when we say a Ahgora, the H between A and G is for human. So we have a human inserted right in the middle of our name because we think that the human factor is key to make companies grow. And as companies grow, people should grow too, not only financially, but they grow in their soft skills, in their relationships, in a way that the work must fulfill a purpose in their lives and allow them to build better lives for themselves and their families and their communities. So I think that if you focus your products considering this, that you’re trying to make an impact in people’s lives and the responsibility that you are doing that, I think this is really important. And you must understand other companies that use AI based on people data should make this explicit what they want with it, what they are trying to achieve. Of course, money is always going to be important for companies. But what else are you trying to do? How are you going to bring those results to your company? And I really think that this really drives the consumers and clients towards your company. It’s what you’re trying to build and what you’re trying to provide to the society as a whole.

Keith Kirkpatrick: And Angelo, just to follow up on that, it would seem that being able to really communicate the benefit, not just talking about the technology, but the benefit to the employee saying, if we’re using this, we can be more fair, we can be more encompassing and really inclusive with everything because we’re actually building off of data rather than just what somebody thinks or what somebody feels.

Angelo Frigeri: Yeah. AI can helps a lot in that because people also have bias and models will eventually makes mistakes of course. But when we have AI working along with people, I believe that this can be much improved. Bias can be resulted to the minimal because of course, people who interpret the data or interpret a forecast of a model, they can also, presuming that they know bias could be presented in AI model as well, they could prevent this from being extended into an action. So this is a safe course of action. They can see what is happening and in one way or another go along with it or just be a critic and say, no, it’s better not going this way. This may be an outlier, this may be a mistake. And at the same time, AI modeling in HR or in high level managers, they can also help to see how if employees are being fairly treated in the levels below.

Keith Kirkpatrick: It sounds like the combination of having all of this data and having the models and being transparent, it really makes sure… It essentially keeps everybody honest to make sure that things are working in the way that they should be working because you have that enhanced transparency and visibility.

Thiago Quadros: If I may add to that. As I said a little bit before, you must have that human data, that human factor.

Keith Kirkpatrick: Absolutely.

Thiago Quadros: You need someone to evaluate that. And AI and analytics and data products can enhance people’s awareness and people must grow to understand that. The ability to understand and to see vast amount of data, different aspects of the reality that are presented through data, they are going to be able to make better decisions. So this is going to be really important.

Keith Kirkpatrick: Great. All right. Well Thiago, Angelo, I really want to thank you both for joining me today. This has been very insightful and enlightening. So, really glad to have you here and we hope to see you again in the future.

Thiago Quadros: Thank you very much.

Angelo Frigeri: Thank you for having us.

Keith Kirkpatrick: All right, everybody hit that subscribe button. Join us here for all of our episodes of the Futurum Tech Webcast and our interview series with insightful leaders from across the technology spectrum. Thanks, and we’ll see you again really soon.