Hello everyone! My name is Ladek and my guest for this episode is Pablo Ferreiro, who joins me in his role as Head of B2B Sales for the Americas for the company ELSA.
Pablo is also a co-founder of Coderhouse, which is backed by Y Combinator, and the Founder & CEO of Nord Learn. Oh, and in his spare time he has also run across the Andes Mountains.
In this ‘cagematch’, Pablo and I talk about
00:00 › Start
3:00 › ELSA’s First Words—Pablo discusses the surprisingly lengthy history of Elsa Speaks and what their focus is today
5:16 › DelegAIt—Pablo talks about how AI is changing the space of education and learning and offers his opinion about which tasks AI can handle better than humans
13:58 › Machine Vs Machine—Pablo discusses the potential issues of “AI talking to AI” with the advent of publicly available Generative AI products
21:14 › All About Ethics—We move into the topic of ethics and Pablo’s views on the importance of this with more than 53,000,000 users a day.
27:31 › Enterprise AI Selling Challenges—Pablo offers the challenges he receives in an enterprise sales process about using AI in recruiting and upskilling
31:45 › Where’s L&D Heading—Where does Pablo see things going over the next few years on that interface between the L&D department and the HR department using a tool like Elsa, and how much is the line blurring between working with humans and working with AI?
This is the eLearn podcast. If you’re passionate about the future of learning, you’re in the right place. The expert guests on this show provide insights into the latest strategies, practices,
and technologies for creating killer online learning outcomes. My name’s Ladic, and I’m your host from OpenLMS. The eLearn podcast is sponsored by eLearn Magazine,
your go -to resource for all things online learning. Click -by -click how -to articles, the latest in edtech, spotlight on successful outcomes, and trends in the marketplace. Subscribe today and never miss a post at elearnmagazine .com And OpenLMS,
a company leveraging open -source software to deliver effective, customized, and engaging learning experiences for schools, universities, companies, and governments around the world since 2005.
Learn more at OpenLMS .net Hello there, my name’s Ladic, and my guest for this episode is Pablo Ferreiro, who joins me in his role as head of B2B sales for the Americas for the company Elsa.
Pablo is also a co -founder of Coder House, which is backed by Y Combinator, and the founder and CEO of Nordler. And in his spare time,
he’s also run across the Andes Mountains in this cage match -like conversation. Pablo and I talk about the surprisingly lengthy history of Elsa Speaks,
and I mean that in the best way, and what their focus is today and how they serve their customers. We also talk about how AI is changing the space of education and learning, and Pablo offers his opinion about which tasks AI can handle better than humans.
Pablo then discusses the potential issues of AI, talking with AI, with the advent of publicly available generative AI products. And then we move into the topic of ethics,
and Pablo’s views on the importance of this with something like more than 53 million users a day on the Elsa platform. Pablo then offers the challenges he receives in an enterprise sales process about using AI in recruiting and upskilling,
and what can be done about those challenges. And then we tie our conversation in a bow at the end, as we usually do, where asking Pablo about, you know, where does he see things going over the next few years in the interface between the L &D department and the HR department using a tool like ELSA and how much is that line blurring between working with humans and working with AI.
And remember, re -record this podcast live so that we can interact with you, our listeners, in real time. So if you’d like to join the fun every week on LinkedIn, on Facebook, on YouTube, just come over to elearnmagazine .com and hit subscribe.
Now I give you Pablo Ferreiro. Hello, everyone. Welcome to the E -Learn podcast. My name is Ladik as you just heard and I’m here with Pablo, whose last name, whose last,
I’m just going to just destroy it. Ferreiro? The close Ferreiro. Ferreiro. It’s a hard one. As a good host, I should have known that,
I should have known that Ferreiro. And then I just realized right before we started, I’m like, I’ve never asked you how you pronounce your family name.
With Elsa, or as some people find, it’s ElsaSpeaks .io or there’s a couple of different ways it pops up on the interwebs. As we always do here, I’m really interested because I have couched this conversation.
I have defined this conversation as a cage match between AI and human tasks. Which ones can, what can AI do better?
What can humans still do better? But, before we get into that, tell us about who you are. Give us the 30, 60 seconds on who you are and the company that you represent. Sure. Well,
thanks so much for giving me the space, super excited to be here with you, to chat, even though it’s a one -on -one, but there’s people out there listening in,
so super excited to be here. Yes, I’m Pablo, right now I’m leading Elsa for what we call the America’s region. so anything from, you know, the US to Brazil.
That’s me. I help Both organizations in the corporate world, but also in the education sector Bring our technology our products to help their workers or their learners improve their communication skills And so that’s really what I do every day And so yeah,
super happy to be here. Super cool. And you know, just so people understand your expertise and Kind of where you come from give us a background on ELSA like how I know that I know the story But I’d love to hear it from your words about,
you know, who started it how it began and then kind of what You know, if I show up at ELSA speaks that I know what am I gonna experience? Sure, I think the the story is Vue the founder,
you know, she originally from Vietnam she went to school and I stand for it obviously Great school, but she realized,
you know, right after she graduated and started working and You know, large large company that she was sort of missing out on certain opportunities due to her her accent due to her Perhaps inability to communicate as fluidly as maybe is someone who’s you know from,
you know, speaks English natively so Out of that pain out of that frustration She thought okay, I gotta I gotta I gotta do something about this and so she went around the world to find a an AI expert and she she found him in Portugal of all places and You know Found the right expert and Xavier it’s his name and You know,
he’s got a hundred peer review papers on on AI linguistics and so together they they built ELSA It really first started as a B2C, right my play so just for consumers So if you go to our app,
you can download it in any app store or Google Play. And download it and run an test band on your English level and then do a couple of exercises to improve your communication skills.
So that’s sort of the background story on that. But yeah, here we are nine years later. There’s been a long journey. And I’ve been just part of it recently about more than a year ago.
But yeah, I’m still very, very fresh. Nice. And so we’re here to talk about, again, that cage match between there is a lot of excitement,
still a lot of consternation, a lot of anxiety around how AI is changing the space of education and learning for the good,
for the bad, for the ugly, for the profound. So I wanted to start right off the top there. So what tasks, in your opinion, can AI handle better than humans?
And when we’re talking about AI, we’re talking about, I assume, large language models. Or is there a different way that we should look at this? Or is it a different definition that you’re putting AI under?
Sure. I think that that definition is OK. What I would say is we are an AI company. But we’re still learning ourselves what the best use of our own technology is,
right? Where every day our product team is figuring that out. What assignments should a student be running? How can we help teachers be more efficient in the way they do lesson planning?
So even as an AI company, we’re still learning what AI is is and can do. I can give you maybe examples on real examples of how we found ways,
both in the corporate world and the education sector, how AI can do a better job than human and where the human still can do a better job than AI.
So if we look at the corporate sector or the corporate world, we help companies run English assessments, candidates,
applying for jobs. So we have a client that has a million candidates applying for their jobs every year. A million candidates. It’s pretty, yeah, it’s a big volume,
right? Oh, it smokes. That is, I just want to let that simmer out there for just a second. Like how do I, like I’m overwhelmed when I have like six people who apply for a job, like a million candidates.
Okay, sure. It’s a little too much. And yeah, all these candidates are applying to jobs that require fluent English. So,
you know, they of course need to make sure that they have that fluency. Before they started working with us, what they did was have their recruiters basically grade the English level of these candidates.
So you can imagine how slow that process would be. Not only slow, but very variable, right? Like, cause, you know, a human is going to have some rubric,
but then, you know, depending upon the person and the listener and the situation, it’s going to, that you’re gonna have to get some very, some considerable variability in those scores. Hi there.
I’m sorry to break into the show right now, but if you’re enjoying this show, if you are challenged, if you’re inspired, if you’re learning something, if you think that you’re gonna be able to get something out of this to put into your practice, do me a quick favor.
Pause right now and just hit subscribe on your podcast player right now. It doesn’t matter which one, just hit subscribe, because that way it’ll make sure that you never miss an episode in the future. Thanks. Now,
back to the show. A hundred percent, a hundred percent, you know, we, we exactly were, you know, humans are not consistent because we have different biases. We’re,
you know, we have different skill sets. So, so a hundred percent, that’s a fact. And, you know, the issue with that is that you might be biased towards something good or something bad.
So maybe you assume that the English level is not up to par when maybe it is and you’re missing out on an important candidate, right, or a potential employee.
So what we did is we came in and we streamlined that process where essentially a candidate goes up to their career site, completes their application and sends a very small audio sample of,
you know, why do you want to join our company in English? They respond and then we analyze that and we’re able to understand, you know, what their English level is at the pronunciation level, at the intonation level,
fluency and a lot of different metrics. Is that, is that, for in that example, are they given a paragraph to read or is it extemporaneous? They, they’re basically making up their answer on the fly.
They’re actually making up their answer, which is a good thing. You know, we want to be able to replicate normal conversation, right? So if I’m a recruiter, I’m asking, hey, why do you want to join our company?
And that’s, that’s what it is. In some cases, some companies just record the call and then give, get the feedback from the candidate right after that call. So they can run the test live in a conversation with the recruiter.
So, so yeah, you can imagine how this can really help streamline a a recruiting process, reuse costs and even recognize some talent that perhaps before wasn’t recognized due to some biases from humans,
let’s say. So I think that’s a pretty non kind of a no brainer example of why AI should be part of it. Now,
on the other hand, if we look at education, while there are some challenges when it comes to having the teacher provide feedback in terms of communication skills in terms of pronunciation to the students,
we can do a good job there, but the teacher still plays a key role for many reasons, one of them being the interpretation of the feedback.
So AI, if you’ve tried to achieve, for example, when you ask them a question, they can be very, they just give you a straight answer,
right? So sometimes you need to be able to interpret that information. So if our AI is telling the student that your pronunciation is 50 % instead of 100%,
the kid might think, well, shouldn’t it be 100 %? Yeah, what does that mean exactly? What does that mean? Exactly, right? It’s like, hey, I got my schwa sound is wrong. It’s like, what is schwa sound?
It’s a phoneme, right? Like who know, I don’t know. Yeah, so the teacher needs to be there. Honestly, I don’t even know what a schwa sound is. I don’t need a speaker of like, when do I say schwa? Ever. Who knows?
Exactly, it’s complicated, right? Even, so in this case, you need the teacher to be there to help interpret that data and not frustrate the student into thinking,
hey, maybe I should get 100%. It’s like, no, it’s, you know what? You’re within your level, right? Maybe you’re an intermediate English spirit. That’s fine. That’s the one way that it’s super important for the teacher to be involved to help interpret the data.
And then obviously, to facilitate the use of the technology for the best use cases. So, you know, AI is not supposed to be used the whole time, right? There’s good times to use it and there’s times that might not make sense to use it.
So the teacher really has to become a facilitator, a technology expert, so that it can really help the students improve their communication skills.
And in another sense, they also need to become sort of data driven, right? They need to look at the data. They need to understand what students are learning,
what they’re having trouble with. And so, I think that the role is changing a little bit. I think the fundamentals are still there in terms of what the value of the teacher is.
But I think the teacher now has more tools to be more powerful, to be more of a super teacher, to, you know, be able to understand their students better.
So I think that’s, in my point of view, obviously, very biased. Yeah, of course. You’ve just struck fear,
I’m sure, into the heart of so many teachers around the world when you’re like, Hey, guess what? Now you’ve got to be data -driven and you actually, you know, the technology and, you know, you need to be technically savvy and whatnot, which,
you know, quite frankly, this is one of these pieces we’re having to grapple with now that we’re moving into what many people are calling the era of AI, the age of AI, right? Because of the generative products that are now,
have been exploding under the scene for the last year. I’m very interested to see how that plays out over the next 10 years, right? I think we’re going to have a very different universe. Let me, let me dial back to something that you,
it’s not a challenge, but I’m wondering if you’ve had this experience or any of your clients have had this experience. Because generative AI products are available at the retail level,
they’re available to every consumer. And they’re pretty darn powerful. Have you run up into instances, let’s just say in your first test with your client,
there does a million applicants there, where people have shown up to the test and have either used an emulator,
or used AI to basically talk to the AI, right? Where they’ve come and said, rather than, okay, clicking on their microphone, they come prepped with that paragraph that is pretty well thought out through an AI model.
And basically you start having AI talk to AI, have you started to experience that at all? So in a sense, maybe a candidate, trying to use a cheerleader. Exactly,
I can go to, there’s plenty of services out there where I can go and I can say a paragraph, and then the AI model will take my voice. And then I can then just sort of, hey, here’s a beautiful paragraph,
give me the audio for that. And then I’m gonna, I’m gonna play that in. I’m just wondering, the only reason I ask that is because we’ve seen in our lives, my life, my wife’s life, whatever, we’ve already started to see CVs,
resumes, becoming absolutely perfect, right? And the answers to typical interview questions becoming just spot on, and interview processes around the world are changing,
as you and I are talking right now, simply because these candidates don’t actually have those skills, or they’re not actually those people, you know. And so I don’t wanna say people are lying,
but they’re definitely able to present a much, much nicer picture. So I’m wondering if that happens as well in this kind of testing model. That’s interesting. I think from,
look, I think in a sense, maybe maybe a strong word would be fraud, right? Oh, I mean, it’s total cheating. It’s absolutely 100 % fraud.
Like if I were to roll in and have a pre -recorded, you know, that’s complete falsehood. Let’s just call it that, yeah, absolutely. No, no, it’s, I’ve heard from one of our clients where a candidate submitted an audio sample and it turned out it wasn’t his voice.
It was a friend of his who maybe had a better English. And so I think, you know, we do have today, right?
You know, fraud prevention and ways to sort of mitigate that. But certainly, you know, that’s not going to stop, right? I mean, I think there’s always been some sort of,
you know, people out there that wanting to sort of gain the system, let’s say, let’s call it, so for sure. But I think the, go ahead. Good, no, no, no, no, go ahead. I just think it’s really interesting because we stumble into this moment where at the end of the day,
we’ve just proven the value of the human, right? In that, you know, if I can put together any process, you know, you’re going to submit a CV, you’re going to submit a paper for a test at a university,
you’re going to work with somebody, you know, in real. As soon as we put the actual humans in the moment though, that’s when, you know, that’s when the reality becomes apparent.
And so if you were to create that beautiful CD or you were to create that beautiful, you know, submit that audio that’s fraud, as soon as you and I start having this conversation, I’m like, oh,
wait, no, that wasn’t you, right? Or maybe I can ask one or two questions and I can realize, oh, you said you had that experience, but you don’t really know what you’re talking about, you know, kind of thing.
Yeah, I know, I think what comes to mind is, you know, I use chatGPT, you know, every day, you know, trying to draft an email. But, you know,
it helps me speed up my day -to -day, but there’s always two things that I need to do to make sure it’s my voice and it’s the best I can be.
One is, my input has to be good, right? So I have to make sure that whatever I ask it to do is very clear, but also the answers are usually not perfect.
Maybe they sound too textbook -y, it sounds too corporate, you know, so the tone is just,
it seems fabricated, right? So that’s the kind of work that needs to happen. You need to become an editor, right?
You can’t just ask for the answer and throw that answer back to maybe your customer, right? I’ve seen uses of chat GPD where, you know, I get an email and it clearly shows that,
you know, they basically copy -pasted the answer. And, you know, I think that’s okay, but if you do that, you lose your voice and you sort of become maybe commoditized,
right? Man, that is, like, you’re speaking to my heart personally, like, you know, my belief system right here was just like, yes,
efficiencies are great, but at the end of the day, where is your personal, like, you know, where’s your personality, right? Like, that’s what makes, if you think of anyone famous,
if you think of anyone that you respect or anything like that, there is something about their personality, there’s something about how they present, there’s something about how they put together their combination of skills that you attach yourself to.
So, yeah, that’s, I think that’s just such a critical piece, you know? For sure, for sure. So, yeah, I think it’s knowing how to use the tool to your benefit, it’s definitely a goal of the human,
but, but I think it’s, yeah, just kind of make sure it’s, it’s really serving you correctly. And I think that’s where the ingenuity of the human comes into play.
But, but I think the only way to really use it correctly is to use it a lot and be a friend of it. And, and I think that’s,
that’s the main challenge maybe for some folks out there that may be a little reluctant to try it because it’s going to replace me, so I’m not going to use it. You know, I think it’s more about just trying it out and see what it really can do and it cannot do.
But, but yeah, I might take on that. We have a person who’s Jenna Carraro has just given us a comment on LinkedIn there.
And she’s saying she’s someone who is neurodiverse. It, I’m assuming she’s talking about AI, really helps her to speed up her connections and cognitive processes for writing CVs, cover letters, etc. But those skills and beliefs are something that are part of her.
And then she does revise them, right? So it comes down to ethics. So like, let’s, let’s talk about that, you know, ethics right there. Like, for example, I’m going to put you on the spot here. I know it’s kind of out of the blue, but does your company else,
do you have a policy on, on the use of AI as an AI company? Like, do you, like, is there a written policy that everybody can look at and say, here’s when we use it, when we don’t, you know, here’s how we use it in communications. Here’s what data we can put in it.
Do you have that like a written statement? Yeah. So, you know, every, every client has sort of a different requirement. So, you know, when it comes to, and that’s, that’s what we try to do, right? We don’t dictate what our customers should or shouldn’t do.
We help them achieve what they want to do. So for example, in Europe, as you know, you know, data privacy laws are really strong. And so we adapt to that, right? In other cases, you know,
when we’re talking about giving access to kids, uh to our AI uh there’s some limitations in terms of how much content they can access for example making sure that if they are running a AI tutor conversation that that conversation doesn’t lead to maybe the use of curse words and things like that so we try to limit that um but honestly you know we’re still learning what that means and and I think the industry in
general also um but uh I think at the end of the day it’s you know our clients choose what they want to do and and hopefully we’re our clients ask us to do what they want um so uh I haven’t I haven’t faced any any kind of ethical dilemma yet I’m I’m actually kind of excited for that time to come to to see what that means and what that is um but so far it hasn’t really been like uh sure yeah 100 percent I’m
that do you have um do each of your clients or the people that you work with do you find that for instance their data use policy is that different you know is does it does it differ radically across each of the clients in terms of what you can put into an AI model AI model and whatnot because as we all know just want to say it out here you know when you put something into chat gpt your bar or whatever it
becomes a part of the system it becomes tech you know and in some cases somebody could write a question sometime in chat bpt and be like oh wow that’s interesting and they’re getting your you there’s the famous Samsung issue where the developers put all the Samsung’s proprietary code in there right to have it check it out and now it’s part of the public database so do you do you do you have you found um those data
use policies to differ radically yeah we basically have uh sort of two two uh models there so some clients don’t want uh us to retain that data so So we don’t retain it.
Some other clients don’t mind. And what do we do with that data? Well, first of all, we started, you know, it’s anonymous, right? So we’re not necessarily tracking per person,
which is what we care about is actually just getting all these different audio recordings and using it to improve our algorithm so that the next student that uses our pronunciation app gets a more accurate response.
So, and that’s really important for us, right? That’s what that’s really our secret sauce is, you know, we are the most accurate because we have 53 million users,
you know, every day using our app, improving our algorithm so that then their experience can be better. So the data is necessary. I mean, we need that data so we can serve the customer.
But again, some customers don’t want us to retain that data. And that’s fine. How much did 53 million users a day tap into this thing and receive feedback about their pronunciation of the English language?
This is something that, you know, I don’t know how to ask this question. At what point does your model become perfect? You know what I mean? Like, does it really continue to evolve and amalgamate every single day in some way?
But I mean, I’m just thinking with 53 million users a day, there’s got to be a point where you’re like, you know what, this is pretty darn solid. Like, you know, there’s only so many ways to say English, you know,
to say the English. Believe it or not, it’s a super hard problem to solve. Yes,
I’m sitting on that one. So, I mean, it’s super complicated. And while we are the forefront of it, you know,
I think there’s still so much room to improve. I mean, today we have about 90 % accuracy. So for every 10 words that you speak on our app and get feedback on,
probably one of those words will be misinterpreted. It won’t understand what you said. And that’s not that bad, but it can be. So even this morning,
I had this call with corporate clients wanting to bring our technology to their workers in India to help them improve their pronunciation.
And they noticed that, yes, one out of 10 words were not correctly given feedback too, right? So that’s still a challenge,
right? It’s not perfect. So then I go back to the role of the teacher, whereas, well, they need to interpret also the response. So because,
again, our technology is not going to be perfect, maybe nine out of 10. But once one out of 10 times, it won’t be correct. And so the teacher will have to have kind of a say there,
right? But yeah, I think there’s still a lot of room to grow. And yeah, that’s what Elsa has been working on for almost 10 years,
and I think it’s going to keep working on. So what are the, you just mentioned kind of one, not objection, but maybe a flag or a pushback from a client.
What other types of objections or questions, concerns, like when you’re out there trying to close a deal, and you’re saying,
look, this is going to really improve your workflow, and this is going to level up your workforce really, like the educational component of leveling up your workforce is profound, what are the things that you get pushed back on?
Other things cost, obviously. Of course. Cost is always important. So I think there’s, there’s maybe camps. I mean,
there are those who have, you know, they’re true believers, right? And the AI and, and but that could be a bad thing, right? Because maybe their expectation is that,
you know, we, with the use of our tool, their workers can improve their English level in, you know, two days, right? I mean,
we can’t do that, right? That’s, that’s not, you know, you know, that’s beyond the expectation that we can meet. And so there’s an education around,
okay, what it is that you can expect, right? So we say, for example, with 10 minutes a day, with our solution, within a span of three months, you can improve one or two English levels,
right? And that’s, you know, that’s realistic. But if they believe that they could maybe in, you know, 10 days with a couple of minutes a day,
that they would improve their English significantly, then that’s not the case. So, so sometimes it’s just about like lowering expectations. On the other hand, then there’s like the,
the, the ones that start off with maybe negativity, let’s say. And my approach there is always to try not to react to that. You know, maybe it’s more of a sales tip out there for sales folks out there.
It’s, you know, not try to speak against what they’re saying, right? But rather try to understand where they’re at.
For instance, this example that I mentioned where there was a lot of focus on the one word out of 10 that we get wrong, instead of focusing on the nine out of 10 that we get right.
So trying to explain, you know, that we’re not perfect, and that you shouldn’t expect us to be perfect. but that the alternative is way worse,
right? Which means you cannot scale with all the humans the improvement of communication skills, right? So I think those are some of the ideas out there.
But, but yeah, there’s and you just named like one of the misconceptions, you know, sort of, hey, you know, you’ve got this, you’ve got this tool. And so I’m going to be able to radically get, you know, better, faster, you know,
so much, you know, so rapidly. I, I love head that also then just again raises that, that flag of, okay, look, we do a lingo as one of,
you know, the world’s most famous examples is like, hey, you know, you can, you can start learning the language there, but guess what? If you don’t show up every day, if you don’t do your, you know, your hour every day, you’re not,
you know, you’re not going to take those steps forward. It’s like any other skill at the end of the day, we are still human, right? And so putting in the work is, it’s,
it’s a non -negotiable, right? It has to be done. Yeah, it’s, we haven’t gotten to the point where, you know, like in the matrix, you just put in a chip and then you know it,
maybe, right? I mean, I feel like everything’s possible. So, but we’re not there yet. You still got to, you know, I sweat a little bit and you got to put in the work. I just don’t know, what would that be like?
You know, you got Elon Musk’s or Elon Musk’s, you know, mine, whatever, that company where they’re plugging into your brain. Like, what would that be like if you could just like push a button and then suddenly,
oh, hey, I, I now have Chinese as a fluency. That’s great. I just, I have no idea what that flipping that switch might look like. It’s crazy. Talk to me like sort of kind of,
kind of bringing us back full circle, like, so where, where do you see things going over the next, let’s just say near term, 18, 16 months, you know, 24 months around that interface between the L and D department and the HR department using a tool like Elsa,
you know, do you see that? I think there’s a lot of fear on that. Do you see those, those staff shrinking and, you know, more being accomplished by, by, by fewer people? Do you feel like it’s more of an evolution in those jobs?
And so they’re going to just be evolving how they’re inputting into the process? Or like, what do you look seeing there? I mean, that’s a hard question,
right, too. And predicting the future, it’s impossible. I think, I kind of think about the present day, right? We all know we are in sort of an economic scenario where profit is really important.
Cost reduction, meaning we got to make sure that whichever investment, whichever activity or initiative we’re putting money into it,
that it’s really giving us value back, right? Return on our investment. So in that, I think it applies to everything in every department. When it comes to learning and development,
I think, first of all, being able to measure ROI has always been super hard, right? And really, that’s a shame because that’s,
I think, one of the reasons why L and D budgets have always been somewhat limited because they haven’t been able to really prove. Now that pressure is even stronger now to be able to do that.
So from my perspective, which is a very small part of the world, which is how do we make sure that investment on making sure learners,
making sure workers are improving their speaking skills? Well, what we do is we’re constantly measuring everything.
So if you’re a worker running, you’re talking to or speaking to your AI tutor, well, we assess that and every other activity that you have.
So we know exactly what your English level is today, and we know at every single activity that you run, how it improves. So we’re able to very easily show that if you start at,
let’s say, a 80 % or like a B1 level, and then in two months, you went to a 90%, it’s very clear that your score has improved, and we can show that.
And I think that just applies to anything around education, that now there’s more pressure to show that result. That’s what I would say is like, let me focus and the focus that I try to bring to the conversation when I talk to L &D folks out there.
Okay, final question for you in this, I love these conversations, but now final question for you is, how much do you see the line blurring between a person,
you know, an individual’s experience in a learning setting like this, between interfacing with a human and interfacing with an AI model. And so this goes to the ability to AI to personalize the situation,
the ability or the need for, because I’m like, I have this vision of, you know, like one, let’s just say one English coach with, you know, 500 students and they’re able to just kind of like,
you know, I had this vision, you know, they’re kind of in that room with 500 screens. And, you know, as the AI model saying, hey, then, you know, you can kind of blip in there, blip in there, you know, like what’s,
how much is that line blurring between working with a human and working with AI? I think,
I think we’re sort of in the midst of that, right? So, so I think we’re still figuring out the role of each and when they should work together or separately.
I mean, what I can say is that the focus that we’re having is we’re trying to solve for the needs of two different interests.
On one hand we want to help the individual right we want to help with their efficiencies when it comes to speaking skills but we also want to be able to help accomplish the goals of the organization what they care about.
So in a simple example, if we have a project manager from India running international projects in English, well what does he care or she care about?
Well, we would need to be able to recommend certain contents just for him or for her and based on his or her deficiencies. And that’s something our art technology can do it can just detect what your issues are and then relevant certain content just for you and it’s also auto generated so the content gets generated just for you for your needs.
But on the other hand, the company cares about the project manager being able to speak very clearly when it comes to the job role so maybe having a meeting where the project manager presents the status report of that meeting.
Well, that’s very specific and it’s job you know it’s job relevant right so we would be needing to provide content that’s related to that job. So we when you meet those both of those needs,
you know you’re improving the individual needs of every single person but at the same time you’re adding content and skill related to the job that they have to do every day. then that becomes very tactical,
right? Because from our perspective, we’re not teaching general English, right? You know, we don’t necessarily care about teaching, you know,
Shakespeare, right? What we need to do is make sure that if you need a specific lingo, if you have a very specific job,
that you can do that job very well in English, right? And that’s how we can actually improve in a short time, right? When you have a very specific use of English, right?
And so that’s sort of our approach to really make sure that workers are improving their English skills and where they need to,
right? And not just sort of a generalistic approach, which yields to kind of poor results. Nice. Yeah, so I love that. So it’s like you’re inserting at the moment or where,
you know, both with specific language, with specific need and specific moments, you know, it’s that learning the flow of work kind of thing and whatnot. Excellent. Indeed. Indeed, it’s cool. Pablo,
I can’t thank you enough for taking time out of your busy day to talk with us about Elsa, about, you know, when AI is appropriate, when we still need humans and, you know, just taking on the challenge of these,
you know, this complex conversation. If somebody wants to talk to you more, how did they get a hold of you? They can reach out via LinkedIn. You can find me there,
just Pablo Ferreiro. And I’m sure you’ll find me as, you know, leading Elsa for the Americas. And I could share my email later also on those comments on the LinkedIn post.
But yeah, thank you so much, Ladik, for the opportunity. It was a fun chat. I hope everyone enjoyed it as well. So yeah, thanks so much. Thank you again for listening to the E -Learn podcast here from OpenLMS.
I just wanted to ask one more time, if you enjoyed this show, if you learned something, if you were inspired, if you were challenged, if you feel like, you know, this is something you can take into your practice, please do me a favor.
And right now on your podcast player, hit subscribe. That way you’re never going to miss a future episode. Also, come over to elearnmagazine .com and subscribe there as well, because we have tons of great information about how to create killer online learning outcomes.