WEBVTT 1 00:00:01.040 --> 00:00:04.640 Hello, and welcome, everybody. Guys, it is getting close to 2 00:00:04.640 --> 00:00:08.400 Christmas. This is Joe from Sudapri. Io, and therefore, I'm bringing 3 00:00:08.400 --> 00:00:12.080 you a very special bonus episode this week 4 00:00:12.080 --> 00:00:15.735 to you very shortly before Christmas. But nonetheless, I would 5 00:00:15.735 --> 00:00:18.235 like to welcome Daniel here. Hey. How are you doing? 6 00:00:19.255 --> 00:00:22.634 Hey, Jan. I'm doing well. Thanks. Thanks for having me. 7 00:00:23.914 --> 00:00:27.550 AI pleasure. We may tell our audience that this, recording 8 00:00:27.550 --> 00:00:30.910 is sponsored by Frankfurt, meaning the business development 9 00:00:30.910 --> 00:00:34.670 agency who is also supporting Frankfurt Forward. And 10 00:00:34.670 --> 00:00:38.270 the reason you are here, you guys won start up of the 11 00:00:38.270 --> 00:00:41.585 year 2024. Congratulations 12 00:00:42.045 --> 00:00:45.645 to that. Did 13 00:00:45.645 --> 00:00:49.405 this recognition first, can 14 00:00:49.405 --> 00:00:53.180 you tell us a little bit about you and your company 15 00:00:53.180 --> 00:00:56.780 before we get into the specific questions? Yeah. 16 00:00:56.780 --> 00:00:59.440 Sure. Yeah. So 17 00:01:00.540 --> 00:01:04.140 Daniel Iglesias is my name. I am from the AI Main 18 00:01:04.140 --> 00:01:07.535 region. I AI married to teacher. My 19 00:01:09.515 --> 00:01:13.295 professional background is in banking technology consulting. 20 00:01:13.515 --> 00:01:16.895 That's what I did before I founded, 21 00:01:17.355 --> 00:01:20.630 Digisapiens in, 2020. 22 00:01:22.290 --> 00:01:25.909 And, what we do at Digi DigiZapiens is 23 00:01:26.050 --> 00:01:29.830 we produce speech recognition systems 24 00:01:30.130 --> 00:01:33.810 that not just recognize what is being 25 00:01:33.810 --> 00:01:37.104 said, rather how it's being said. So 26 00:01:37.485 --> 00:01:41.085 it's a speech recognition system that is geared towards 27 00:01:41.085 --> 00:01:44.545 special use cases, that 28 00:01:45.405 --> 00:01:48.930 are relevant to measure how well somebody speaks or 29 00:01:48.930 --> 00:01:52.630 reads. And with the help of this speech 30 00:01:52.770 --> 00:01:55.739 recognition technology that, is AI, 31 00:01:56.610 --> 00:02:00.450 so that we've built ourselves, you can build 32 00:02:00.450 --> 00:02:04.005 use cases in especially in the area of 33 00:02:04.005 --> 00:02:07.765 education. So that's where we started. And the first use 34 00:02:07.765 --> 00:02:11.144 case that was built with our speech recognition technology 35 00:02:11.444 --> 00:02:14.910 was the LALI 2. LALI 2 36 00:02:15.070 --> 00:02:18.610 stands for loud laser tutor in German, which is, 37 00:02:19.470 --> 00:02:22.050 reading aloud tutors. So it's it's a tool, 38 00:02:22.910 --> 00:02:26.450 that listens to students in schools 39 00:02:26.750 --> 00:02:29.410 from 2nd to 7th grade in Germany. 40 00:02:30.965 --> 00:02:34.584 While they read aloud, it analyzes how 41 00:02:34.644 --> 00:02:38.265 well they do that, so it carries out a diagnosis. 42 00:02:39.204 --> 00:02:42.105 And after that diagnosis, they are, 43 00:02:43.364 --> 00:02:47.000 being trained to become better readers 44 00:02:47.140 --> 00:02:50.579 based on the, diagnosis that was carried out before 45 00:02:50.579 --> 00:02:54.099 that. So basically, AI looking for a question here. Sure. 46 00:02:54.659 --> 00:02:58.099 When you said they are your tool 47 00:02:58.099 --> 00:03:01.795 helps, how well they are speaking 48 00:03:01.795 --> 00:03:05.495 their AI the tone of voice recognition. Is it only 49 00:03:05.875 --> 00:03:09.635 how well they vocalize the tones? Or is 50 00:03:09.635 --> 00:03:13.310 it also that you can deduce some some 51 00:03:13.450 --> 00:03:16.590 level of their understanding of what they're reading? 52 00:03:18.810 --> 00:03:21.950 The speech recognition technology itself is, 53 00:03:23.130 --> 00:03:26.825 audio technology, so it listens. So we it detects everything 54 00:03:26.885 --> 00:03:29.705 that can be caught by a microphone. 55 00:03:30.725 --> 00:03:34.405 And, in contrast to, I 56 00:03:34.405 --> 00:03:37.765 call it, regular speech recognition systems where the goal is to 57 00:03:37.765 --> 00:03:41.460 detect the probable intent. So what would 58 00:03:41.460 --> 00:03:44.900 be this what are you probably meaning? So, 59 00:03:45.220 --> 00:03:48.980 the goal is to carry out a a 60 00:03:48.980 --> 00:03:52.500 task or, AI I'm 61 00:03:52.500 --> 00:03:56.200 looking, a command. You have to carry out a command. Mhmm. Yeah. 62 00:03:56.495 --> 00:04:00.254 Play this and that song or whatever. In contrast to 63 00:04:00.254 --> 00:04:04.015 that, we really transcribe and list listen to what has 64 00:04:04.015 --> 00:04:07.555 actually been said. So this includes 65 00:04:07.694 --> 00:04:11.350 arrows AI my that I already said now a few 66 00:04:11.730 --> 00:04:15.250 times. And repetitions, or 67 00:04:15.250 --> 00:04:18.769 text so regarding text repetitions, things that have 68 00:04:18.769 --> 00:04:22.470 been left out or added, the tonality 69 00:04:22.770 --> 00:04:26.435 of things, and also whether you pronounce words 70 00:04:26.435 --> 00:04:29.955 correctly. Yeah. So that's what the, speech 71 00:04:29.955 --> 00:04:32.775 recognition technology does. But we also develop, 72 00:04:33.875 --> 00:04:37.530 systems that help understand help 73 00:04:37.770 --> 00:04:41.150 students understand the text better by 74 00:04:41.690 --> 00:04:42.190 generating, 75 00:04:45.130 --> 00:04:48.810 differentiated quizzes. So, you cannot ask 76 00:04:48.810 --> 00:04:52.455 every reader the same question. It must be adapted to 77 00:04:52.455 --> 00:04:55.915 his reading level and also understanding 78 00:04:56.375 --> 00:04:59.835 capabilities. So the complexity of the questions 79 00:05:00.455 --> 00:05:03.835 and the possible answers also must be adapted. So, 80 00:05:04.615 --> 00:05:08.360 all in all, we're in the business of providing 81 00:05:10.020 --> 00:05:13.700 education, specialists, with, 82 00:05:14.100 --> 00:05:17.160 with the necessary tools to build very innovative 83 00:05:17.940 --> 00:05:21.160 adaptive tools, for learning 84 00:05:22.005 --> 00:05:23.945 reading skills or language 85 00:05:25.525 --> 00:05:28.825 skills? That is exactly what I had in mind. 86 00:05:29.685 --> 00:05:33.445 Vividly remember when, for the first time 87 00:05:33.445 --> 00:05:36.820 in my life, I understood a Chinese joke 88 00:05:36.820 --> 00:05:40.340 about foreigners, instead of how 89 00:05:40.340 --> 00:05:44.180 how, she said, how how. So the the 1 means 90 00:05:44.180 --> 00:05:47.940 good good, the other means mouse mouse, different tones. So the the 91 00:05:47.940 --> 00:05:51.625 the the question is here, how 92 00:05:51.625 --> 00:05:55.384 many languages can you do? And isn't something like 93 00:05:55.384 --> 00:05:58.845 Chinese where there's a different if I ask or 94 00:06:00.505 --> 00:06:04.264 the 1 we means please ask. So if you ask for direction, 95 00:06:04.264 --> 00:06:07.990 the other one's the other 1 means please kiss. Well, I made 96 00:06:07.990 --> 00:06:11.290 an older Chinese lady on the streets of Beijing really blush. 97 00:06:12.390 --> 00:06:15.750 How many of those differences could you actually do? 98 00:06:15.750 --> 00:06:19.110 Because on the top of my mind, yes, of course, English is a little bit 99 00:06:19.110 --> 00:06:22.925 difficult to pronounce Spanish as well. But if the tone really makes a 100 00:06:22.925 --> 00:06:26.544 difference, like in languages, like Chinese, Cantonese, and so on and so forth, 101 00:06:27.004 --> 00:06:30.685 that would weigh your AI way 2 would really come 102 00:06:30.685 --> 00:06:34.284 in handy. So my question would be, how 103 00:06:34.284 --> 00:06:37.850 much can you do there? Could you, give us an 104 00:06:37.850 --> 00:06:41.610 idea of the granularity and languages you cover? Yes. 105 00:06:41.610 --> 00:06:44.830 We so as of now, we're covering German and English. 106 00:06:46.169 --> 00:06:49.925 We did not experiment with Chinese since the 107 00:06:49.925 --> 00:06:53.465 Chinese education technology market is highly regulated 108 00:06:53.685 --> 00:06:56.664 and basically close towards foreigners. 109 00:06:58.245 --> 00:07:01.224 But, we my team, I have 110 00:07:02.170 --> 00:07:04.270 dedicated experts in, 111 00:07:05.530 --> 00:07:09.210 working with non Roman 112 00:07:09.210 --> 00:07:12.110 languages, so especially also Indian languages. 113 00:07:12.570 --> 00:07:16.395 There's 40 something. Sorry that I cannot recall 114 00:07:16.395 --> 00:07:19.835 the exact number, but there's, more than 40 115 00:07:19.835 --> 00:07:23.275 languages, official languages spoken in India. We can work with 116 00:07:23.275 --> 00:07:26.575 those. We can also work with, Arabic 117 00:07:28.979 --> 00:07:32.440 character sets. But as of now, in terms of 118 00:07:33.379 --> 00:07:37.159 solutions that are at hand, we can work with 119 00:07:37.379 --> 00:07:40.885 Germany, German language, and, we 120 00:07:41.365 --> 00:07:44.905 or next year, we will also launch, the English version 121 00:07:45.665 --> 00:07:48.505 AI API, and other languages. 122 00:07:49.445 --> 00:07:53.205 We have the skills and capabilities to train and 123 00:07:53.205 --> 00:07:56.860 fine tune models, with a short pilot 124 00:07:56.860 --> 00:08:00.080 project that we need to carry out with potential customers. 125 00:08:01.740 --> 00:08:05.420 So that means you already, by the way, I linked it in the show 126 00:08:05.420 --> 00:08:08.595 notes. La Lalu. Who is it? 127 00:08:08.835 --> 00:08:12.292 La la la la la la la la la la la la la la la 128 00:08:12.292 --> 00:08:13.235 la la la la la la la la la la la la la la la 129 00:08:13.235 --> 00:08:16.455 la la. It's There's a there's a, a 130 00:08:18.355 --> 00:08:21.555 song that you sing to children before they go to bed, 131 00:08:22.445 --> 00:08:24.700 AI. It's called. Yeah? 132 00:08:26.680 --> 00:08:30.300 And, we have some similarities there. So it's called 133 00:08:30.440 --> 00:08:34.220 too. It's about 3. Mhmm. 134 00:08:35.640 --> 00:08:39.025 So, sorry. Just 135 00:08:39.025 --> 00:08:42.804 typing here, that we also linked the 136 00:08:43.105 --> 00:08:46.865 song here, in the show notes. My my quest 137 00:08:47.985 --> 00:08:51.600 so this already establishes something we 138 00:08:51.600 --> 00:08:55.120 could deduce from what you're saying. So, basically, you are 139 00:08:55.120 --> 00:08:58.900 not a customer facing product. You're 1 of the tools, the APIs, 140 00:09:00.080 --> 00:09:03.705 others could include, could work with in 141 00:09:03.705 --> 00:09:07.385 developing their own client facing b to b, b to c, b to 142 00:09:07.385 --> 00:09:10.285 g tools. Right? Yes. 143 00:09:10.985 --> 00:09:14.425 With the asterisk, we we have 144 00:09:14.425 --> 00:09:17.840 developed the LALI 2, for our partner at hence, 145 00:09:18.480 --> 00:09:21.780 So we do develop platforms and applications, 146 00:09:22.160 --> 00:09:25.940 but we rather license our the speech technology 147 00:09:27.440 --> 00:09:30.420 to partners, b to b, b to g, b to c, whatever. 148 00:09:31.274 --> 00:09:34.714 We are already discussing, the tool. But 149 00:09:34.714 --> 00:09:38.555 before that, I I actually wanted to be because I have so 150 00:09:38.555 --> 00:09:42.334 many questions. Before that, I would actually, wanted 151 00:09:42.394 --> 00:09:46.060 to ask you 152 00:09:46.120 --> 00:09:49.640 where this idea is coming from. I do believe I have an idea since you're 153 00:09:49.640 --> 00:09:53.180 married to your teacher, but you have been working in banking, 154 00:09:53.480 --> 00:09:57.320 finance, technology, triangle consulting. So 155 00:09:57.320 --> 00:10:01.105 so, where did the idea come from? How did you get 156 00:10:01.105 --> 00:10:04.644 that? And especially the question, when did you decide 157 00:10:04.704 --> 00:10:07.605 jump ship to really do this full time? 158 00:10:08.945 --> 00:10:12.660 Okay. So I was, like you said, I 159 00:10:12.660 --> 00:10:15.880 was about 17 or 18 years into banking technology, 160 00:10:16.980 --> 00:10:20.500 and, I always had the goal 161 00:10:20.500 --> 00:10:23.160 to promote young people 162 00:10:23.860 --> 00:10:27.255 in achieving, how do you call 163 00:10:27.255 --> 00:10:31.014 it, higher education. Let's put it this way, to 164 00:10:31.014 --> 00:10:34.135 get the most out of their potential. So I did, 165 00:10:34.454 --> 00:10:37.675 trainings in schools for how to apply to a job, 166 00:10:38.579 --> 00:10:42.339 how do I choose the right job for me, etcetera etcetera. And 167 00:10:42.339 --> 00:10:45.160 I always had this passion for helping young people. 168 00:10:45.940 --> 00:10:49.720 So, that's that that passion was always there. But in 169 00:10:49.860 --> 00:10:53.645 2019, that was the time where I really, 170 00:10:55.225 --> 00:10:58.825 thought about what can I do with the skills that I have and the 171 00:10:58.825 --> 00:11:02.665 knowledge that I have? And, to to really have an 172 00:11:02.665 --> 00:11:06.460 impact on our youth in a bigger scale. Because 173 00:11:06.460 --> 00:11:10.220 this is the AI, shortly after AI became a father of 174 00:11:10.220 --> 00:11:12.960 a of a daughter, and I have observed 175 00:11:13.740 --> 00:11:17.420 certain trends in our society in Germany. 40 to 176 00:11:17.420 --> 00:11:19.440 50% of the children have 177 00:11:21.575 --> 00:11:24.955 background, with non German parents or migration 178 00:11:25.095 --> 00:11:28.555 background, as you would call it in Germany. And this 179 00:11:28.695 --> 00:11:32.215 leads to some hurdles and 180 00:11:32.215 --> 00:11:35.580 some difficulties, in, 181 00:11:35.920 --> 00:11:39.360 in the school system, and and and since 182 00:11:39.360 --> 00:11:43.200 we also, at the same time, have a shortage of 183 00:11:43.200 --> 00:11:46.880 shortage of teachers. And I have also 184 00:11:46.880 --> 00:11:49.220 observed what happens in the market regarding 185 00:11:50.824 --> 00:11:54.365 the upcoming AI of AI, robotics, automation. 186 00:11:54.665 --> 00:11:58.125 I was part of it in banking. So I added 187 00:11:59.944 --> 00:12:02.925 the deterioration of reading skills 188 00:12:03.785 --> 00:12:07.050 towards higher requirements 189 00:12:07.270 --> 00:12:10.950 regarding job skills and came up, 190 00:12:11.430 --> 00:12:14.410 with, with a 191 00:12:14.870 --> 00:12:18.565 perspective that I didn't like for the future of my children. So that's 192 00:12:18.565 --> 00:12:22.105 why I decided to take my skills, 193 00:12:22.404 --> 00:12:25.845 which is general management, business 194 00:12:25.845 --> 00:12:29.225 development, technology understanding skills, and 195 00:12:30.005 --> 00:12:33.280 work together with the best experts I can find, 196 00:12:33.740 --> 00:12:36.560 in terms of reading capabilities 197 00:12:37.100 --> 00:12:40.540 and, reading training and, 198 00:12:41.500 --> 00:12:45.100 excellent techno technology experts bring those things together and 199 00:12:45.100 --> 00:12:47.040 build what you find today. 200 00:12:48.565 --> 00:12:52.105 Mhmm. The tone of voice recognition 201 00:12:52.165 --> 00:12:55.765 is fascinating. You already told 202 00:12:55.765 --> 00:12:59.525 us the, loud reading tutor is 203 00:12:59.525 --> 00:13:03.290 something you develop for a customer. Could you also 204 00:13:03.290 --> 00:13:06.670 share another example already where an external, 205 00:13:08.250 --> 00:13:12.010 client is using your tool? The 206 00:13:12.010 --> 00:13:15.770 second example is in the making. It's not ready to be 207 00:13:15.770 --> 00:13:19.275 shared publicly, but, I must openly say the 208 00:13:19.275 --> 00:13:22.255 past was, highly, 209 00:13:23.195 --> 00:13:26.875 we were highly invested into building 210 00:13:26.875 --> 00:13:30.335 that tool, which is 1 of a kind. It took a lot of 211 00:13:30.950 --> 00:13:34.550 attention and all of our resources to get it running in 212 00:13:34.550 --> 00:13:37.530 time and, make it 213 00:13:38.150 --> 00:13:41.990 scalable and stable and user friendly. So 214 00:13:41.990 --> 00:13:45.605 now that the product is fully marketable and 215 00:13:45.745 --> 00:13:48.725 most of the almost all bugs are fixed, yeah, 216 00:13:50.065 --> 00:13:53.505 now we are able and ready to focus in on 217 00:13:53.505 --> 00:13:56.085 new, projects and partners. 218 00:14:00.080 --> 00:14:02.640 AI see. Talked about 219 00:14:03.660 --> 00:14:07.340 partners here and winning new clients. Winning the Frankfurt Forward award 220 00:14:07.340 --> 00:14:11.095 is a huge accomplishment. What do you think made DigiSapiens 221 00:14:11.395 --> 00:14:13.895 stand out among the competition this year? 222 00:14:16.401 --> 00:14:18.615 AI think it's the social impact, 223 00:14:20.275 --> 00:14:23.815 dimension of what we're doing. We are for profit social enterprise, 224 00:14:23.955 --> 00:14:27.000 so we are here to do good 225 00:14:29.060 --> 00:14:32.900 and, earn some money at the same time. And I think the audience 226 00:14:32.960 --> 00:14:36.660 AI that idea, and I think there's a lot of ad 227 00:14:36.660 --> 00:14:40.475 techs out there, but pulling it off in the way 228 00:14:40.475 --> 00:14:44.015 that we did, by partnering with such a, 229 00:14:45.115 --> 00:14:47.775 renowned brand AI Ernst Kedfalak, 230 00:14:49.115 --> 00:14:52.810 as a first initial project. And at the same time, 231 00:14:53.830 --> 00:14:57.590 building such a unique technology like we do. 232 00:14:57.590 --> 00:15:01.050 I think that's what impressed the jury and, 233 00:15:02.710 --> 00:15:06.255 caused a lot of, support in the audience. 234 00:15:09.035 --> 00:15:12.394 I see. I see. Your 235 00:15:12.394 --> 00:15:15.855 technology has potential across industries. 236 00:15:17.435 --> 00:15:19.935 We already know you're working and focused 237 00:15:21.420 --> 00:15:25.200 on the education, but could you also see some other 238 00:15:25.340 --> 00:15:29.100 industries where you could, like, in the future, a few years down the road, 239 00:15:29.100 --> 00:15:29.840 apply it? 240 00:15:32.785 --> 00:15:36.005 Yes. Outside of, education, 241 00:15:37.265 --> 00:15:40.645 there is also the entertainment and gaming industry 242 00:15:41.185 --> 00:15:44.485 that could work with our technology, 243 00:15:46.540 --> 00:15:50.240 where you could you could use it to build games that, 244 00:15:51.100 --> 00:15:54.720 based on reading skills, which would 245 00:15:55.420 --> 00:15:59.100 be some sort of, yeah, educational games at at the same 246 00:15:59.100 --> 00:16:02.685 time, but you can also use it to, 247 00:16:04.264 --> 00:16:07.485 to build, how do you call it? 248 00:16:08.665 --> 00:16:11.945 A a presentation trainer or speech trainer that, 249 00:16:12.505 --> 00:16:16.320 helps you become a better speaker for public speeches. 250 00:16:18.320 --> 00:16:21.780 The creativity and opportunities are unlimited, 251 00:16:22.000 --> 00:16:25.700 but everything that involves the ability 252 00:16:25.840 --> 00:16:29.654 to carefully listen to what and how things are 253 00:16:29.654 --> 00:16:32.395 being said. You know? Mhmm. 254 00:16:34.695 --> 00:16:38.295 I AI was wondering, AI sure you 255 00:16:38.295 --> 00:16:41.920 thought a lot about potential use cases. Could you share, like, 256 00:16:41.920 --> 00:16:45.060 the the the the the the most interesting, the most, 257 00:16:46.240 --> 00:16:48.820 quirky 1 you already came up with? 258 00:16:50.399 --> 00:16:53.300 No no need that it actually AI, but you thought, 259 00:16:54.015 --> 00:16:57.155 theoretically, our idea could also be applied too. 260 00:16:58.815 --> 00:17:02.415 Yes. So we have applied it to reading 261 00:17:02.415 --> 00:17:05.714 learning or reading promotion, but, very, 262 00:17:06.734 --> 00:17:10.250 very relevant use cases also in the area 263 00:17:10.250 --> 00:17:13.790 of language learning. You have seen 264 00:17:14.089 --> 00:17:17.849 big companies like Bubble, etcetera use it in 265 00:17:17.849 --> 00:17:20.990 some way. Yeah. And we envision 266 00:17:22.204 --> 00:17:25.645 other ways that are much more focused on 267 00:17:25.645 --> 00:17:29.485 dialogues, that our technology could be 268 00:17:29.485 --> 00:17:33.085 used for to promote language skills or learn a new 269 00:17:33.085 --> 00:17:36.840 language. So we are able to evaluate how 270 00:17:36.840 --> 00:17:39.820 things are being pronounced. That's 1 major skill. We can 271 00:17:40.279 --> 00:17:44.120 analyze, literally what has been said. 272 00:17:44.120 --> 00:17:46.860 We can analyze 4 ds, all areas 273 00:17:47.575 --> 00:17:51.355 of, language that we can analyze. 274 00:17:51.495 --> 00:17:55.255 And I think this is relevant if you want to learn a language 275 00:17:55.255 --> 00:17:58.615 properly. You have, you have given some example from 276 00:17:58.615 --> 00:18:02.320 Chinese. If you listen to people 277 00:18:02.320 --> 00:18:05.840 talking German with all their accents and 278 00:18:05.840 --> 00:18:09.380 dialects, we also have ways to 279 00:18:10.480 --> 00:18:12.320 tackle, dialect, 280 00:18:14.080 --> 00:18:17.875 dialects because the way you use your mouth, your 281 00:18:17.875 --> 00:18:21.635 tongue, your teeth. So your 282 00:18:21.635 --> 00:18:25.155 complete speech apparatus is also something that we can 283 00:18:25.155 --> 00:18:28.780 derive from the audio signal and combining all 284 00:18:28.780 --> 00:18:32.080 those, all those measurements 285 00:18:33.020 --> 00:18:36.160 into a cohesive didactical 286 00:18:36.620 --> 00:18:39.040 concept is something really unique, 287 00:18:40.300 --> 00:18:43.815 that we haven't seen so far. Mhmm. 288 00:18:45.715 --> 00:18:49.395 Going into a little bit different topic because every everybody talks 289 00:18:49.395 --> 00:18:52.995 about accuracy, like fantasizing AI and 290 00:18:52.995 --> 00:18:56.675 ethical use of AI. With analyzing of tones, you're 291 00:18:56.675 --> 00:19:00.130 you're, collecting potentially 292 00:19:00.190 --> 00:19:03.870 sensitive information or your clients do and you process 293 00:19:03.870 --> 00:19:07.649 this. How do you, ensure 294 00:19:07.870 --> 00:19:11.330 the ethical use and the accuracy when handling the sensitive 295 00:19:11.389 --> 00:19:14.205 data? Yeah. So 296 00:19:15.705 --> 00:19:19.405 the in Germany, it's always a relevant question whether it's a 297 00:19:21.385 --> 00:19:25.085 personal personality AI identity related data. 298 00:19:25.330 --> 00:19:28.530 You know? That's 1 major question. And, 299 00:19:29.570 --> 00:19:33.170 the the thing is, if voice 300 00:19:33.170 --> 00:19:36.850 really is such data, you would 301 00:19:36.850 --> 00:19:40.445 need to have some registry 302 00:19:41.145 --> 00:19:44.525 of confirmed identities 303 00:19:45.065 --> 00:19:48.845 that are linked to a voice profile or voice biometrics profile 304 00:19:49.785 --> 00:19:52.765 to to pose some danger 305 00:19:54.419 --> 00:19:57.780 to a data leak or whatever. You know? Mhmm. 306 00:19:58.179 --> 00:20:01.940 And this is not the case and will never be the case. We will 307 00:20:02.020 --> 00:20:05.140 I well, let's let's not say it will never be the case. I don't know 308 00:20:05.140 --> 00:20:08.985 what happens in, the year 21 100. But as of now, we don't 309 00:20:08.985 --> 00:20:12.365 have a voice register, a public 1. 310 00:20:12.585 --> 00:20:16.285 And, the question is also, even if this existed 311 00:20:16.585 --> 00:20:20.105 on a government level, the question is also, do 312 00:20:20.105 --> 00:20:23.390 companies, do other individuals, 313 00:20:24.090 --> 00:20:27.850 criminals have access to this registry, and can they use it to harm 314 00:20:27.850 --> 00:20:31.690 you? And the I don't I don't see 315 00:20:31.690 --> 00:20:35.295 that. When you ask me about ethics 316 00:20:35.535 --> 00:20:39.375 in my context, we regard the topic of 317 00:20:39.375 --> 00:20:42.835 ethics in terms of accessibility to 318 00:20:43.215 --> 00:20:46.675 our solutions. So can somebody from Bavaria 319 00:20:46.895 --> 00:20:50.270 use it, as the same way as somebody from 320 00:20:50.270 --> 00:20:53.890 Saxony can use it? And can somebody 321 00:20:54.110 --> 00:20:57.870 with Turkish or Arab 322 00:20:57.870 --> 00:21:01.490 accent use it AI somebody from, Hanover 323 00:21:01.950 --> 00:21:05.785 without any accent? And the answer is yes. 324 00:21:05.845 --> 00:21:09.605 So we AI to and we put a lot of 325 00:21:09.605 --> 00:21:13.305 effort into avoiding any biases 326 00:21:13.685 --> 00:21:17.460 in our speech recognition system by 327 00:21:17.460 --> 00:21:20.520 training it very profoundly 328 00:21:20.980 --> 00:21:24.520 with different accents and dialects 329 00:21:25.380 --> 00:21:28.280 to make sure that it works with every 330 00:21:29.220 --> 00:21:32.935 user. Yeah. So that's how we look at that. 331 00:21:33.875 --> 00:21:37.655 So you you you put a lot of effort into that, 332 00:21:37.715 --> 00:21:41.315 making it possible for everybody to understand. We may tell the 333 00:21:41.315 --> 00:21:44.935 audience that there are some people who speak very, hefty, 334 00:21:45.475 --> 00:21:49.276 local accents, not only from Bavaria, but AI, 335 00:21:49.420 --> 00:21:53.180 Thales, Saxony, and so on and so forth, but also Platych 336 00:21:53.180 --> 00:21:56.780 in the very north. It's really hard for you to understand when you're from a 337 00:21:56.780 --> 00:22:00.000 different area. Many Americans will understand, 338 00:22:01.335 --> 00:22:05.095 will, have an idea when I talk about somebody with a very 339 00:22:05.095 --> 00:22:08.934 heavy southern draw or something that's also hard to understand. So 340 00:22:08.934 --> 00:22:12.535 you took care to cover all those 341 00:22:12.535 --> 00:22:16.169 peoples and not disqualify somebody there. So I do believe there 342 00:22:16.169 --> 00:22:19.309 was a lot of development work going into. 343 00:22:20.010 --> 00:22:23.770 What challenges did you face, and how did you overcome this in 344 00:22:23.770 --> 00:22:27.370 developing digital DigiSapiens, not digital. 345 00:22:27.370 --> 00:22:30.365 DigiSapiens. Sorry. Yes. So 346 00:22:32.345 --> 00:22:36.045 yeah. There there were a lot. So which 1 can I 347 00:22:37.145 --> 00:22:40.665 yeah? So we started as a company that wanted 348 00:22:40.665 --> 00:22:43.965 to provide speech recognition systems only, 349 00:22:44.470 --> 00:22:47.990 And then we were suddenly in the position to develop a whole 350 00:22:47.990 --> 00:22:51.690 platform. So in a short time, we had to 351 00:22:51.750 --> 00:22:54.890 set up a team that was able to do that, build a 352 00:22:55.910 --> 00:22:59.655 product team or, UX and 353 00:22:59.655 --> 00:23:03.275 front end development team around our core technology 354 00:23:03.415 --> 00:23:05.415 team in a very short time and, 355 00:23:08.055 --> 00:23:11.415 build a product that fulfills high 356 00:23:11.415 --> 00:23:15.040 expectations. And this, was a 357 00:23:15.040 --> 00:23:18.660 real challenge. To be honest, we, 358 00:23:19.440 --> 00:23:22.900 like a lot of other startups, were 359 00:23:23.520 --> 00:23:27.200 in the forced to publish a product a 360 00:23:27.200 --> 00:23:31.034 year ago that was not perfect, far from perfect. 361 00:23:31.815 --> 00:23:35.495 So we got that feedback in the beginning, but we worked 362 00:23:35.495 --> 00:23:36.715 really, really hard, 363 00:23:39.255 --> 00:23:42.970 with our team, which also includes the first level 364 00:23:42.970 --> 00:23:46.750 support who's directly in contact with us, our schools, 365 00:23:47.370 --> 00:23:51.210 that use it and the partners to really get the first 366 00:23:51.210 --> 00:23:54.429 hand impression of what is 367 00:23:55.485 --> 00:23:58.065 working well and whatnot, and they are 368 00:23:59.565 --> 00:24:02.465 very much integrated into our development process. 369 00:24:03.485 --> 00:24:06.925 And we take we took every feedback very 370 00:24:06.925 --> 00:24:10.520 seriously. And, yeah, now a 371 00:24:10.520 --> 00:24:14.220 year after we've launched, I am confident to say that 372 00:24:14.280 --> 00:24:17.880 we have a very unique, innovative, and highly 373 00:24:17.880 --> 00:24:21.580 effective, reading promotion solution 374 00:24:21.720 --> 00:24:25.495 that is, yeah, that we can 375 00:24:25.495 --> 00:24:29.115 be really proud of. I see. 376 00:24:29.975 --> 00:24:33.815 The AI world is developing pretty rapidly. You 377 00:24:33.815 --> 00:24:37.520 are right now, I would say, on the cutting edge of development. 378 00:24:38.380 --> 00:24:42.220 How do you make sure that you remain 379 00:24:42.220 --> 00:24:46.060 there as 1 of the top solutions keeping up with the with 380 00:24:46.060 --> 00:24:48.880 the AI developments? We 381 00:24:49.414 --> 00:24:53.255 thanks, for that feedback. Yeah. And we and 382 00:24:53.255 --> 00:24:56.615 I, we really work hard to be seen this 383 00:24:56.615 --> 00:25:00.375 way. So what we do is we invest a lot a 384 00:25:00.375 --> 00:25:03.735 lot into r and d. Most of our money goes into r and 385 00:25:03.735 --> 00:25:07.000 d. We publish papers. We participate 386 00:25:07.220 --> 00:25:11.000 in international conferences, where we 387 00:25:12.020 --> 00:25:13.880 also do take over 388 00:25:15.460 --> 00:25:18.435 tasks and compete against other teams 389 00:25:19.315 --> 00:25:23.075 in optimizing models, quantifying models, and 390 00:25:23.075 --> 00:25:26.595 raising accuracy. And we always, come 391 00:25:26.595 --> 00:25:30.135 up on top, also, leaving 392 00:25:30.355 --> 00:25:34.090 huge names behind us. So we, 393 00:25:36.389 --> 00:25:38.809 we regard this as a sport to develop 394 00:25:40.389 --> 00:25:42.409 new methods, overcome, 395 00:25:44.070 --> 00:25:47.304 overcome hurdles 396 00:25:47.684 --> 00:25:51.385 and, yeah, really try to be on the cutting or beating edge 397 00:25:51.605 --> 00:25:54.905 when it comes to sophisticated speech recognition 398 00:25:55.205 --> 00:25:58.905 and NLP solutions. Yes. 399 00:25:59.045 --> 00:26:02.740 Daniel, I'm sure there will be questions on 400 00:26:02.740 --> 00:26:06.580 where are the papers. I do have a few suspects where you mentioned something like 401 00:26:06.580 --> 00:26:10.419 this. You will give me after the official end of the 402 00:26:10.419 --> 00:26:14.179 interview, you will give me, the link, and I'll post it in the show 403 00:26:14.179 --> 00:26:17.895 notes. I do have a few certain suspects that always request something 404 00:26:17.895 --> 00:26:21.515 like this. Yes, Claude from Paris. I'm looking at you. Exactly. 405 00:26:22.455 --> 00:26:25.195 And then they they can dig through it. So, 406 00:26:26.820 --> 00:26:30.340 let us go into the very last part of the 407 00:26:30.340 --> 00:26:33.700 interview because I'm now already bothering you for, like, 408 00:26:33.700 --> 00:26:36.679 almost 40 minutes in this online meeting and, 409 00:26:37.779 --> 00:26:41.065 more than 25 minutes in actual interview. So, 410 00:26:41.684 --> 00:26:44.585 don't worry. There there are only a few more questions left. 411 00:26:45.684 --> 00:26:49.525 I was wondering winning an award like Frankfurt Forward often 412 00:26:49.525 --> 00:26:52.885 reflects strong local support. How has the 413 00:26:52.885 --> 00:26:56.340 Frankfurt ecosystem contributed to your journey so far? 414 00:26:56.340 --> 00:27:00.020 There are a lot. 415 00:27:00.340 --> 00:27:03.559 So the our ecosystem is 416 00:27:03.860 --> 00:27:07.240 very regional. Our investor is regional. Our 417 00:27:07.380 --> 00:27:09.960 network helped us find our first customer, 418 00:27:11.934 --> 00:27:15.615 is from the region. I we get a lot of 419 00:27:15.615 --> 00:27:19.235 recognition and inquiries due to that, price. 420 00:27:19.615 --> 00:27:23.235 So, it's mostly visibility 421 00:27:23.855 --> 00:27:27.500 and also recognition. So it's when you approach somebody 422 00:27:27.720 --> 00:27:31.100 and, he asks you who you are, what you do, and, 423 00:27:32.120 --> 00:27:35.640 you mentioned the start up of the year world from Frankfurt forward, 424 00:27:35.640 --> 00:27:38.860 especially in the region. Everybody stops questioning 425 00:27:39.000 --> 00:27:42.705 whether you are or whether what you are doing is 426 00:27:42.705 --> 00:27:45.445 sound and makes sense. So, 427 00:27:46.865 --> 00:27:50.644 the intro and entrees into conversation building partnerships 428 00:27:50.784 --> 00:27:54.200 is much easier. Mhmm. And 429 00:27:54.200 --> 00:27:56.600 now only 3 more questions left. 430 00:27:57.880 --> 00:28:01.400 You are AI now a leader in a very specialized niche in 431 00:28:01.649 --> 00:28:05.400 AI. But what advice would you offer to other start ups 432 00:28:05.400 --> 00:28:08.955 looking to innovate in the AI and tech space? Some kind 433 00:28:08.955 --> 00:28:12.735 of skills, like processes you have learned so far, 434 00:28:13.434 --> 00:28:17.215 not only taking the, your business ideas from influencers 435 00:28:17.355 --> 00:28:20.815 on Instagram. Yeah. 436 00:28:21.710 --> 00:28:24.670 I don't know if I'm the right 1 to give advice. Yeah. I'm not, 437 00:28:25.310 --> 00:28:29.070 the tech guy in my company. But what I would generally suggest, when 438 00:28:29.070 --> 00:28:32.850 looking at tech is not looking at a hype or the technology, 439 00:28:33.070 --> 00:28:36.290 but rather trying to solve real world problems. 440 00:28:37.995 --> 00:28:41.695 So AI in Germany, we have 25% of the children, 441 00:28:42.475 --> 00:28:45.995 that at the age of 15 do not understand what they 442 00:28:45.995 --> 00:28:49.675 read, and they don't understand what they read because they are not fluent 443 00:28:49.675 --> 00:28:53.180 in reading. So that's a huge societal 444 00:28:53.320 --> 00:28:57.000 problem, and that's, that was the initial thing 445 00:28:57.000 --> 00:28:59.900 that led me to found it found DigiZapiens. 446 00:29:02.200 --> 00:29:05.955 And I would really not focus on the 447 00:29:05.955 --> 00:29:09.635 technology or the AI for the purpose of the technology in 448 00:29:09.635 --> 00:29:13.235 AI, but rather using tech as a 449 00:29:13.235 --> 00:29:16.835 means to solve a real problem. Yeah. And if it takes 450 00:29:16.835 --> 00:29:18.934 AI, it's fine. If not, not. 451 00:29:20.720 --> 00:29:24.080 But 1 also has to say in German, there can 452 00:29:24.080 --> 00:29:27.540 be very difficult, 453 00:29:28.640 --> 00:29:31.824 sentences and structures. III 454 00:29:32.160 --> 00:29:35.845 everybody who's who tried either in German or in a 455 00:29:35.845 --> 00:29:39.365 translated version to read Kafka should know what we're 456 00:29:39.365 --> 00:29:43.125 talking about here. So, therefore, it 457 00:29:43.125 --> 00:29:45.705 can be difficult. There's a saying in German, 458 00:29:47.125 --> 00:29:50.220 German language, difficult language. So, actually, 459 00:29:51.240 --> 00:29:54.919 it it it says good things about you guys that you started 460 00:29:54.919 --> 00:29:58.760 with the German language and mastered it for your, for 461 00:29:58.760 --> 00:30:00.940 your, tool, DigiSapiens. 462 00:30:03.025 --> 00:30:06.865 We usually close out with 2 more questions, and they're usually 463 00:30:06.865 --> 00:30:10.705 pretty simple and usually end with a yes. But I'll ask them 464 00:30:10.705 --> 00:30:13.285 anyway. Are you open to talk to new investors? 465 00:30:16.940 --> 00:30:20.700 And as always, I'll link your LinkedIn profile down here in the show 466 00:30:20.700 --> 00:30:24.540 notes wherever you're looking this wherever you're watching this. No. 467 00:30:24.540 --> 00:30:27.760 Sorry. This AI, no watching. But, 468 00:30:28.155 --> 00:30:31.995 anyways, you you either directly in your tool. I'm 469 00:30:31.995 --> 00:30:35.835 sorry. Not every tool allows links you can then click. 470 00:30:35.835 --> 00:30:39.675 So, basically, you could go to our blog, standard break. 471 00:30:39.675 --> 00:30:43.215 Ioforward/block, and there we link Daniel's 472 00:30:43.390 --> 00:30:47.150 profile. Plus, when you are 473 00:30:47.150 --> 00:30:50.750 expanding, when you're growing as a young company, I am sure you're 474 00:30:50.750 --> 00:30:54.270 all so open to have applications from potential new 475 00:30:54.270 --> 00:30:57.870 employees. Right? Yes. Yes. Yes. Yes. Yes. We're 476 00:30:57.870 --> 00:31:00.265 looking for we're looking for good people always. 477 00:31:01.985 --> 00:31:05.765 AI, is there a career website that I could launch, 478 00:31:05.765 --> 00:31:09.365 or should the people simply, that I could link, or should the 479 00:31:09.365 --> 00:31:12.665 people, simply reach out to you via email? 480 00:31:13.340 --> 00:31:16.960 Yeah. The latter. Directly reach out to us. We, 481 00:31:17.740 --> 00:31:21.420 we could, we we have a way to go regarding an 482 00:31:21.420 --> 00:31:25.180 HR department, so everything's handled by the team, depending 483 00:31:25.180 --> 00:31:28.715 on the competency somebody's applying for. So 484 00:31:28.934 --> 00:31:32.775 can't use the public channels or directly contact me, and I will forward 485 00:31:32.775 --> 00:31:36.375 it to the colleagues. Again, go to the blog to the LinkedIn 486 00:31:36.375 --> 00:31:39.895 profile. Guys, it was a 487 00:31:39.895 --> 00:31:43.539 pleasure talking to you, Daniel. Thank you very much. Thank you very much for 488 00:31:43.539 --> 00:31:47.000 answering more than 30 minutes difficult questions here and, 489 00:31:48.100 --> 00:31:51.860 keeping up my stupid interjection interruptions. Thank you very 490 00:31:51.860 --> 00:31:55.595 much. Thank you. Nice being here. 491 00:32:00.775 --> 00:32:04.535 Yeah. So and yeah. Thanks for having me here, and, yeah, 492 00:32:04.535 --> 00:32:07.355 Merry Christmas and a happy New Year to everyone. 493 00:32:09.672 --> 00:32:13.512 Thank you. Yeah. Merry Christmas. Happy New Year from you as well. 494 00:32:13.512 --> 00:32:14.892 Thank you, guys. Bye bye.