WEBVTT 1 00:00:00.240 --> 00:00:03.680 Maximum pain relief. I use it all the time because 2 00:00:03.680 --> 00:00:07.200 I'm always active, playing golf, working out, fixing up my 3 00:00:07.280 --> 00:00:09.800 place right now, and I put it on in the 4 00:00:09.839 --> 00:00:14.599 evening around my back and it gives me maximum pain relief. 5 00:00:15.119 --> 00:00:16.039 It's amazing. 6 00:00:16.280 --> 00:00:18.000 You have to get it if you're in pain. 7 00:00:18.359 --> 00:00:21.800 To find out more about ice bod, go to icebodactive 8 00:00:21.839 --> 00:00:25.039 dot com and get yours today. This week they are 9 00:00:25.039 --> 00:00:27.440 having a flash sale where you can save twenty five 10 00:00:27.480 --> 00:00:31.760 percent off by using the promo code KCAA, plus they 11 00:00:31.800 --> 00:00:35.119 have free shipping. Go to ice bodactive dot com. You're 12 00:00:35.159 --> 00:00:37.479 gonna be pain free and you're gonna love it. 13 00:00:37.799 --> 00:00:41.840 Kc AA Radio, Loma Linda, where no listener is ever 14 00:00:42.079 --> 00:00:42.840 left behind. 15 00:00:43.759 --> 00:00:47.359 The information economy has a ride. The world is teeming 16 00:00:47.399 --> 00:00:51.600 with innovation as new business models reinvent every industry industry. 17 00:00:51.679 --> 00:00:56.039 Inside Analysis is your source of information and insight about 18 00:00:56.039 --> 00:00:58.600 how to make the most of this exciting new era. 19 00:00:58.840 --> 00:01:03.000 Learn more and insight Analysis dot Comside Analysis dot com. 20 00:01:03.000 --> 00:01:05.920 And now here's your host, through Eric Kavanaugh. 21 00:01:10.560 --> 00:01:12.959 All Right, folks, hello, and welcome back once again to 22 00:01:13.040 --> 00:01:17.079 the only nationally syndicated radio show all about the information economy. 23 00:01:17.120 --> 00:01:20.000 It's called Inside Analysis or truly Eric Kavanaugh is here 24 00:01:20.200 --> 00:01:23.519 and I'm very pleased to have Arun Baradarajohn. He is 25 00:01:23.640 --> 00:01:27.640 the founder and chief commercial officer for a company called Assendian, 26 00:01:28.040 --> 00:01:31.120 and we're going to talk about a new development at Paradigm, 27 00:01:31.480 --> 00:01:34.599 So just to give some context for that. As anyone 28 00:01:34.640 --> 00:01:38.120 who has worked with professional services firms know, they typically 29 00:01:38.200 --> 00:01:40.079 make their money by the hour, and if you want 30 00:01:40.079 --> 00:01:42.120 to do something new, they're like, oh, we need more people, 31 00:01:42.519 --> 00:01:44.959 more money, got to check out the price, so we'll 32 00:01:44.959 --> 00:01:47.599 make more money and give you new stuff. And then 33 00:01:47.599 --> 00:01:50.400 there's the other side of the equation, which is that 34 00:01:50.599 --> 00:01:52.359 a lot of times if you want to go faster 35 00:01:52.560 --> 00:01:54.760 or for less money, they're like, well, it's going to 36 00:01:54.799 --> 00:01:56.599 cost you in terms of quality, So you have to 37 00:01:56.599 --> 00:02:01.799 give up something either quality or price, money or ideas, 38 00:02:01.840 --> 00:02:03.920 like even don't even start your project, as maybe you 39 00:02:03.920 --> 00:02:06.719 can't afford it. And I think Arun has come up 40 00:02:06.719 --> 00:02:09.840 with a pretty clever way to address that, and it's 41 00:02:09.879 --> 00:02:12.159 an end to end engineering platform. 42 00:02:12.560 --> 00:02:13.479 They figured out. 43 00:02:13.280 --> 00:02:15.360 There are all these friction points one hundred and fifty 44 00:02:15.360 --> 00:02:17.840 I think is what I heard him say, friction points 45 00:02:17.960 --> 00:02:21.680 from idea to production. So what are those and how 46 00:02:21.680 --> 00:02:24.159 do you solve them while we live in a very 47 00:02:24.199 --> 00:02:27.520 interesting world these days where you can have teams work 48 00:02:27.599 --> 00:02:31.000 together on projects without holding each other up. I give 49 00:02:31.039 --> 00:02:34.800 an example all the time about how Google blindsided Microsoft 50 00:02:34.919 --> 00:02:37.599 with Google Docs allowing multiple people to work on the 51 00:02:37.639 --> 00:02:41.000 same document at the same time, which was an absolutely 52 00:02:41.120 --> 00:02:44.960 brilliant innovation. Saves tremendous amounts of time and effort. You 53 00:02:45.000 --> 00:02:46.840 don't have to throw things over the wall anymore. You 54 00:02:46.840 --> 00:02:49.960 can also be looking at the same document, typing, editing, 55 00:02:50.400 --> 00:02:53.159 and the technology tracks who did what. So you have 56 00:02:53.199 --> 00:02:57.879 all these wonderful guardrails and safeguards built into the technology, 57 00:02:57.919 --> 00:02:59.879 and that's what you want. And I think they have 58 00:03:00.120 --> 00:03:02.360 something like that for a development platform. But with that, 59 00:03:02.479 --> 00:03:04.039 I run, I'll throw it over to you. Tell us 60 00:03:04.080 --> 00:03:07.000 a bit about ascending and what are these friction points? 61 00:03:07.479 --> 00:03:11.520 Absolutely so so, Eric, The Cindy On journey started about 62 00:03:11.520 --> 00:03:15.479 four years ago, and we had a group of us 63 00:03:15.560 --> 00:03:20.159 who came from the the overall engineering and software industry, 64 00:03:20.800 --> 00:03:24.240 and we had realized that there were fundamentally three issues 65 00:03:24.280 --> 00:03:27.520 that clients were facing. There was there was a crisis 66 00:03:27.560 --> 00:03:31.719 of trust, there was a crisis of speed, and what 67 00:03:31.800 --> 00:03:33.360 I call is a crisis of capital. 68 00:03:34.520 --> 00:03:36.199 The crisis of trust was you. 69 00:03:36.120 --> 00:03:38.719 Know what, these guys will show up at my door 70 00:03:38.759 --> 00:03:40.800 and say, hey, don't worry, I'll get this done for you. 71 00:03:41.360 --> 00:03:42.280 This will be the cost. 72 00:03:42.360 --> 00:03:45.319 I've got these accelerators, I've got these really good talent. 73 00:03:45.439 --> 00:03:48.000 I can get this work done for you. And then 74 00:03:48.039 --> 00:03:49.400 they show up at the door and say that. 75 00:03:49.479 --> 00:03:53.280 And then when the reality you know what do I 76 00:03:53.319 --> 00:03:56.719 say manifests, they find that, oh my god, I'm being 77 00:03:56.800 --> 00:03:58.120 charged more than what I was told. 78 00:03:58.360 --> 00:03:59.400 You know, I'm not. 79 00:03:59.520 --> 00:04:03.560 I don't have full transparency into how things are being done. 80 00:04:03.840 --> 00:04:06.199 I don't know what your engineers are doing. It's a 81 00:04:06.280 --> 00:04:09.520 black box. I have no clue. The second is a 82 00:04:09.560 --> 00:04:12.080 crisis of speed. Everybody comes and says, yeah, we will 83 00:04:12.159 --> 00:04:15.080 accelerate you. We will drive more more value to you. 84 00:04:15.520 --> 00:04:18.360 And then lo behold, when you say I want to accelerate, 85 00:04:19.000 --> 00:04:21.000 the first thing the service provider says is you need 86 00:04:21.040 --> 00:04:24.279 to add another hundred people right that cost. 87 00:04:24.560 --> 00:04:25.319 And then listen. 88 00:04:26.439 --> 00:04:28.439 One thing I can tell you is we add more 89 00:04:28.480 --> 00:04:33.319 people into a software engineering project, it actually slows you down. 90 00:04:33.360 --> 00:04:33.879 It doesn't. 91 00:04:34.040 --> 00:04:36.959 It doesn't increase your speed because you add more humans 92 00:04:36.959 --> 00:04:40.959 into the software process, which is very it's very nebulous, 93 00:04:41.519 --> 00:04:44.360 not superly defined. You will find that there are more 94 00:04:44.439 --> 00:04:47.879 errors creeping in and you're busy fixing those errors. Right, Yeah, 95 00:04:48.000 --> 00:04:50.959 So it's a it's actually a it's actually a negative 96 00:04:51.120 --> 00:04:54.160 move to add more people to projects. The third verse 97 00:04:54.199 --> 00:04:57.920 crisis of capital. Now, what is the crisis of capital? 98 00:04:58.399 --> 00:05:03.480 Over the years, because of the first two issues, clients 99 00:05:03.480 --> 00:05:08.199 have been building software with a lot of technical problems, 100 00:05:08.199 --> 00:05:09.759 and we call it technical. 101 00:05:09.360 --> 00:05:11.560 Debt, a lot of debt. 102 00:05:12.399 --> 00:05:15.959 So what happens is your capital is locked on these 103 00:05:16.079 --> 00:05:20.000 legacy old code and you're spending a lot of money 104 00:05:20.079 --> 00:05:23.279 just maintaining it, and you don't have enough money to 105 00:05:23.360 --> 00:05:24.680 do new things and innovation. 106 00:05:25.199 --> 00:05:26.920 And I call that the crisis a capital. 107 00:05:27.800 --> 00:05:30.600 So all this was creating, in my opinion, a logjam 108 00:05:30.639 --> 00:05:33.959 and a guardian not as I call it. Okay, which 109 00:05:34.000 --> 00:05:36.680 the cord is yes, And I said, it's time to 110 00:05:36.759 --> 00:05:41.240 untangle this. So the thesis statement for Cyndion was, let's 111 00:05:41.240 --> 00:05:44.480 go back and understand what is causing this problem. And 112 00:05:44.519 --> 00:05:46.920 to me, I went back to the manufacturing world, and 113 00:05:46.959 --> 00:05:51.120 I saw how lean manufacturing transformed the entire manufacturing world. 114 00:05:51.759 --> 00:05:52.040 Right. 115 00:05:52.319 --> 00:05:54.519 Otherwise, earlier, if you had to do a line change, 116 00:05:54.600 --> 00:05:56.519 or a die change or a product change. 117 00:05:56.600 --> 00:05:58.519 It took you days to do it. 118 00:05:59.079 --> 00:06:01.439 You had issues with product quality, you had a whole 119 00:06:01.519 --> 00:06:05.720 bunch of things with lean manufacturing, where automation and to 120 00:06:05.759 --> 00:06:10.399 a large extent, AI fundamentally transform manufacturing. And I said, 121 00:06:10.399 --> 00:06:12.720 if manufacturing can do it, why are we lagging behind? 122 00:06:13.319 --> 00:06:15.959 Why can't we do the same thing with software. So 123 00:06:16.120 --> 00:06:19.040 we went back to the drawing board and really looked 124 00:06:19.040 --> 00:06:23.759 at the entire engineering value chain, right from ideation through 125 00:06:23.800 --> 00:06:27.560 to production, through to even post production, and said, what 126 00:06:27.680 --> 00:06:30.959 is causing these what is causing this friction and problem? 127 00:06:31.000 --> 00:06:33.319 And to a large extent, I'm sorry to say that 128 00:06:33.439 --> 00:06:34.720 it is the humans. 129 00:06:35.319 --> 00:06:35.600 Right. 130 00:06:36.000 --> 00:06:39.839 If you give a human, if you give four humans 131 00:06:40.639 --> 00:06:44.560 one requirement saying hey, I want to do this very 132 00:06:44.560 --> 00:06:47.319 simple requirement. It could be a very two to three 133 00:06:47.319 --> 00:06:50.319 lines of description of what you want the software to do, 134 00:06:50.360 --> 00:06:53.079 and you give it to four different engineers, you will 135 00:06:53.079 --> 00:06:55.279 get four different code. 136 00:06:55.040 --> 00:06:57.120 Bases for the requirement. 137 00:06:57.519 --> 00:07:01.800 Sure you will also get ten different bugs that each 138 00:07:01.839 --> 00:07:05.560 of them will will come up with. And I said, 139 00:07:05.639 --> 00:07:08.720 this is ridiculous. It's time for us to change. So 140 00:07:08.759 --> 00:07:12.199 that's when Sindion we was started with the notion that we 141 00:07:12.240 --> 00:07:16.120 are going to build a platform. The platform is fundamentally 142 00:07:16.160 --> 00:07:20.000 going to have the ability to support the different actors 143 00:07:20.240 --> 00:07:23.759 in the engineering value chain, use machine learning and AI 144 00:07:24.000 --> 00:07:29.360 and try high degree of repeatability, standardization, and really get 145 00:07:29.399 --> 00:07:32.720 the process. See if you look at the software process today, 146 00:07:33.199 --> 00:07:36.959 it is all in documents. With the platform, what we're 147 00:07:37.000 --> 00:07:40.519 doing is we're taking the process and instantiating it into 148 00:07:40.560 --> 00:07:46.000 a physical platform that will drive the process and ensure 149 00:07:46.000 --> 00:07:49.120 that there is standardization and scale. That is what we 150 00:07:49.160 --> 00:07:53.000 have and thanks to all the AI work that has 151 00:07:53.040 --> 00:07:55.759 been going on over the last two to three years, 152 00:07:56.319 --> 00:07:58.639 we've been able to take this to a different level 153 00:07:58.680 --> 00:07:59.959 and I'm happy to talk more about it. 154 00:08:01.439 --> 00:08:01.680 Yeah. 155 00:08:01.720 --> 00:08:05.920 Well, so standard components, that's one thing you can focus on, right, 156 00:08:06.000 --> 00:08:09.959 making making solutions modular. Maybe walk through what are some 157 00:08:10.040 --> 00:08:14.639 of the foundational components in the platform and maybe explain 158 00:08:15.079 --> 00:08:18.360 how it is. I'm just guessing here that developers can 159 00:08:18.399 --> 00:08:21.399 work on different parts of the process without causing each 160 00:08:21.399 --> 00:08:24.199 other trouble and without delaying each other, right, because that's 161 00:08:24.240 --> 00:08:27.759 one of the huge challenges is when one group has 162 00:08:27.800 --> 00:08:30.600 to wait for some other group to finish something and 163 00:08:30.639 --> 00:08:32.759 then they get delayed. So now these guys are off 164 00:08:33.039 --> 00:08:35.799 the ranch basically just twiddling their thumbs. 165 00:08:36.039 --> 00:08:36.600 You don't want that. 166 00:08:36.639 --> 00:08:39.360 You want everybody working at the same time without disrupting 167 00:08:39.399 --> 00:08:42.480 each other. What are some of the component parts or 168 00:08:43.120 --> 00:08:46.759 buckets if you will, of development that you've isolated. 169 00:08:47.559 --> 00:08:51.759 So we have gone even more radical, right, So our 170 00:08:51.879 --> 00:08:55.679 radical thinking is that why do we even need humans 171 00:08:55.720 --> 00:08:59.080 to actually do any of this? And we are thinking 172 00:08:59.120 --> 00:09:02.360 differently actually saying that. So if you look at our industry, 173 00:09:02.360 --> 00:09:06.039 we do different things for our lives. So sometimes we'll 174 00:09:06.039 --> 00:09:08.440 come in and build a new platform for a client 175 00:09:08.480 --> 00:09:10.840 where we have to take it from idea to production. 176 00:09:11.039 --> 00:09:12.000 As you mentioned. 177 00:09:11.759 --> 00:09:13.759 Earlier, all the client will say is and I have 178 00:09:13.799 --> 00:09:17.240 an old platform. I want you to modernize it. You know, 179 00:09:17.559 --> 00:09:20.240 the old platform could be written on something like Cobol 180 00:09:20.360 --> 00:09:24.039 or pl one, which is forty fifty years old old language, 181 00:09:24.399 --> 00:09:26.639 and they want to move it to a modern construct. 182 00:09:26.799 --> 00:09:31.240 So there are several such asks from our lives and 183 00:09:31.360 --> 00:09:34.159 my view was why can't I get AI to do 184 00:09:34.200 --> 00:09:36.919 all of this? So today, if you look at our platform, 185 00:09:37.360 --> 00:09:39.799 our platform is what is known as an agent tick 186 00:09:39.919 --> 00:09:44.240 AI platform, So I literally have an agent that can 187 00:09:44.320 --> 00:09:48.080 do anything you wanted to do, and I can configure 188 00:09:48.120 --> 00:09:52.320 this agent to say, okay, agent read this. So a 189 00:09:52.399 --> 00:09:55.519 requirement in our world is called a user story, basically 190 00:09:55.559 --> 00:09:58.399 saying this is the story that you will present to 191 00:09:58.440 --> 00:10:01.120 the user. Right this, this is how the user will 192 00:10:01.200 --> 00:10:03.639 use that use the software. 193 00:10:04.600 --> 00:10:06.960 So the user story is a very important input. 194 00:10:07.600 --> 00:10:10.159 Now I have an agent that will read the user 195 00:10:10.200 --> 00:10:15.080 story and create the code. I have an agent that 196 00:10:15.159 --> 00:10:18.519 will read the user story and create the testing that 197 00:10:18.679 --> 00:10:21.799 needs to be done on the code and execute the testing. 198 00:10:22.600 --> 00:10:22.879 Right. 199 00:10:23.200 --> 00:10:25.200 So what we are doing is and with this, what 200 00:10:25.639 --> 00:10:28.919 we're fundamentally doing, Eric is now when you write a 201 00:10:29.080 --> 00:10:31.600 user story. If I'm a product manager, my job is 202 00:10:31.639 --> 00:10:32.720 to write a user story. 203 00:10:33.039 --> 00:10:34.960 That's my role to write user stories. 204 00:10:35.320 --> 00:10:38.080 But the funny thing is if I again, just like 205 00:10:38.159 --> 00:10:42.039 I mentioned earlier about developers or engineers, if I give 206 00:10:42.159 --> 00:10:44.919 somebody an input to write a user story, and the 207 00:10:44.919 --> 00:10:46.840 input isically typically. 208 00:10:46.440 --> 00:10:47.120 Called an epic. 209 00:10:47.799 --> 00:10:50.000 So if I give you an epic and I say, 210 00:10:50.080 --> 00:10:53.159 please write user stories for this epic, and if I 211 00:10:53.240 --> 00:10:55.679 give it to four different product managers, I will get 212 00:10:55.679 --> 00:10:59.679 the user stories written in four different ways and will 213 00:10:59.720 --> 00:11:03.840 not cover all of the needs of that epic. Because 214 00:11:03.879 --> 00:11:07.480 the human can only think so much. Right, But what 215 00:11:07.519 --> 00:11:09.879 I'm doing with AI is I tell the AI A 216 00:11:10.519 --> 00:11:13.000 here's the epic, generate user stories, and I have live 217 00:11:13.080 --> 00:11:16.840 examples where I've taken I've done this time and motion 218 00:11:17.000 --> 00:11:20.879 study and comparison between what the human output is and 219 00:11:20.879 --> 00:11:23.480 what the AI output is. Right, So when I give 220 00:11:23.480 --> 00:11:25.679 a user, when I give an EPIC to one of 221 00:11:25.720 --> 00:11:28.759 my AI agents and say generate user stories, it may 222 00:11:29.360 --> 00:11:33.240 generate fourteen fifteen user stories that cover all of the 223 00:11:33.279 --> 00:11:36.000 aspects of what the system needs to do in the 224 00:11:36.039 --> 00:11:40.399 first iteration, Whereas if I give that to a human, 225 00:11:41.120 --> 00:11:43.440 I sometimes see them coming up with seven or eight. 226 00:11:43.440 --> 00:11:47.039 They're not able to think about the edge, the edge conditions, 227 00:11:47.120 --> 00:11:50.519 the limit. You know, when you think about software, you've 228 00:11:50.559 --> 00:11:53.120 got to think about all the eventualities and all the 229 00:11:53.159 --> 00:11:55.639 possibilities that you may have to encounter as a user, 230 00:11:56.360 --> 00:12:00.480 and many times as a human mind, we have these limitations. Right, 231 00:12:01.159 --> 00:12:03.919 that gets taken away, and then I can do this 232 00:12:04.000 --> 00:12:08.080 at scale instead of having you know, each time, let's 233 00:12:08.080 --> 00:12:10.120 say I need I have tons of epics, and I 234 00:12:10.159 --> 00:12:12.960 need to attack all these epics. I need to add 235 00:12:12.960 --> 00:12:16.879 more product managers. Right here, I just add more agents, 236 00:12:18.080 --> 00:12:20.600 and I still need product managers because I will have 237 00:12:20.639 --> 00:12:24.919 them eyeball what the agent creates, taking the human away 238 00:12:24.960 --> 00:12:27.759 from the loop. But I'm fundamentally saying, hey, if you're 239 00:12:28.039 --> 00:12:31.039 if I had ten product managers doing the job, maybe 240 00:12:31.039 --> 00:12:31.759 I just need. 241 00:12:31.600 --> 00:12:33.720 To m hm and more. 242 00:12:34.480 --> 00:12:36.840 These product mats are really. 243 00:12:36.679 --> 00:12:40.759 Improving the improving the agent's ability and efficacy to do 244 00:12:40.799 --> 00:12:41.919 this even better. 245 00:12:43.000 --> 00:12:45.360 So if you don't mind my asking the the agents 246 00:12:45.399 --> 00:12:49.000 that you've designed, how what what language did you. 247 00:12:49.039 --> 00:12:50.039 Use to design them? 248 00:12:50.200 --> 00:12:50.879 How do they run? 249 00:12:51.120 --> 00:12:54.279 They do they run in containers? Is this a containerized environment? 250 00:12:54.320 --> 00:12:55.159 Tell me a bit about that. 251 00:12:56.000 --> 00:13:00.759 So our platform is completely containerized. Okay, follow all the 252 00:13:00.799 --> 00:13:05.120 modern methods. It's a platform. So our platform allows you 253 00:13:05.240 --> 00:13:10.279 to create agents, and I can create an agent for 254 00:13:10.559 --> 00:13:14.360 different use cases. So somebody may come to me, Like 255 00:13:14.399 --> 00:13:16.080 one of my clients came to me the other day 256 00:13:16.120 --> 00:13:21.159 and said, listen, I've got these Pearl scripts. Now peerl 257 00:13:21.279 --> 00:13:24.000 is again one of the old scripting languages that nobody 258 00:13:24.080 --> 00:13:28.360 uses anyone, right, So he said, I've got Pearl scripts. 259 00:13:28.919 --> 00:13:33.080 I really want to convert them into Java. And he said, 260 00:13:33.120 --> 00:13:35.039 I'm trying to get my engineers to do it. It's 261 00:13:35.080 --> 00:13:37.120 taking way too much time. And I've got tons of 262 00:13:37.159 --> 00:13:40.799 Pearl scripts. I've got these four hundred applications running in 263 00:13:40.840 --> 00:13:42.440 my data center and I want to move them to 264 00:13:42.440 --> 00:13:44.440 the cloud, and I can't really use PELS scripts if 265 00:13:44.440 --> 00:13:46.279 I move into the cloud. I want to move it 266 00:13:46.320 --> 00:13:49.639 to the new language Java. In the old days, what 267 00:13:50.159 --> 00:13:51.559 we would have done is we would have tried to 268 00:13:51.639 --> 00:13:54.759 use these code converters that will take the peerl scripts 269 00:13:54.799 --> 00:14:00.559 and converted into whatever language that is not working for 270 00:14:00.679 --> 00:14:03.679 us anymore because we want to move to micro services 271 00:14:04.159 --> 00:14:08.919 and containized component based applications. What we're doing is we 272 00:14:08.919 --> 00:14:11.720 have changed the process. So I have an agent which 273 00:14:11.759 --> 00:14:15.240 goes and interestingly I put two of my principal engineers 274 00:14:15.240 --> 00:14:17.799 with our platform and we were able to get this 275 00:14:17.919 --> 00:14:20.799 done in weeks for our clients who actually took months 276 00:14:20.840 --> 00:14:23.279 to try and do this. So what we did was 277 00:14:23.799 --> 00:14:26.879 we had what is known as a reverse engineering agent 278 00:14:27.039 --> 00:14:31.720 that came in and read the Pearl scripts and understood. 279 00:14:31.879 --> 00:14:34.279 And these Pearl scripts can be very messy. They can 280 00:14:34.320 --> 00:14:37.240 have routines. I mean, I'm not going to get super 281 00:14:37.320 --> 00:14:40.159 I don't know how technical I can get here, but fundamentally, 282 00:14:40.360 --> 00:14:43.279 when you have these scripts, these they call each other 283 00:14:43.320 --> 00:14:47.399 there's a lot of linkages here and there. Right, So 284 00:14:47.480 --> 00:14:50.320 we actually put an agent that went and read all 285 00:14:50.360 --> 00:14:54.279 of this and generated what we call is a simple 286 00:14:54.840 --> 00:14:58.799 ascendion language, that's what we call it, which is just 287 00:14:58.960 --> 00:15:02.840 English and nim on that describes the Perl script's logic. 288 00:15:04.519 --> 00:15:07.240 And then I take that, and then I get another 289 00:15:07.279 --> 00:15:10.720 agent to come in and convert that language, and I 290 00:15:11.240 --> 00:15:13.720 go and tell the client, does this logic make sense 291 00:15:13.759 --> 00:15:17.559 to you? Right? Forget about the Pearl scripts, none of 292 00:15:17.639 --> 00:15:20.440 us know how to read it. But does the logic 293 00:15:20.480 --> 00:15:22.320 of what is getting done in the Perl script make. 294 00:15:22.240 --> 00:15:22.679 Sense to you? 295 00:15:23.600 --> 00:15:26.240 Says yeah, this part makes sense. This one I don't 296 00:15:26.279 --> 00:15:31.159 even think I use. This part looks good. So I'm 297 00:15:31.200 --> 00:15:33.360 able to also say, you know what, I don't even 298 00:15:33.399 --> 00:15:37.000 need to waste money converting these Pearl scripts because they. 299 00:15:36.879 --> 00:15:37.759 Are no longer used. 300 00:15:39.159 --> 00:15:43.080 So I'm now a document that is super efficient that 301 00:15:43.120 --> 00:15:47.039 I reverse engineered, so I can now forward engineer it 302 00:15:47.120 --> 00:15:50.000 whichever way you want. So I can take this document 303 00:15:50.200 --> 00:15:52.840 convert them again into what I call as user stories, 304 00:15:52.919 --> 00:15:55.559 which is the input, as I said, the requirement input 305 00:15:55.639 --> 00:15:59.840 for development. And once I create the user stories, I 306 00:16:00.080 --> 00:16:02.679 can I have an agent that converts one user's story 307 00:16:02.720 --> 00:16:06.399 into Java. I can have one agent that can convert 308 00:16:06.440 --> 00:16:09.919 it into c SHAP or whatever you want, and moving 309 00:16:10.320 --> 00:16:13.440 completely into micro services architecture. And while I'm doing that, 310 00:16:13.799 --> 00:16:17.480 in parallel, I have agents that can create the test 311 00:16:18.320 --> 00:16:22.960 scripts and I can automate the testing too. So with 312 00:16:23.679 --> 00:16:26.799 what is happening, Eric is, I'm my teams and I 313 00:16:27.000 --> 00:16:29.519 we are completely reimagining how. 314 00:16:29.440 --> 00:16:34.360 Work gets done right. There's no need for rework. Now. 315 00:16:34.480 --> 00:16:35.519 The biggest I'll. 316 00:16:35.320 --> 00:16:41.519 Tell you thirty person of project costs is rework because 317 00:16:42.200 --> 00:16:46.720 the developer didn't understand the user story. The used story 318 00:16:46.879 --> 00:16:50.679 was not written correctly because the product manager did not 319 00:16:50.840 --> 00:16:53.879 understand the user needs and did not elaborate the scenario. 320 00:16:53.879 --> 00:16:58.679 As well, the testers did not test the software correctly. 321 00:16:59.440 --> 00:17:02.799 These are all all the pitfalls we are eliminating by 322 00:17:03.720 --> 00:17:09.119 running these agents and bringing in high degree of standardization 323 00:17:09.200 --> 00:17:09.680 and scale. 324 00:17:11.519 --> 00:17:15.559 That's very interesting. You know, I understand conceptually what you're 325 00:17:15.599 --> 00:17:19.400 talking about. Did you use jen Ai for this reverse engineering? 326 00:17:19.480 --> 00:17:22.480 Is that how you loaded a Pearl script? You said, hey, 327 00:17:22.559 --> 00:17:24.960 what is this thing actually doing? Was that a Genai application? 328 00:17:24.960 --> 00:17:26.720 We got one minute left in this segment, Go ahead, 329 00:17:27.400 --> 00:17:28.240 we use. 330 00:17:28.160 --> 00:17:31.759 A combination of GENI and other AI. So we use 331 00:17:31.920 --> 00:17:35.079 machine learning, we use you know, deep learning, and we 332 00:17:35.200 --> 00:17:38.960 use GENI depending on the type of problems we're. 333 00:17:38.759 --> 00:17:39.359 Trying to solve. 334 00:17:39.400 --> 00:17:41.400 With GENAI, we've been able to take this to the 335 00:17:41.440 --> 00:17:41.960 next level. 336 00:17:42.400 --> 00:17:45.079 Mm hmm, well I can imagine. I mean I remember 337 00:17:45.160 --> 00:17:47.279 playing around with it whenever it came out a year 338 00:17:47.279 --> 00:17:50.039 and a half ago or so, and immediately drawing the 339 00:17:50.079 --> 00:17:52.319 conclusion that if it can write in French and German 340 00:17:52.400 --> 00:17:54.799 and Spanish and English, that you can probably write in 341 00:17:54.839 --> 00:17:58.880 cobal and C sharp and C plus plus and Java 342 00:17:59.279 --> 00:18:02.480 and these other languages. And guess what it can. You know, 343 00:18:03.240 --> 00:18:06.039 my understanding, having researched this a good bit is that 344 00:18:06.079 --> 00:18:09.000 it'll get you eighty percent of the way there typically, 345 00:18:09.039 --> 00:18:11.319 and then the last twenty percent is the mile you 346 00:18:11.400 --> 00:18:14.839 have to do fine tuning, manually checking things. 347 00:18:14.880 --> 00:18:15.880 You do want to make. 348 00:18:15.759 --> 00:18:18.039 Sure that, as you've already suggested, you have a human 349 00:18:18.079 --> 00:18:21.960 being monitoring things, making sure that the outputs are accurate. 350 00:18:22.039 --> 00:18:24.559 But the beautiful thing about software that works is that 351 00:18:24.759 --> 00:18:27.279 it works, and if it doesn't work, then it doesn't work. 352 00:18:27.319 --> 00:18:30.319 And you know very clearly in a binary fashion, is 353 00:18:30.319 --> 00:18:33.279 this accomplishing the task or not. And if it's a no, 354 00:18:33.400 --> 00:18:34.839 you have to go back to the drawing board. But 355 00:18:35.200 --> 00:18:37.279 this very interesting stuff, well, don't touch one, folks. We'll 356 00:18:37.279 --> 00:18:40.640 be right back talking with ascending about a new way 357 00:18:40.799 --> 00:18:44.240 of doing software. It's a lot faster and probably a 358 00:18:44.240 --> 00:18:46.359 lot more efficient. Will get into the details in our 359 00:18:46.400 --> 00:18:46.799 next segment. 360 00:18:46.839 --> 00:18:47.640 We'll be right back. 361 00:18:47.680 --> 00:18:48.960 You're listening to Inside Analysis. 362 00:18:49.359 --> 00:18:49.880 React. 363 00:18:55.519 --> 00:19:00.400 Welcome back to Inside Analysis. Here's your host, Erica. 364 00:19:03.039 --> 00:19:05.119 All right, folks, back here on Inside Analysis talking to 365 00:19:05.319 --> 00:19:09.319 Ruin Varadarajan. He is the CCO and founder of a 366 00:19:09.359 --> 00:19:12.200 company called Ascendian. It's just like it sounds ascend with 367 00:19:12.279 --> 00:19:14.960 io n at the end. And they have developed a 368 00:19:14.960 --> 00:19:18.160 do a platform for software development that is agentic. They 369 00:19:18.160 --> 00:19:20.759 have all sorts of AI agents out there. 370 00:19:20.799 --> 00:19:21.039 Now. 371 00:19:21.160 --> 00:19:23.920 This is the talk of the town in Silicon Valley 372 00:19:23.960 --> 00:19:26.960 and around the world quite frankly, because if a machine 373 00:19:27.000 --> 00:19:29.519 can do a job better than a human, let the 374 00:19:29.559 --> 00:19:32.240 machine do the job. I mean, humans make mistakes. The 375 00:19:32.359 --> 00:19:36.440 machines make mistakes too, Jennai we've talked about famously makes mistakes. 376 00:19:36.440 --> 00:19:40.119 It generates things, so you have to do all sorts 377 00:19:40.160 --> 00:19:43.920 of work to train it and to ground it. Basically, 378 00:19:44.319 --> 00:19:46.039 but that's not really what we're talking about here with 379 00:19:46.079 --> 00:19:48.720 this agentic stuff, though, you still want to monitor what 380 00:19:48.759 --> 00:19:52.039 they do and guide them. And frankly, this is my 381 00:19:52.160 --> 00:19:56.240 big question about agentic AI is how do we monitor 382 00:19:56.279 --> 00:19:59.799 their behavior, correct their behavior when they're wrong? How do 383 00:19:59.799 --> 00:20:02.039 we orchestrate their behavior? Because you've got a bunch of 384 00:20:02.079 --> 00:20:04.960 little guys out there doing stuff? Now do the overlap? 385 00:20:05.359 --> 00:20:06.200 Is it log. 386 00:20:06.000 --> 00:20:07.880 Files that gets spun out of these things? How do 387 00:20:07.920 --> 00:20:11.319 you actually know what they've done to where you can 388 00:20:11.519 --> 00:20:13.519 correct them or optimize what they're doing? 389 00:20:14.160 --> 00:20:16.759 Amazing, amazing question, unbelievable question. 390 00:20:17.319 --> 00:20:21.160 So what we are when we first of all, let 391 00:20:21.200 --> 00:20:24.279 me explain to you what is an agent architecture? 392 00:20:24.599 --> 00:20:28.279 Okay, yeah, complete, Right. 393 00:20:28.759 --> 00:20:31.559 So what we have done is in our we can 394 00:20:31.599 --> 00:20:35.480 build agents for anything under the sun. Right, but we 395 00:20:35.640 --> 00:20:37.839 decided as a company that we're going to focus on 396 00:20:37.960 --> 00:20:41.960 building this agentic platform around one value chain, which is 397 00:20:42.000 --> 00:20:44.960 software engineering, because we said we want to be super 398 00:20:45.000 --> 00:20:46.160 specialized in this area. 399 00:20:46.519 --> 00:20:46.880 Mm hm. 400 00:20:47.440 --> 00:20:50.000 So if you really look at an agent, an agent 401 00:20:50.079 --> 00:20:54.640