Muhammad Umar built CodeNinja's own internal platform called Hyper in three to four weeks, keeping full source-code ownership and integrating Microsoft Copilot for AI features. He argues that vendor-locked SaaS and CRM tools charge too much per seat, but AI now makes it cheap enough for any company to build and maintain their own software instead.
Chapters
There's a resource center within codeninja if you share the story of Resource center as well. So codeninja is an AI services and platform business. But a significant portion of what we do is that we integrate these services into customers workflows, which is generally called professional services. There's a way to manage professional services operations in organizations like ours. And we spoke to one of the platform company who have this incredible platform that we want to roll your, roll out your platform into our business. They don't work with any business that's less than 1,000 people. We were at the time 200 people, I think 200 people, 200 engineers at the time at coding. And then we said okay, I mean what to do now? And then I spoke to again, Ibrah that I mean we're building this platform Hyper. I mean these are the very early days. Even the name wasn't there at the time. Abra said, I can, I mean build this probably in weeks for you and literally you using Hyper. We bootstrapped this entire resource center in probably three to four weeks. And then it had all the functionality that was required. And then when we handed it over to professional services team, they were all engineers. Our professorships is led by Sultan and because we had all the source code, we had all the platform with us. They used Ms. Copilot using MCP and built this incredible agentic experience on top of it. I can guarantee you that this resource center, this professional services platform that we're using, Driver would work I think a lot better than that. I won't name it that. Big business, right? I can tell you because it's built for our needs, number one. And second, we have an absolute control how to deploy AI on top of it. We control the meta layer on top of which like, I mean a couple of engineers pulled off these incredible agentic experiences on top of mscopilot using this. So this is an incredible case study. Other than that, I mean now we're trying to replace every other software that we use and we're paying, I mean we're using one of the CRMs that's costing us millions a year. I mean there's no way, I mean, I don't, because I lead sales too. Out of 200 screens, they offer probably five, use three. Right? And every month they charge us like these $50 and $70 a seat. For what? I don't use those. So this SaaS model, this has to be, I mean I think we had a great run for that, but I think this model is over because building platform building software is now easy even for businesses who have deployed some of the most expensive let's say CRMs or ERPs whatever right. There's no point of paying these expensive monthly and yearly fees and why the transition is hard I mean it's not like that we have realized this everyone has realized it the transition is hard because of data because of the existing deep integration of these businesses into these CRM businesses or ERP businesses into these organizations and there's no for them for businesses who have deployed these large systems these ERP systems are it's very hard for them to get out of these traps I would call it now traps right But I think with AI I think that shift is now given you can't I think we had a great run for these enterprise applications but I think now I don't see why a business in 25, 26, 27 would deploy any SaaS software into the organization won't build its own I still have to struggle for it probably they don't know much about it but I don't know Yeah.
