@Nick Puruczky my email id is [email protected]

After 2 days of locking In and 3 days of explaining my model to more than 50 people in the Genesis Model Making Competition at St. Aloysius (Deemed to be University), Mangaluru, we won 1st place in the CS category and 2nd place in the overall 1st PU, competing with more than 50 Teams!! (Certificate coming on 22nd Aug… until then, we made our own🥀) Met some really awesome people. Improved my communication skills ALOT. Looked into some other Insane models. Explained our AI FrontDesk OS to a lot of students, parents and some other awesome guys who were into this field. Well, after this, I need to start working on another competition on the 22nd of this month called Campus Insights. It basically involves collecting data from students on a specific topic, analysing it, and presenting the findings and insights through charts on a chart paper. This was some really awesome event, BUT I think it's time I lock in on my real stuff, which is my Internship, Roamoir and my Agency. Here's the GitHub repo for my project: https://github.com/Arhan-SE/AI-FRONT-DESK-OS

Came across an interesting new AI project called Rakazo. Instead of having one AI chatbot, Rakazo lets you create multiple AI teammates. Each bot can have its own computer, memory, routines and conversation history. Bots can even create other bots or use sub-agents to complete work. What I find particularly interesting is that it is open source and self-hostable. You can choose your own AI model and sandbox rather than being locked into one provider. It currently supports options including OpenAI Codex, GitHub Copilot, SuperGrok and models through OpenRouter. Imagine building a small AI team: 🤖 Research Agent 🤖 Sales Agent 🤖 Inbox Manager 🤖 Operations Agent 🤖 Finance Assistant Each with its own role, memory and routines, but working together. It is still in beta, so I would treat it as something to experiment with rather than immediately depend on for critical business processes. But the direction is interesting. We may be moving from using one AI assistant to managing a team of AI workers. For anyone in AI Accelerator experimenting with agents, this one may be worth watching. https://rakazo.com/

As I mentioned in my introduction, I’m actively reaching out to founders to understand their business problems before building things that nobody actually needs. I’ve reached out to a few founders, introduced myself and what I do, and thankfully, some of them have responded and shared what they’re building and the problems they’re trying to solve. Now I’m at the next step 👇 One founder actually asked me to understand his business and tell him where AI could help him generate more revenue. This is where I need some guidance from people who already have experience getting clients. How should I approach these conversations? - What questions should I ask to properly understand their business? - How do I identify where AI can actually create value? - What kind of solution/proposal should I present? - And most importantly, what proof or portfolio should I have before trying to sell my offer? I don’t want to randomly pitch “AI automation” just because it’s trending. I want to understand the problem first, find where AI can genuinely help, and then build a solution around it. Would really appreciate advice from people who have already gone through this journey. What would you do if you were in my position? 🙌

Alexis Kingsbury created an experimental AI accountancy firm staffed by 11 AI colleagues. Each AI had its own role—from Managing Partner to Head of Compliance—and could do more than just answer questions. These agents could manage files, prepare SOPs, update systems and work on longer-term assignments with limited supervision. But the experiment also exposed the risks. At one point, the AI team was ready to take on paying clients without considering basic requirements such as AML registration and professional indemnity insurance. The outputs could look professional while still being completely wrong. To improve quality, AI agents were made to review each other’s work. But even then, human judgment remained the final control. For me, that’s the biggest takeaway. The future may not be an AI-only professional firm. A more practical model could be an AI-first firm—where AI agents do much of the research, documentation, routine compliance and administrative work, while qualified professionals handle judgment, review, client relationships and final responsibility. That could fundamentally change how accounting and professional firms operate. Would you trust an AI colleague to prepare 80–90% of the work if a qualified professional reviewed the final output? @Nick Puruczky What are you thoughts on this ? How practical or possible it is to achieve this?