We Just Launched Kaman. Here's What Happened.
Two years ago, I started building something I couldn't fully explain to most people around me. Not because it was secretive — but because it was still becoming. A platform where any business owner, any entrepreneur, any person with an idea could build AI agents that actually work for their business. No code. No cloud infrastructure expertise. No monthly invoices to five different AI tools that still don't talk to each other.
On February 23rd, we showed it to the world for the first time.
This is the story of that day.
The Format: No Slides, No Pitch Deck
We made a deliberate choice going into the launch event: no polished marketing deck, no scripted keynote, no "the future of AI is here" theatrics. Just Rajeeb Mohapatra and me in conversation, with our team members joining us live to build and demo real things, in real time.
If something broke, it would break on stream. That was the deal.
Rajeeb — entrepreneur, investor, someone who understands the ground reality of Indian SMEs — pushed back on every claim I made. Not combatively, but honestly. He asked the questions a business owner would actually ask: Does this work for a garment dealer in Surat? What happens when a customer sends a voice note? What if I don't have a tech team?
Those questions made the demos sharper. They made the event real.
What We Built, Live
We ran through seven demonstrations. Each one was built or operated live on stream by a member of the Yoctotta team. Here's what we showed:
1. WhatsApp Sales Agent + KDL as CRM
Every salesperson in India runs their business on WhatsApp. The problem is there's no system behind it — leads get lost, follow-ups don't happen, and the owner's personal WhatsApp becomes a 500-chat graveyard of potential revenue.
We built an agent live that qualifies inbound leads on WhatsApp, extracts structured data — product interest, budget, timeline — and stores it in KDL, our built-in data lake. The result: a fully functional CRM running directly inside WhatsApp, with no Salesforce subscription required.
Rajeeb's reaction: "Every garment shop, electronics dealer, and real estate broker in India runs sales exactly like this."
2. WhatsApp Shop/Menu Assistant
The flip side of sales — a customer-facing agent that handles inbound queries for a business. We showed a restaurant menu assistant that could answer questions, take orders, handle customizations, and capture customer preferences — all over WhatsApp, automatically.
3. Appointment Booking Agent
Salons, clinics, tutors, dentists — anyone whose business runs on scheduled time — needs a booking system that doesn't require a receptionist on call 24/7. The agent handles WhatsApp booking, sends reminders, manages rescheduling, and follows up on no-shows.
4. Invoice & Payment Follow-up Agent
The most awkward part of running a small business: chasing payments. We showed an agent that tracks pending invoices in KDL, sends progressively firmer WhatsApp reminders, and can trigger Razorpay payment links automatically. Uncomfortable conversations, automated.
5. Hiring & Recruitment Screener
Candidates apply via WhatsApp. The agent screens them with structured questions, scores and ranks responses in KDL, and delivers a shortlist. No HR software subscription. No ATS. Just Kaman and a WhatsApp number.
6. Customer Support + FAQ Agent
Feed your product documentation, policies, or service guides into KDL — and the agent handles 80% of inbound support queries instantly. The remaining 20% get escalated to a human with full context. The ratio flips over time as the knowledge base grows.
7. Vendor & Procurement Coordinator
The final demo showed a more complex workflow: inventory counts coming in via WhatsApp from floor staff, the agent checking KDL for approved vendors and pricing, auto-generating purchase orders, and following up on delivery timelines. Multi-step, multi-agent orchestration — built in minutes on the platform.
The Moment That Mattered
Somewhere around Demo 4, a viewer asked in the chat: "Is this a ChatGPT wrapper?"
I stopped the demo to answer it directly.
Kaman is not a ChatGPT wrapper. It's a platform — multi-agent orchestration, a full knowledge management system, a built-in data lake (KDL), a workflow engine, multi-channel deployment across WhatsApp, Telegram, web, voice, and email, and RBAC-based enterprise controls. We've been building the architecture for two years. The LLM is one component inside a much larger system.
That distinction matters because wrappers break when the underlying model changes. Platforms don't.
What's Next
The launch was the starting line, not the finish line.
Kaman is now open for early access at kaman.ai. We're onboarding SMEs, enthusiasts, and builders who want to create agents that actually do something — not just chat.
We're also preparing enterprise demos: an ESG reporting platform built on seven coordinated Kaman agents, and an AI-native SDLC management platform that shows what software delivery looks like when every stage has intelligence behind it.
If you want AI that works for your business — not AI that impresses you in a demo and confuses you in production — that's what we're building.
<iframe width="1217" height="685" src="https://www.youtube.com/embed/C1Wc0qkLsak" title="Kaman.ai Launch: 7 Live AI Demos to Transform Your Business" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
A Personal Note
This launch took two years of building, funded entirely by Yoctotta's consulting business. Most of the architecture was built solo, late nights, weekends, and flights between client engagements. The team did the client work that paid the bills; I built Kaman on the margins of everything else.
I don't say this for sympathy. I say it because I know the cost of the thing, and I know its value.
On February 23rd, the world got to see it for the first time.
It felt exactly like it was supposed to.
Surajbhan Satpathy is the founder and CEO of Yoctotta Technologies and the creator of Kaman. Follow the journey at kaman.ai.
Enjoyed this article? Share it with others!
