Kavin Bharti Mittal, founder of Hike and son of Bharti Enterprises Chairman Sunil Bharti Mittal, claimed that advances in agent-based artificial intelligence are compressing product development timelines, and said he has built several full-scale applications in just a week using OpenClaw and Anthropic’s Claude Code.
In a detailed post on X (formerly Twitter), Mittal said that in the past, that accomplishment would have required a large team, months of work, and a lot of money.
“I set up OpenClaw just 7 days ago. Here’s what I’ve built since then: In 2025, this would have taken a team of 10-20 people, 6-9 months, and over $1 million in funding,” he wrote. “7 days. Just me + OpenClaw + Claude Code… Total cost ~$600 (tokens, compute, etc.)”
Build multiple consumer apps in days
Mittal cited three products developed during this period, including a cross-platform note-taking tool and an AI assistant.
The first is called Simple Notes and is described as “a full-fledged Apple Notes replacement with full syncing across all your iOS, Mac, Web, and Android devices.”
Its second product, Lumenote, aims to rival Obsidian, a knowledge management platform with built-in AI capabilities. He described it as “a full-fledged Obsidian replacement…with better markdown handling and fully integrated AI to interact with my knowledge base.”
The third tool, CleoAI, is a conversational assistant offered through Telegram rather than a standalone app. “You just talk to it, you don’t need an app,” he said.
Mittal added that OpenClaw itself is still too technical for mainstream users and a simplified interface needs to be developed. “OC is too technical for the average person, so here’s a version that my girlfriend and brother can use.”
He said the app is awaiting approval from Apple developers for distribution and is built using cloud services such as Vercel, Railway and Supabase.
Autonomous agents as “digital staff”
Beyond apps, Mittal highlighted a network of AI agents that perform tasks typically reserved for employees or service providers.
His primary agent, “Sam”, serves as Chief of Staff with broad operational authority.
“Sam is my new lead… optimizing email, calendaring, booking, shopping, code and building… personal CRM… infrastructure,” Mittal wrote, adding that the agent maintains memory across sessions and can autonomously deploy subagents.
He said the system was able to process over 1,000 emails, manage reservations, book restaurants through browser automation, and even participate in group chats. “I don’t think they know he’s an agent (!),” he said.
Sam also has his own communication channels, including “his email and his phone number,” as well as voice capabilities powered by ElevenLabs.
Autonomous cryptocurrency trading system
Another agent called “Midas” manages cryptocurrency portfolios using predefined strategies and risk management.
Mittal said the system will execute trades directly on the exchange, perform yield farming, and generate daily performance reports.
“Midas has full exchange API access…but you will never be able to withdraw your funds,” he said, adding that it operates under a locked strategic framework and continuously monitors the market.
AI for scientific research
The third agent, “Ritam”, focuses on theoretical physics and the study of gravity. Mittal explained that it is about integrating insights across modern science and ancient texts, scanning academic papers and patents.
“Every theoretical insight is subjected to a pressure test: ‘What can we build then?'” he writes.
Signs of agent AI shift
Mittal’s post highlights the rapid rise of agent-based AI systems that can plan, execute, and coordinate complex tasks with minimal human intervention, a trend that is gaining momentum across the technology industry.
He said additional specialist agents, including marketing and distribution teams, are already in development.
“At least a dozen more agents are in the works,” Mittal wrote. “We just started a team of agents for marketing and distribution, and we have no idea what’s going to happen.”
