I build practical AI ventures around real workflow gaps.
I am Os Ishmael, founder of 2nd Bell and builder of a small portfolio of AI-enabled ventures and services across revenue workflows, learning, focus, websites, and emerging data infrastructure.
The surface areas are different. The pattern is consistent: find a real behaviour gap, design the workflow around it, and use AI only where it makes the system clearer.
The common thread
Different projects, same operating pattern.
I tend to build where a useful thing is being lost in the handoff between people, tools, and decisions.
Service businesses lose opportunities after first contact.
Students use AI before they know how to test, explain, or improve the work.
Workers lose rhythm after breaks, context switching, or unfocused sessions.
Hong Kong SMEs need websites that create clarity and credibility faster.
Robotics teams may need structured, consented real-world task data before models can be evaluated well.
The work starts with the constraint, not the tool.
Current work
The portfolio I am building.
2nd Bell is the flagship. The rest of the portfolio explores the same practical system-building instinct in adjacent places.
2nd Bell
Revenue workflow visibility for service businesses that need to see where enquiry, follow-up, booking, sale, payment, or repeat business gets stuck.
Visit 2nd Bell →AI Skills Tutor HK
Practical AI habits for Hong Kong students, families, teachers, and schools that want confidence without hype.
Explore AI learning →Qithym
A focus and return-to-work tool for people who lose momentum after breaks, context switching, or long work sessions.
Follow the build →Website HK
A website and digital presence service for Hong Kong SMEs that need clarity, credibility, and stronger enquiry paths.
Explore Website HK →Venture in development
Robotics Data Venture
Exploring structured, consented real-world task data for robotics and embodied-AI teams working across Asia.
Background
What shaped how I build.
The work is portfolio-shaped now, but the foundation comes from years spent around the systems that decide whether opportunities, habits, and operational momentum actually move forward.
Approach
How I decide where AI belongs.
AI is useful when it helps a real system become clearer, faster, or more accountable. It is not the starting point.
Start with the behaviour
What are people actually trying to do, where does momentum break, and what would make the next action easier?
Make the system visible
A good workflow should clarify ownership, status, next steps, and the point where the work is getting stuck.
Use AI where it earns its place
AI should reduce friction, improve judgement, or support follow-through. If it does not, the simpler system is usually better.
Useful context
A few things people usually want to place.
What is the main focus?
Building practical AI-enabled ventures and services around real workflow gaps.
What is the flagship venture?
2nd Bell. It is focused on helping service businesses find where revenue gets stuck and what to fix first.
What kind of conversation fits?
Founders, operators, schools, SMEs, investors, and partners with a concrete workflow, learning need, commercial system, or venture thread to explore.
Start a conversation
If one of these threads is relevant, start there.
I am most useful when there is a real workflow, learning need, commercial system, or venture idea that needs clearer structure.