External research, not Kihon results or guarantees. These figures measure different outcomes and cannot be compared or added together.
What this looks like
Real work. A practical starting point.
01 / Build your team
From trying AI to working well with it.
ADX Asia
Four weeks of workshops, department clinics, and practice on employees’ own work.
The goal: a workflow each department can run, review, and share.
Week 01Use AI in everyday work+
Choose a useful task, give it context, check the answer, and improve the result.
Week 02Repeat what works+
Organise the references and instructions behind a recurring task.
Week 03Build a workflow with AI+
Test the steps, define human review, and handle missing or conflicting information.
Week 04Demo Day+
Each department showcases its workflow, demonstrates how it works, and shares what they learned.
02 / Build your workflows
A clearer path from budget to spending.
ADX Asia
An internal system connecting budgets, purchasing, and spending requests, with AI support along the way.
People retain approval and payment decisions.
The workflow being built
The AI layer is being designed to:
01Help prepare requests+
Read receipts and suggest details for people to confirm.
02Flag things to check+
Highlight mismatches, possible duplicates, and missing information.
03Answer business questions+
Ask about spending and budgets, with answers linked to records.
04Get answers where people work+
Ask questions and check status through the ADX internal chat app.
Results will be shared after completion and measurement.
Elbert Toha / Founder, Kihon
Who is behind the work
Business judgment. Hands-on engineering.
I’ve spent 16 years running my own businesses and 14 years building software. I understand the distance between a promising idea and something a team can use every day.
My experience spans McKinsey, business ownership, and software engineering. I hold a master’s degree in business from Nanyang Technological University (NTU), with a specialisation in technological innovation.
16 years
Running businesses
14 years
Software engineering
Let’s put AI to work
Start with a conversation.
Bring one task your team struggles with or one process slowing you down. Let’s find a useful first step.