Case Studies
Practical AI delivery stories, shown with clear evidence boundaries.
These case studies separate what is live, what is in progress, and what remains intentionally unclaimed.
Current case studies
Each card opens a detail page with the story, evidence trail, claim boundary, and current status.
No time to read? Turn documents or pasted Thai–English–Chinese text into an MP3 to listen to whenever it suits you.
Invitation code required. Speech generation is available while the host is online; pronunciation is still being tested.
Local validation report →Test results, review decisions and what remains before launch.Local composite implementation proven across 11 tables and 805 bounded rows; modeling and Azure Databricks implementation not started.
Local composite evidence only. No real Telco customer data, model, Azure Databricks, production latency, operator validity, or production-readiness claim.
A time-boxed learning and demonstration use case for a governed customer-churn retention decision-support workflow.
Sample data only. No production accuracy, causal impact, campaign uplift, or real-operator validity claim.
A developing case study on turning delivery evidence, assumptions, and recovery constraints into a safe public story before an entry is treated as complete.
Not yet a completed case study or validation claim.
A bounded evidence case study on how AI adoption work becomes measurable, traceable, and claim-safe across controlled execution rounds.
Not a production, ROI, full automation, or real-world business outcome claim.