FEATURE: MOBILE-FIRST AI SOLUTION
WE HAD TO CREATE INTELLIGENT CHECKS THAT IMMEDIATELY FLAG IF THE IMAGE ISN’ T GOOD ENOUGH AND PROMPT THE USER TO RETAKE IT. THAT ALONE REDUCED ERRORS BY A HUGE MARGIN.
This case study examines the development of Smyl AI, its application in real-world patient care and why it may be a precursor to the future of frontline healthcare.
Building for the real world, not just the ideal user
For Anand Panchbhai, CTO of Logy. AI, the biggest challenge wasn’ t training the AI. It was designed to work in real-world conditions, where lighting is poor, camera angles are awkward and users often don’ t follow instructions precisely.
“ We weren’ t dealing with perfect dental X-rays. These are mobile phone photos – sometimes blurry, overexposed, or taken at strange angles,” said Panchbhai.“ So we had to create intelligent checks that immediately flag if the image isn’ t good enough and prompt the user to retake it. That alone reduced errors by a huge margin.”
36 INTELLIGENTCIO MIDDLE EAST www www.. intelligentcio. com. com