Skip to main content

11A – Idea Napkin No. 1


1) Who I am: I'm a nerd who loves 3d printing and robots and AI and all things cutting edge tech, yet my job right now is typing prescriptions all day while I daydream about robots and machine learning. I feel like a robot should take my job and I feel like it can do so feasibly and open up gateways to other options such as handwriting detection in general. I'd more be interested in the AI than entrepreneurship though. 2) What does it do: It transcribes prescriptions automatically and a company using it only needs to use human error checking rather than an entire central utility technician fleet. 3) The consumers: at first just the pharmacists and the pharmacy industry, but once the algorithm gets good at reading doctor chickenscratch, it can be used to transcribe generic handwritten notes like for students. 4) Why they should care: don't have to pay for the central utility humans' wages, and also the notetakers would save time in transcription. 5) What's different: Technically I don't have anything that anyone else doesn't have access to or could do, but nobody has attempted to implement AI handwriting recognition technology into central utility pharmacy stuff, and with millions of correct training data for the algorithm, it can be trained easier than one would expect. I believe my elements all fit together because it's very specific to one industry with some potential in other areas, but probably the weakest point would be the actual difficulty in the initial accuracy of the training algorithm and the time it takes for it to train it and error correct. One would need a research team before a entrepreneurship team.

Comments

Post a Comment

Popular posts from this blog

10B- Elevator Pitch No. 1

Haha look at me I'm pretending I'm a CapITaLISt HO HO LOL GIvE mE MonEy I hATe EmPLoYeEs Words? So there I am at work, tap tap tap, typing away at a computer. Sounds like a normal desk job right? It's a bit too normal. All I do is type up prescriptions digitally. Ten hours a day, four days a week. It's bad enough for me with ADHD, but with 500 employees on the floor doing this day in and day out, it can be pretty mind numbing for anyone. And I'm thinking to myself as I am working, that I felt like a robot, and not in a good way. Then it hit me. There's literally millions upon millions of these prescriptions entered and stored. And we have machine learning algorithms to take thousands of images and learn what's in them and export them to text or even a new image. What I'm saying is why don't we literally let a robot take my job? Walgreens would be saving by not having to employ 500 monkeys at typewriters. What could go wrong with the employees an...
1 comment
Read more

8A - Solving the Problem

My job is weird. I don't mean this negatively. I mean this in the sense why are humans still doing this? Data entry of prescriptions? It seems like this is something a robot should be doing. And I don't mean this in a "robots should take our jobs" sense (even though they should). I mean this in the sense of why are we relying on humans to mash out what's written on prescriptions when the technology exists to better extract data from handwritten text and even better so from printed text. My point is we already have millions of training examples correctly documented. If we used these as training data for a machine learning algorithm, it would be able to create immediate more accurate prescriptions for pharmacists. Not only that, but imagine how good AI could become at reading handwriting if it can decipher doctor cursive! This potentially has many uses outside of the pharmacy just in terms of technological advancement alone. Now this is just one possible solution...
2 comments
Read more

12A – Figuring Out Buyer Behavior No. 1

Since it is easier to seek out people who need the service as a notetaking transcription rather than actually interviewing supervisers at work, I used this segment for my interviews. I find that students desperately seem to want a service to be able to take their notes and digitize them, so that their scrawl is easily readable later on and so that they can share notes with friends. They seem to look for services that'll either detect their handwriting (which are rudimentary at best and barely able to read print) or for services for other students or such to type up their notes for them so that they save time. This becomes especially apparent around exam time when students are desperately trying to organize their notes but simply don't have the time. An image detection tech trained on the universally most messiest handwritings would be able to detect and transcribe student notes with just a push of a button. It could be put on phones or on library computers with scanners.
Post a Comment
Read more