AnsisMalins

It is not clear what exactly do you have in mind. If you're serious about it, you should write a design document, listing all the features and how they're supposed to work. Imagine yourself using it step by step. Draw fake screenshots of it, or comic sketches of specific use cases of the assistant. To illustrate what I mean, an example breakdown of one feature could be: The assistant shall detect the user's face using the webcam and greet them. If an unknown face is detected, the assistant shall lock the screen. There shall be a control panel for the assistant where, among other things, known faces can be specified. Adding a new known face will require a picture of the user be taken using the webcam, similar to how Skype lets you take a picture of yourself to use as an avatar. The assistant shall greet known faces at most once per day each and use a greeting appropriate for the time of the day. If any unknown and no known faces are seen in the webcam, the assistant shall lock the screen. Particularly, if at least one known face is seen, the screen shall not be locked - this is a situation when the owner is showing their screen to somebody else, so we don't want to lock them out! Risks associated with this feature: need to find a working face detection algorithm, it could be too inaccurate to be useful (e.g. people with glasses), it could require too much performance from the computer, there are privacy risks associated with filming the user, the assistant could be fooled with a photograph of a known face, and how many people have webcams anyway? And so on for every little thing you want the assistant to be able to do. Expect it to be a hundred plus page document. From it you will be able, with the help of a programmer, start to estimate what will it take to implement it, and which features you should focus on first, and what can be left for later or dropped entirely, and help you define the so called minimum viable product.