| Some unorganized thoughts: When I look at it, it sounds like it is "text based siri with a memory." When siri has memory, siri must do it through a specific app interface that is often not always directly clear how to query. Siri does not have memory, the reminders app has memory. I'm not sure you have any secret sauce that apple can't copy in a couple weeks. I am not clear what value it has over siri to make it worth $100 a year. I don't think any of the voice assistants have really caught on as a generalized assistant, but I could be wrong. I Suppose your 5 second intro is: AI Personal assistant with a memory. I suspect you will find one use case that people really want, and then you will pivot for that niche. The way my mom uses siri is for very specific use cases after I have told her that siri can do it. So she uses it for timers, messaging when her hands are busy/dirty, initiating calls, and google searching. She wouldn't think to do anything else with it. She never would have used it if I didn't give her a personal lesson. There is probably a real onboarding/education problem. Learning about a product is a pretty high upfront cost to using it. Hyper optimzing for 1 case this does really well and slowly onboarding someone to the more powerful features might be a way to get people into it. What problem does the app solve immediately? How does it make my life better immediately? How does it do those things immediately and better than competitive products. With the notes app, I know exactly where everything is and can organize things the way that makes sense to me. Having a nebulous cloud of stuff with a nebulous interface makes it hard for me to imagine using it. A lot of the stuff shown is actually structured data that I would like to see a real representation of. I would want to be able to see a calendar of birthdays, I would want to see a bio sheet on the people the app knows about. I would want to see a list of pending appointment and todos, etc. Importantly, not in a text based format. Strides is an app that I have used before for keeping track of habits I want to form. It is the interface that provides value to me. It's like a dashboard for my habits. So what are the features that Things, Strides, Siri, Google Calendar, Notes, etc. have that would make them the chosen apps over this one? It is the presentation of the data involved and ease of managing it and viewing it that gets me to use them. I don't think I've seen an FRM (like a CRM for social relationships), but it looks like in the examples, that is kind of the object of the memobot queries. FWIW, seeing an index of explicitly supported and thought out use cases in a documented format would be helpful. The maximum that it might be able to do and creative use cases it might handle is much less interesting than the minimum it will definitely do. One major significant flaw of non discrete interfaces is that any time the intention to do something does not result in that thing happening, it makes someone not want to go to the app for help. I don't know how you are measuring failure to perform task based on misunderstanding, maybe repeated queries for similar things or requests that result in low confidence behavior? That seems like a very important metric. Maybe an option for more technical than conversational responses would be nice. If I had a list of career goals, I would want bullet points rather than a comma separated list, for example. Also it might be useful to bold or italicize proper nouns in your system that can be referred to. for example: MOM, LONG_TERM_CAREER_GOALS, REMINDER. Emphasizing the things your app can do or manipulate in the responses might make it more flowy to use. |
Thank you hayst4ck, this is just amazing!
Definitely a lot to think about.