Pretend Product: Socratic Siri
Pretend Product is my musing on a product that I think should exist in the world or improvements I’d offer to an already existing product
For this Pretend Product, we’re creating something new: Siri for language learners.
Apple wants to make money and incentive more people to switch to iOS.
Gen Z and Millenials who are no longer in school taking foreign language courses but still want to maintain or develop fluency in a second or third language.
As a user, I want to develop fluency in another language in a way that is natural and enjoyable.
Note: adults can learn languages well. Do not be deceived.
Duolingo and language learning gamification apps, language practicing clubs/meetups, listening to movies and media in a different language
What they miss:
Apps are often too simple/slow in building you up to fluency and may provide nonsensical phrases that you won’t practically use in day to day conversation.
Language practicing clubs/meetups aren’t happening in person (as with a lot of things) due to COVID and speaking with others can be intimidating or awkward.
Listening to movies and media in a different language may have speakers who are going too fast, discussing a niche topic, or speaking at a comprehension level that is past my current level.
The specific design that I have in mind follows users' current schema of activating Siri. Socratic Siri is activated by saying “Hola Siri” (the hello would be in the target language as more language packs are added). The conversation would flow as a conversation does.
How are you?
What are you up to now?
How was your day?
What did you do today?
What do you wish would’ve gone differently today?
What do you want to do tomorrow?
This diary format allows the user to practice in different conversational tenses and would incorporate language that is actually relevant to a user’s life by having them talk about, well, their life.
The general structure is that Siri would ask you a question in Spanish, give you an example sentence structure, repeat the question then wait for your response. In your response, Siri would listen to distinguish the English phrases from the Spanish ones to see what words you ‘missed.’ Siri would then restate what you just said in the grammatically correct structure and with the English words switched out for the Spanish word that you were looking for.
Me: “Hola Siri!”
Siri: “Hola Phyllis! Coma estas? If you want to say ‘I am good, thanks!’ you can say ‘estoy bien.’ You can also say ‘Estoy mas o menos’ if you’re just okay, or ‘Estoy mal’ if you’re not good. So, ‘como estas Phyllis?’
Me: “Estoy tired.”
Siri: “Estas cansada. Tired in Spanish is cansada. Por que estas cansada?”
And so on, adjusting over time as you naturally incorporate more and more words that you’re learning into your conversations with Siri.
This is not too far from the English experience where Siri teaches you it’s capabilities as a virtual assistant by prompting you for some possible next requests.
The recordings are not stored in your account for you to reference but the data contribute in aggregate to Socratic Siri’s improvement.
If you’re like me, you can write an essay in 3 languages but converse in only one. Speaking is the best way to learn a language and I believe speaking to Siri is a good way forward to do that. This experience ties back to the user goal by allowing you to have a conversational partner that doesn’t flake, doesn’t laugh at your mistakes, and utilizes NLP and ML (natural language processing and machine learning) over time to adapt its comprehension level as you build your fluency. It’s incorporated into a tool that you already use and the specified rollout plan actually tests on users who have demonstrated activity with using Siri in their lives already so the only excuse is your own excuses or issues with the specific design solution.
Ties back to the business goal by having tiered features. Users pay more to have more than one language in Socratic Siri. If you’re learning both Spanish and French, Socratic Siri can only partner with you in one at the lowest tier. A mid-level tier offers 2 languages. Premium tier offers unlimited. The free offering is a daily 2-minute sample conversation.
Start with Spanish. Release to 50% (A/B test) of users who have used Siri at least 3 times per week for the past 6 months in order to filter for users who do not utilize Siri in the first place.
A use of Siri is measured as an activation by ‘Hey Siri’ where a task, such as setting an alarm, sending a text, or playing a song, was completed successfully.
Successfully means that Siri did not respond with ‘I’m sorry. I didn’t quite get that.’ or with ‘Here are some results on the web for…’
10% lift in activation of Siri in the variant.
Increased revenue for Apple.
60% of users who activate Socratic Siri at least 3 times per week for the first 14 days of the beta launch continue to activate Socratic Siri at least 3 times per week at 60 days.
Rollout if test shows success:
Phase 2: rollout to 100% of users who used Siri at least 3 times per week for the past 6 months before the phase 1 launch.
Phase 3: rollout to all iOS users.
Offer Socratic Siri as part of a discounted bundle with other Apple products like Apple Music in order to upsell and increase revenue.
Add more languages over time. Add English for non-natives.
What do you think? What excites you about this? Why wouldn’t this work? There are plenty of reasons to both sides. I’d love to hear your thoughts! Feel free to message me on my LinkedIn or write your own Medium post with a better solution to the user problem!