Product Manager Interview: Design an ATM for the Blind
Vložit
- čas přidán 21. 07. 2024
- ⭐️ FREE resources for you! ⭐️
✅ Full step-by-step guide to break into Product Management: pmdiego.substack.com/p/if-you...
✅ Survival guide for new Product Managers to be an amazing PM in no time:
pmdiego.substack.com/p/-survi...
----------
#ProductManagement #Interview: Design an ATM for the Blind
💻 Visit my Website for all the information you need about Product Management!
www.pmdiego.com/
💎 Did you know I post content every week on LinkedIn about Product Management? Let's connect!
/ diegogranadosh
📝 Check out my Newsletter on Substack!
Learn how to find Hiring Managers on LinkedIN and how to build side projects!
pmdiego.substack.com/
📚 Did you know I'm writing a book? Check it out here!
www.pmdiego.com/my-book
-----------
00:00 Introduction
00:22 How would you design an ATM for the Blind?
01:30 Clarifying Questions
02:51 Goal
05:35 Customer segmentation
06:58 Use Cases / Pain Points
09:13 Solutions to the Pain Points & Prioritization
16:34 Risks for one of the Features
17:52 Metrics for Success
19:31 Summarizing the problem
20:48 Interview Feedback
🤔 Why is the face blurred?
Participants in Mock Interviews are either current Product Managers or aspiring Product Managers. Each interviewee decides whether they want their face blurred for the video - I believe that this also allows us to have a more candid interaction and feedback session without the pressure of having their face in a CZcams video.
✅ Would you like to participate in a Mock Interview with me?
✅ Sign up here: forms.gle/H6G13kAjvtBS8NA48
📝 Want to download the framework? Here's the PDF Format: drive.google.com/file/d/1r9bE...
💻 Want to see an explanation about the Framework? • How to answer Product ...
----------------
🤔 Have questions about Product Management? Let's connect on LinkedIn or ask me on TikTok & Instagram!
👉 linktr.ee/pmdiegogranados
Check out the different playlists I created to help you get into Product Management:
📹 Check out other Mock Interviews: • PM Mock Interviews
📹 Awesome guests from Amazing Companies: • Awesome guests from Am...
📹 Ace your PM Interview?: • Ace your PM Interview!
📹 Product Management 101: • Product Management 101
📹 Resume Advice: • Resume and LinkedIn Ad...
----------------
#pminterview #tech #googlepm #facebookpm #applepm #microsoftpm #productmanager #product #prodmgmt #productdesign
Share your moonshot idea below in the comments!
What do you think of this Mock Interview?
Want to participate in one? Signup here! forms.gle/ZSTNhZ3bYABrTtPbA
@@NexusAIInnovations hi! For sure :)
Here’s a video on how to answer case questions: How to answer Product Case Questions (Part 1) - A framework to ace your interviews!
Best of luck with the interview and let me know how it goes!
@@NexusAIInnovations happy to help with that!
To learn more about what is Product Management you can view this video I did: What is Product Management?
czcams.com/video/dJmg4ODH0YY/video.html
And to know the interview for PMs will be at most companies, I have this playlist: Ace your PM Interview!
czcams.com/play/PLcMv0LrPMqFUjPzGYFyquwxbkMyx81GGD.html
In general, PM interviews will be more centered around business, strategy, product improvement questions from a holistic view (understanding goals, customers, pain points, trade offs, etc.)
Engineering interviews are more focused on the technical skills and experience. I’m no expert in engineering interviews (since I’m a PM), but interview questions will be more focused on technical cases, algorithms, etc. not so much on the business, strategy, and other aspects that will be part of the PM interview.
Let me know if this helps and if you have any other questions
Really loved this. Thank you. Apurv
Great job.
Interviewer as well as the interviewee
My moonshot idea - Assuming these blind folks already have a bank a/c, the bank can fetch data like facial scan, voice pitch and finger prints ( all fingers) and leverage it for future use - transaction here. Now incase the user's are using a phone, whenever the scan is done they will receive a call and they can direct the algo to work on voice commands, like an virtual assistant, and unless the transaction is complete the phone won't be hanged, only when confirmed then the phone call will disconnect, here we are saving headfone cost and leveraging existing resources.
Let me know your thoughts.
That's a great suggestion Abhijeet! this solution can work for anyone! You can also pre-approve everything and as you are in front of the ATM, with Facial recognition or fingerprint you can simply retrieve the money - since everything was already pre-approved :)
Thanks for the suggestion!
I like this idea and was thinking the same. To use the smartphone for audio communication vs speaker of an ATM.
In addition I think an important consideration is regulatory and compliance concerns in fintech. This could have been mentioned as a risk consideration
Thanks for the mock interview!
Some ideas from me.
1) sound proof cubicle
2) fingerprint/facial scan for identification
3) Voice command (no tactile/braille buttons needed) - should be relatively straightforward if the use case is just withdrawal.
I really appreciate these mock interviews. It helps a lot - I went through this with Diego and the interviewee, pausing to come up with my own user segments, pain points, solutions, etc., and found it really useful.
This is refreshing! He did a good job and Diego is such a great interviewer. The anonymity throws me off a bit, but I understand. I love how you have shown key pieces to keep in mind during an interview. Thank you for the video, Diego!
Thank you, Abel! The anonymity is optional - the next mock interview won’t be! But it’s up to the participant :)
@Diego Granados, on the point of mobile app just noting that it would need to be ADA compliant. Thank you for the time & effort you put into making these videos . Great learning experience.
Indeed, Sri! There are so many nuances and details to keep in mind. Thanks for sharing this reminder!
@PM Diego Granados Can we add a functionality wherein the user uses a headphone to answer the questions of the ATM and at the end of the transaction, ATM asks the visually impaired individual to rate the experience on a scale of 5 which the user can speak out, and at the end of the month, we can take an average of these ratings to get an understanding of how satisfied the user is with our product which would essentially provide us with a KPI to measure user satisfaction?
Also, I think, to avoid privacy concerns associated with AI in general and face unlock in particular and to also avoid issues wherein face detection sometimes fails (consider the case wherein the blind person has now grown a lot of beard), we can ask the user to scan his 2 fingers. For example, when he plugs his headphone, he would be asked to scan the index finger of the left hand and then the thumb of the right hand. This would add a lot more security as we are randomly scanning any 2 fingers and also does away with the privacy concerns. Please correct me if I am wrong.
Love this segment! I am a PM myself and it still takes a lot of courage to do a public interview like this and not feeling like drowning in a million decision points.
It seems to me that if the candidate had spent time going through the entire process/screens of withdrawing money, the customers pain points might have become more apparent. To me, there are two main pain points:
- consuming information (which aligns with your question about audio)
- inputting information (which was what he initially focused on)
If these two problems can be solved, we should have a pretty decent MVP. Such a classic PM question! Good job.
Thank you! I'm glad that you liked the video !! 😁
Thank you for the video! Two ways:
1 - Perhaps, we can have headphones with a microphone, to guide them step by step. They can input their amount on the keypad of the ATM which has braille. You don't have to say it aloud, and it doesn't display anywhere. Only beeps and a feedback beep.
2 - We can make use of an IVR call to their mobile phone, once their fingerprint/facial scan is done. They can input the amount to withdraw on their mobile and also continue to complete the transaction.
Such a great interview, I got 70% of the things correct, still I have so much to improve. But learnt a lot from this exercise. Thanks and keep these coming!
Thank you, Rajath! Have you checked out the other interviews? They are also great!
One of the key areas especially because this is something to do with the money of a person who cannot see, first onus is on providing an ATM card with an indication to the ATM machine (Ai) that here is a person who is blind..he swipes it to start with before entering.
- the machine immediately guides him on if he is all alone inside the cabin, which is the key , security.
- after this it's all AI, guides on what the machine looks like, a contact less swipe, would first suggest him his balance details etc and we can use headphones here to keep information that can be directly heard.
For me, security is the key, the card communicating to the machine that the person is blind, ofcourse it's camera and AI
Thank you for sharing these ideas @Ranjith! Question for you - I believe these solutions are great, but we tend to focus a lot on the technology part of these case questions. Do you think people would be ok having a card that identifies them as having a visual impairment? What if someone doesn't want the card, how would the ATM identify them? 🤔 Can you think of another way to identify this segment of users without needing a specific card?
@@PMDiegoGranados
First: the blind person stands in front of the atm, that is built with an intelligent AI based camera right at the top,that reads the person's movements
Second: the machine will always start with a " enter the code on the screen to continue", this is for any user.
Third: AI camera reads the person's movements and Voice assistant starts the timer for say 5 seconds to receive the entry, failing which the AI enabled camera sends a message to the main control that the person is visually impaired.
Fourth: A specially designed Case comes out , one with a small ear piece, Hard number pad designed for blind, A opening from which the cash can come out.
fifth: Assistance is all via the ear piece specially designed for blind, that takes through the entire flow from Asking for pin, other details, to balance to the amount and dispatches the amount and once the AI sensor camera confirms to the control the txn is done, the specially designed Case goes back to its place.
Sixth: In events of security breach, the case automatically closes upon the signals from the AI sensor camera , this is to ensure the blind person is not cheated of theft.
What if we setup ATM vestibules which can address the security concerns of hearing someone(we can use headphones as well to address this) and my concern would be someone taking the blind person's money. If the person is in enclosed glass room then we can make sure he gets the money safely. cons of setting up ATM vestibules is it would be very expensive.
We can add few more use cases as 1. Changing ATM pin 2. Checking last 5 transactions 3. Utilities bill payment.
Can have a Bluetooth which can be automatically connected to the phone and the features of phone can be utilised.
Thoughts?
I think there can be an IVR type of a mechanism through which the machine communicates with the user. The user understands all the commands that the machine can understand. Every command can be mapped to a button on the screen.
Also, instead of communicating the balance directly to the user, the user can get some form of a voice mail on his mobile phone or a call which communicates the same to him.
In the above interview, there was no mention of card. As card is always needed to take money out from ATM. Facial recognition can be used for authorization, but card will be needed for authentication.
This is great, thanks for the idea Animesh!
I like the idea of the phone call to make it more private.
If there is facial recognition and the ATM understands who is in front... would you still require the card? Or would that be enough to let the user get access to their money?
@@PMDiegoGranados A person may have bank accounts with different banks. If the assumption is that the ATM is for a single bank, then the facial recognition should be fine. Else the person will have to present his bank's card at the ATM. The later gives more flexibility to the users but adds a layer of complexity.
This is great, thanks for the ideas Animesh!
@PM Diego Granados Can we add a functionality wherein the user uses a headphone to answer the questions of the ATM and at the end of the transaction, ATM asks the visually impaired individual to rate the experience on a scale of 5 which the user can speak out, and at the end of the month, we can take an average of these ratings to get an understanding of how satisfied the user is with our product which would essentially provide us with a KPI to measure user satisfaction?
Also, I think, to avoid privacy concerns associated with AI in general and face unlock in particular and to also avoid issues wherein face detection sometimes fails (consider the case wherein the blind person has now grown a lot of beard), we can ask the user to scan his 2 fingers. So, when he plugs his headphone, he would be asked to scan the index finger of the left hand and then the thumb of the right hand. This would add a lot more security as we are randomly scanning any 2 fingers and also does away with the privacy concerns. Please correct me if I am wrong.
I feel that facial scans are a good idea but again inhibit security because people get misused very easily in countries like the one I come from, even to the point that someone brainwashed them to withdraw their money and then steal it and it is just easier with fingerprints and facescans rather than a memory based code. I would recommend having more privacy over the sound and use that as the main solution. If we have an common headphones connected to the ATM which wouldn't cost a lot more so that while withdrawing the money the person can wear it and we place it at a maximum acessible location to use it, so it is easy to find. Having a headphone provides clearer sound/ more privacy/ and also if there is braille then I don't think it shall be hard to enter a 4 digit pin only known by a user. With obviously the headphone conveying to them the steps they have used. I am just a new beginner looking to express my ideas.
Also I loved this interview the most, the interviewee figured out a lot more and better segmentation of the users than I could and also the use cases, but I couldn't relate to his soln maybe cuz of the place I come from or the issues in tier 2 countries we need to take care of while designing a product
I think a good combination of moonshot idea and voice command solution should be a 3D cam that scans the perimeter and alert the user is there's someone close to them or Behind them trying to eavesdrop
Hi Diego, thanks for the video. Great job on your channel!
For solutions, I have an idea of an ATM that "comes towards you". It might work as an ATM on demand. Say the user taps on the phone, the someone will carry an ATM to your house (or a place that is easier for the blind user to access).
My question: is this a good "moonshot" idea?
In general, what's your recommendation on coming up with "good" moonshot idea? Are we looking for "creative" solution that will be de-prioritized anyway, or genuine solution that might work given the right technology?
Thank you so much!
Great question Ming! Always go with very creative solutions that MIGHT be prioritized last but that are genuine solutions.
Bringing the atm to your house might be too cumbersome to execute, but what about sending to your home a package that can only be opened with your phone and in no other way?
That being said… remember that the prompt here is “how would you design an atm…”, and not: how would you make it easier to retrieve money? 😉 your solution solves a different problem
Hi, thank you for this video! I had a doubt. Is it a good thing to talk about the practicality of our ideas during the interview? For example, in case of identification methods, the ideas that the interviewee suggested(fingerprint, facial scanner) were those that are already in use and can be easily implemented but the moonshot ideas that you talked about in the feedback(for ex : a panel that pops out braille) is a relatively newer solution whose implementability is not yet known. So if i were to suggest both of these ideas in an interview, should I explicitly talk about the practicality of each idea or would that deviate from the discussion on steps of product design?
Hi Argos! Thanks for asking :)
It’s always good to acknowledge that certain things such as moonshot ideas might require more time of R&D, and that perhaps it’s not the immediate solution. You can always use that when prioritizing features, for instance you can say that:
You’ll prioritize simple features that will have a big impact fast and you’ll start R&D for your moonshot idea.
You only have 30-45 minutes so they key here is to be concise. As long as you and your interviewer are on the same page, it should be ok.
@@PMDiegoGranados Thank you!
Just attach a phone receiver 📞 on the ATM. The blind can speak on the receiver 📞 about the money he wants without letting others listen. The speaker on the receiver will tell the blind in his ears about the remuneration
My moonshot idea - If we build a new ATM, then it could be a small, insulated digital cabin, you enter the cabin, the door will be closed and you're in your "personal bank", secured from other eyes and noise. And since it's an isolated cabin, it could be right next to the store doors for easy access, not inside. The cabin itself is a voice assistant and the actions are on the front wall, partially or wholly touchable, you just need to raise your hand and find it. The assistant asks you to touch your card or place your fingerprint anywhere on the wall (considering different heights, etc..), which will fetch your data. You can tell what action you want to take, then get the money from the dedicated box. The box for the money could be adjusted vertically or automatically adjust with the height of the person. The experience resets when confirmed by the user or when the door is opened.
The Pros: this solves security, noise, safety issues and could reduce cognitive load (e.g. don't have to find the card or fingerprint box, etc. ). This will also work for any type of user.
The Cons: Considering it's digital and automated, it might need long-term high maintenance, and manual adjustments for emergencies, e.g. opening/locking the door manually etc.
My moon shot idea is on the card scan : My idea on the card scan will be : Design a pocket where user can place his hand along with the card for scan, and provide a finger print scan in the same pocket. think of this as a slot in the machine where user can place his palm holding his/her card. Another important part adding headphones for the voice commands,. No display on the screen for anything, just pure voice commands for safety.. Finally put the money in an envelope so it is not exposed to someone how much is withdrawn ...
I am still not comfortable with facial recognition for the banking transaction, the technology is great but it is not 100% accurate at this point and we need solid tech for the specialized community..
I like the privacy of not showing anything on screen!! Thank you for sharing these ideas!
Question - how would the atm identify if the user is blind? Once they put their hand in the pocket, how do you distinguish between someone visually impaired and someone who does need to see the screen?
diego i'd love to see you push back harder on him just to simulate a real interview. there were definitely times where he talked about medium effort but didn't exactly explain why and like you said we never went through the user journey which i think is key to the painpoint section as well. his structure was sound but like you sort of point out in the video the actual content is a bit lacking.
Moonshot idea: Bring the phone boot back!
ATM will be a video telephony system with webcams and the user's personal phone (assuming they are using traditional button phones) in a modified phone booth. The booth is pressure or motion activated and features a sound guide to assist passerby blind users inside (think crosswalks). The system will connect to an AI-powered Virtual Accessibility Representative (VAR) with computer vision, who will ask the user to validate their identity, purpose, and validate that they are alone. The booth will close once items are confirmed, and the soundproofed, smart-glass will become opaque. The transaction will be phone-based, using user's phone interface for pins, transaction amount, or confidential information authorization to add additional privacy/security (think 2FA/MFA), and once completed, the door will open up, and the smart-glass will become transparent again and the VAR will guide the user out.
You could probably use headphones to take care of how the system would interact with the customer. Headphones are something you can easily wear using just the feedback from your hands .
Headphones would have sensors near each of the ears which could help scanners further locate the position of the head more precisely to enable an accurate facial scan. These sensors could also help an AI based camera guide the customer better with respect to using the machine
That’s a very interesting idea! Thank you for sharing Yash!
Do you see any potential risks about using AI in a camera that is always tracking whomever is near the ATM? If yes, how would you solve it?
@@PMDiegoGranados love your channel and the practicality with which you do your videos. Thank you !
@@PMDiegoGranados Well yes, some may see this as a potential risk but this is where I believe the sensors on the headphones come into play. They could force the AI into only tracking the customer wearing them. Does that make sense ?
@@yashgupta3336 yes! Great clarification! When talking about AI, we have to be careful about users data or tracking. Using the headphones can be the way to opt in and consent to be tracked.
Before getting into the features, if he has gone thru the user experience, it could be broken up into . For a good portion intially, he focussed a lot on Authentication
1. Getting to the ATM
2. Authentication
3. Navgation to the function and select the function (Withdraw cash as the focus here)
4. Interact with the Function (how much to withdraw)
5. Get cash
6. Sign out
video quality is very poor, cant read tips