Interviewee 1:

1. Engineer, not fluent in English legal jargon 
   1. Personal: 
      1. Will detaily read all legal documents
      2. Decrease time of reading
      3. Common legal documents
         1. Tax
         2. Insurance
         3. Trust
         4. Will
         5. Real estate
      4. Will receive a legal document every 3 months
      5. Will definitely double check with other key documents before 100% trusting
      6. Will translate in paragraphs, not the whole document
         1. Don’t believe that the whole document will be translated accurately
         2. But if the whole document will translate accurate and efficiently, will chose to translate the entire document
      7. Will highlight important points
         1. Related to the specific industry
         2. Original and translated double highlight the important parts
      8. Concerns of privacy
         1. Because the contracts are private
         2. Save the previously past history in personal account
      9. Cost:
         1. If per document – does the document’s length matter
         2. Does not use a lot for annual
      10. Ads on the website doesn’t matter
          1. only pertaining to video
          2. No pop ups PLEASE
   2. Business:
      1. No a lot of usage

 

Interviewee 2:

1. New Grads
   1. Ask parents to read it
   2. Need translator to have good reviews
   3. Assume that is secure
   4. Delete the documents 
   5. Purchase of big assets
   6. $100 – per document 
   7. Lawyer would know better 
   8. Would absolutely hire a lawyer if a lot of money were on the table
   9. watch a 2 min ads to receive a free document 

 

Interviewee 3:

1. Banker at HSBC
   1. Personal
      1. Common legal documents
         1. Insurance:
            1. Car
            2. Medical
            3. Tickets
         2. Real estate
         3. Large asset purchase
      2. Twice a year will perceive these documents
      3. Time consuming
      4. Need credibility
         1. Will use on smaller documents
      5. Translate the entire document
         1. Highlight by taking them out and separately referring to them at the end of document
      6. Initially may have concerns but if everyone uses it will believe it
      7. Safe majority of document
         1. But stuff like be SSN be erased
      8. Per document payment
         1. More important:$100 
         2. Less important:$20
      9. Ads watching is ok
         1. Have warning
   2. Business
      1. When you open a new account there is a customer disclosure
         1. Bank and client agreement
         2. Include a lot of charges, financial jargon
      2. Clients may be happy due to better service
      3. 10% customers may need this support
         1. 5 – 6 people a day – on the higher end of spectrum
      4. Google glasses has a scanning and translating device development
      5. Current resolution:
         1. Call service
         2. Banker diversity (bilingual bankers to help out

 

Interviewee 4:

Amazon warehouse worker, 55 y/o, ESL

• Documents are too long and too dense, can be confusing. Maybe what we’re trying to do can help out especially people who don’t have time to read details
• Main thing is that it can summarize without leaving out important details
• Helps out older people without losing meaning
• Sometimes when you sign agreements, you’re not gonna read every word. Can be useful for things that people generally skip
• Can be useful––she would pay for it 
• Not concerned about data 
• A lot of people my age aren’t concerned about data! (Older/younger people)
• Older people develop trust with Facebook

 

Interviewee 5:

STS Major at Stanford:

• Skepticism about changing the specificities of the meaning
• Deceptive and non-deceptive legal documents might be mixed by the AI
• The main thing is to not lose meaning–– really need to research what the deceptive methods in legal documents are that actually exist so that technical meanings are not lost in translation
• There’s a reason documents are long––because they’re very specific
• Would rather hire a professional than use the tool
• Tiering system is similar to other things that she’s seen
• Market for english speaking people who live in other countries
• Tier idea: Free: x-amt of documents, middle: every additional after free costs x amt if you choose not to do subscription model, premium: unlimited documents
• Make sure to reference reasonings
• Older millennials are more concerned about data and privacy
• Give people a choice between storing and destroying their documents
• As someone who’s data-conscious, it’ll be enough for a disclaimer that says we don’t store the data. Would prefer the option but understands data used to teach algorithms but no selling or giving away
• Grammarly interface
• Under the yes option, elaborate why our model would be benefit from storing data
• Mobile-app (can be short-term or long-term goal, interchangeable with website)-everyone has a phone and it’s accessible; but website is better so you can have features like documents side by side
• Medium-term: paraphrasing; long-term: very thorough translation with accuracy
• Whenever she suggests something for business that’s beneficial, people seem to put it on backburner (to reprioritize longterm goals)–– give people customizability with accs

 

Interviewee 6:

Econ Major at Stanford:

• He knows a person that’s working on a Startup called “Tome”–– feeding legal documents and it regurgitates the basic term sheet. Doesn’t know if it has an MVP or not
• Cultural standpoint: Two summers ago, he worked at a small business administration that worked to figure out how to improve AAPI citizens’ accustomably
  • Language and technology barrier are huge. Maybe affluent immigrants would be able to take advantage of this, but if a refugee or older person they may have a difficult time using the software
• One thing to ask yourself: Will businesses and well-educated people have concerns about data privacy? Suggests looking at other companies like DocuSign and see how they deal with very similar issues, since you do have to upload your contract
  • Research the issues of data privacy storage here
• Parents were poor affluent immigrants coming to America, now are more fluent and wealthier–– would rather pay for a tool than a lawyer
• The tool would be cool if it could also write things up in legal contract form (long-term)
  • Useful for things like contract renewal 
• Maybe do by word-count 

 

Interviewee 7:

MS&E Major at Stanford:

• B2B is a potentially more profitable market, but you would be a lot more likely to get sued unless you have a really strong disclaimer
  • Selling to individuals will have you less likely to pursue legal actions
  • Justify that both are very strong: Businesses would be more inclined to pay for the service bc they have more documents that they need to scan
• GPT-3 charges by tokens generated (could charge by token instead of document, which is like a syllable)
• You can’t bring a crappy product to the market (short-term) so agrees with just trying to highlight and emphasize key terms
  • Think the product can be a bit more bare-bones to begin with 
  • Thinks that a product is only worth releasing if short-term is very effective, otherwise for mid-term and long-term 
• Features in mind:
  • What if it could write up a summary of a video
  • Program– feed it a textbook and generate test questions
    • Teachers spend a lot of time making test questions and writing a new test, it’d be way more time efficient if they could just curate the answers

Interviewee 8:

Background: Upper class, highly educated family— CS major — has significant medical 

Paperwork

In most case she just consents things 

Usually only reads things when regarding finances

Usually signs medical documents

Ore recently have been running into financial documents but while she reads it, she asks her dad first – trust him over her – she trusts him more because he usually takes care of stuff before – doesn’t trust herself into investing with her family being invested into – plans to take more of this

Would be hesitant to use AI since she does not trust it – if she has the capabilities to do than there are few things that she would pay for even if she could afford it – Withtime she’d trust AI like how she doesnt trust a car that drives its self – She would need a human in the process at some time – She is concerned about algorithmic basis – Field is dominated but white men so time is necessarily in diversifying the text – She trusts n the insitution, what are the chances your doctor is trying to harm you? – if her husband wanted a prenup, she wanted to try it out before she can you – reputation matters and she would be ok with a free trial or ads just to try things out – needs referral to trust – Would like to see reviews of people 

Interviewee 9:

Background: FLI student studying CS in the process of obtaining citizenship – Immigrated from Vietnam and is the “Acting” older sibling – deals with significant paperwork – parents know limited English

Would trust the system if someone told them about it – Would not know whether or not what was outputted is true or false – wants to see how things work internally – this would help in digesting but also signs a lot of paperwork without looking 

Recognizes that paperwork processing is a lot of of money but does not know if this would necessarily help elevate cost or contribute to it

More helpful when they first had stuff due to language barriers, but most places give forms in multiple languages

Documents are also closer to forms than they are documents so having options to edit the pdf and add text 

Probably would not pay a significant amount to use until they establish its features and usefulness

 

Reflection and Key Insights

Overall, it seems that most of the people that we interviewed were generally optimistic about the tool. Initially, we were apprehensive that people’s concern for data privacy would be a hindrance in the success of our product. However, we learned that the population of people who are actually data-conscious are relatively small. Because of this, we will put a less of an emphasis on optimizing our user interface in the short-term (to have users trust our product), and will focus more on adding features to our product. Moreover, another user corroborated this by suggesting that ads on the website would not affect whether they found the website to be too sketchy to use. While privacy was less of an issue, accuracy was a concern for many. There were prevalent concerns with whether users trust the effectiveness of our tool, especially if they were paying for it. We found that even a lower-income user who has dealt with the burden of being an immigrant child would likely not pay much for the tool until it proved to be effective without a doubt. This suggested a larger need for a referral or review system. People were also hesitant to price it without first using it for free to identify whether or not it fit with their needs. This solidifies our need for a free tier. We also got several perspectives on the pricing model for our product. Some of our suggestions include a system similar to GPT-3 and constraining users by a set number of tokens, an elaborate three-tier subscription model, and pricing by the document. 

 

Participatory Roadmap

 

Competitors (4 – 6)

Axis of 2×2

 

This 2×2 competitors’ matrix demonstrates the functionality of our market. As we can see from the graph, all existing digital competitors either focus on language translation or sentence paraphrasing (specifically focused on homework help). As a result of their generic focuses, all existing tools translate casual everyday languages. However, one of our main competitors – law firms – conduct very similar practices to our product. They have just as much expertise in professional languages and can very fluently explain hard concepts. Here is the main difference between Sign it! And law firms are able to translate texts into almost any language, while it is impossible for any law firms to perform this translation task.

 

 

 

We further examine the market space of a professional language simplification and translation product, we create the above graph. It is very clear that there is a lack of similar products/services out there. It is also very important to note that our product is vastly different from one of our strongest competitors – lawyers. While law firms have higher credibility than our product in its initial stages, it is vastly more expensive and slow in providing its services. We on the other hand, rely on models and computers to run all services thus we are able to provide cheaper and faster translations. Another important feature, our product holds is that with time and practice, our model will be more and more accurate in language simplification and translation, perhaps even surpassing human abilities.