Amrit Kumar
Product Designer
Below area has horizontal scroll within scroll. Use the side space to get scroll on full page.
Project 1
B2B SAAS Product
Review of similar contract documents
Lawyers often find themselves reviewing two similar documents multiple times, even when the content is largely identical. This redundancy leads to inefficiencies in the review process. AI technology can analyze the content and context of documents, identifying similarities and assigning them a similarity score.
This score offers users contextual information about the extent of similarity between the documents. This project aims to leverage this technology to streamline the review process for lawyers, minimising the need for repetitive review of similar documents.
AI can analyse the content and context of documents to identify similarities, assigning them a similarity score
When the system identifies similar reviewed documents, users experience the following.
Assumptions
Mental Status:
Users may feel reassured by the similarity between docs, leading to a sense of familiarity and confidence in the review process.
Queries:
Users may have questions regarding the degree of similarity, the status of reviewed values, and any potential discrepancies or differences between the documents.
Below is the mind map to capture the thoughts user might have after seeing the AI summary/prompt
Conclusion of mind map exercise
Users have lots of why on similar and dissimilar values?
User acknowledges the time-saving aspect but legal is all about the data accuracy.
Allow user to revert their choices.
After understanding users mind, i audited the current design.
Finding gaps in older existing designs
Existing design pain points
Locating the appropriate document for copying values is challenging.
The UI design is fragmented, exacerbating the complexity of this task.
The copy metadata pop-up presents multiple values without clear segregation.
There is a gap in user experience when striving for quicker reviews with similar documents.
The copy metadata pop-up presents multiple values without clear segregation.
There is a gap in user experience when striving for quicker reviews with similar documents.
Design goal
Design a solution to reduce review effort by leveraging previously reviewed values from similar documents.
The initiation and visibility of this feature
We recognise the gain in review efficiency this feature can provide.
The effort of browsing through each extracted values, making decisions, and processing value corrections can now be skipped.
So, It should be prominently communicated to encourage users shift their natural/current document review behavior.
Below is the user flow illustrating where this feature should integrate.
Designing the first idea
Showing values doc v/s doc
We designed our initial concept to allow users to easily see values between documents, enabling them to quickly identify similarities and differences.
Both document and its values are displayed on UI.User has full visibility of values in respective document.
Tested the above design with internal lawyers team.
Conclusion
It's cumbersome to match each values for similarity and dissimilarity.
Lot of data to compare.
Increases the human effort.
To reduce the data clutter.
Implemented an UI affordance to highlight dissimilar values.
Dissimilar Values On Doc A and Doc B is highlighted on UI
To further simplify the information,
Learned the following insights in a feedback session.
The focus needs to be on enhancing review efficiency.
Allow user focus by
Segregating similar and dissimilar values completely.
Possible solution
Display the values on specific tabs "Similar Values" and "Dissimilar Values.
Allowing user to compare one value type and make right choices of data.
Conclusion
The design here supports viewing similar values, segregating the value types.
Reduces context switching.
Press sync and this will mark all the similar values as review complete for current document.
But, how to help lawyers validate these values?
So we know, Lawyers need to view the surrounding text to validate data correctness and understand the context.
On hover user can see the value and its proximity text, This helps user decide its correctness.
From cross icon user may choose to delete the values which seems wrong for syncing.
On hover the card has values in two similar documents.Letting user compare between Doc A v/s Doc B
To assure that data accuracy is not compromised, created these below check points.
1. Data selection should happen from the closest reviewed similar document.
2. System should auto segregate dissimilar and similar values with design in support.
3. Display details of extracted from context page number to dig deeper.
4. The Hover card offers visual comparison, helping for decision-making.
5. Upfront indication of a reviewed tick (Review status) on values, Reassures that this value is prior reviewed.
Since similar values use case is covered
Further we needed clarity on dissimilar values
Should we allow syncing of dissimilar values?
How dissimilar values can add in gaining review efficiency?
To test these thoughts.
I assumed the cases that creates dissimilarity in values and created the below list.
Assuming reasons for dissimilarity giving user the reason to decide why the value is dissimilar.
I had three ideas to proceed
I collaborated for a brainstorm session with product and other stakeholders.
After the session we got these pointers
The final flow
The flow is reiterate as we discovered syncing dissimilar is not helping for review efficiency.
It should be view only.
Designing components
The List v/s Card
To show the list of document similar to current document. I redesigned it to be more informative and showing only relevant info.
For next iteration of above component,
Made it more action oriented.
I noticed that to initiate sync data there are two criteria which user looks to choose best document
1. The document must be having highest similarity score.
2. The document must be in review complete state.
So redesign version has auto selected document out of whole list and added bifurcation of other similar document.
Prototype walkthrough Video
view in laptop (recommended)
Exploring card visuals
Landing page of similar documents when none of similar docs are reviewed.
View Project 2 : AE Turbo.
View in Laptop/Desktop Recommended.
Project 2
Contract Review with AE Turbo
AE Turbo enhances lawyers' contract review with an AI-powered extraction model, reducing data errors and boosting efficiency. The sleek, user-friendly design ensures minimal learning curve, guiding users through each step for error-free, effective review.
Figma embedded link below (Similar docs review) you may open it in an new tab from a link given below.
or click expand icon from right corner while hovering inside embedded areaThis below area is scroll inside scroll. Scroll on side to get full page scroll
Project 3 (Self initiated)
Temperature Monitoring Dashboard concept.
