Building Lexis+AI.
When Startup speed meets Corporate life
Building Lexis+AI.
When Startup speed meets Corporate life
Building Lexis+AI.
When Startup speed meets Corporate life
When a global industry leader like LexisNexis forges into the AI field, with 3.8 petabytes of data (around 790 times the size of Wikipedia), you can understand why navigating content is, well, difficult.
When a global industry leader like LexisNexis forges into the AI field, with 3.8 petabytes of data (around 790 times the size of Wikipedia), you can understand why navigating content is, well, difficult.
When a global industry leader like LexisNexis forges into the AI field, with 3.8 petabytes of data (around 790 times the size of Wikipedia), you can understand why navigating content is, well, difficult.
Client: Lexis Nexis Legal Solutions
Time Frame: 6 months
Type: Internal Design and Strategy
Client: Lexis Nexis Legal Solutions
Time Frame: 6 months
Type: Internal Design and Strategy
Client: Lexis Nexis Legal Solutions
Time Frame: 6 months
Type: Internal Design and Strategy
Workflow
Identify the user need
Validate with research
Build initial wireframes
Validate designs
Implement
Metrics
Trust Score on AI-generated content
Net Promoter Score (NPS)
Onboarding Efficiency
Customer Satisfaction (CSAT)
Team
George Adams, Research
Zaina Ali, Design
Andreea Bralia, Design
Natalie Baillie, Design
Madeliene Graham, Product
Product
Lexis Create for Microsoft Word
Legal professionals are consumed with information-dense legacy process wrapped in tradition, impending deadlines lost in billable hours with the pressure of perhaps changing someones life immeasurably for years to come. With AI, technology that presents visceral risk to the profession for those who don’t adapt fast enough.
Building a responsible platform that is adoption and usability was key. We needed to make sure that people who weren’t necessarily tech savvy had confidence and trust in what we were building for them. They needed to understand where the generated content came from, be able to cross reference it. The last piece of the puzzle was that we needed it to be fast. It had to be faster that what they could do themselves.
Building a responsible platform that is adoption and usability was key. We needed to make sure that people who weren’t necessarily tech savvy had confidence and trust in what we were building for them. They needed to understand where the generated content came from, be able to cross reference it. The last piece of the puzzle was that we needed it to be fast. It had to be faster that what they could do themselves.
Building a responsible platform that is adoption and usability was key. We needed to make sure that people who weren’t necessarily tech savvy had confidence and trust in what we were building for them. They needed to understand where the generated content came from, be able to cross reference it. The last piece of the puzzle was that we needed it to be fast. It had to be faster that what they could do themselves.
93%
Felt confident to start exploring Lexis+AI for their daily document drafting workflow
7+ Hours
Average time saved per week in calculations with initial beta testers
12,000
Participants surveyed in the first 6 months of the exercise
^750%
Orders of LexisNexis products in Microsoft with addition of AI features
Intense research investment and picking the low hanging fruit meant we were able to get a viable product published. The most important thing was building trust between AI and the users.
Intense research investment and picking the low hanging fruit meant we were able to get a viable product published. The most important thing was building trust between AI and the users.
Intense research investment and picking the low hanging fruit meant we were able to get a viable product published. The most important thing was building trust between AI and the users.
There were 3 steps to build trust in these responses:
1. Qualify the cases that the AI was referencing, making sure that the user could click through to verify everything that the AI was computing.
2. Minimise hallucinations by using our own unmatched library of content written and verified by lawyers and legal professionals, and making sure we limited conversations to concise and clear conversations, sometime forcing the user to start new conversations.
3. Guide the user, but don’t drive the user, which meant the navigation had to be clear, offer suggestions but make sure that the user could easily override them.
There were 3 steps to build trust in these responses:
1. Qualify the cases that the AI was referencing, making sure that the user could click through to verify everything that the AI was computing.
2. Minimise hallucinations by using our own unmatched library of content written and verified by lawyers and legal professionals, and making sure we limited conversations to concise and clear conversations, sometime forcing the user to start new conversations.
3. Guide the user, but don’t drive the user, which meant the navigation had to be clear, offer suggestions but make sure that the user could easily override them.
There were 3 steps to build trust in these responses:
1. Qualify the cases that the AI was referencing, making sure that the user could click through to verify everything that the AI was computing.
2. Minimise hallucinations by using our own unmatched library of content written and verified by lawyers and legal professionals, and making sure we limited conversations to concise and clear conversations, sometime forcing the user to start new conversations.
3. Guide the user, but don’t drive the user, which meant the navigation had to be clear, offer suggestions but make sure that the user could easily override them.
"You can’t accomplish anything meaningful by incremental thinking; the next Bill Gates will not build an operating system. If you’re going to start something new, it needs to be 10x better than what already exists.”
"You can’t accomplish anything meaningful by incremental thinking; the next Bill Gates will not build an operating system. If you’re going to start something new, it needs to be 10x better than what already exists.”
"You can’t accomplish anything meaningful by incremental thinking; the next Bill Gates will not build an operating system. If you’re going to start something new, it needs to be 10x better than what already exists.”
Peter Thiel ‘Zero to One’
Peter Thiel ‘Zero to One’
Peter Thiel ‘Zero to One’
I’m here to
help you.
I’m here to help you.
I’m here to help you.
Send me an email
Send me an email
Send me an email
Stevie Rodger
Fun fact: Stevie started his career designing kids surf wear in Australia.
Stevie Rodger
Fun fact: Stevie started his career designing kids surf wear in Australia.
Stevie Rodger
Fun fact: Stevie started his career designing kids surf wear in Australia.