AI Autofill

AI Autofill

Save hours of manual data entry by automatically populating fields with AI

Save hours of manual data entry by automatically populating fields with AI

Save hours of manual data entry by automatically populating fields with AI

TEAM MEMBERS

3 product managers, 4 engineers, 5 data scientists, 1 legal domain expert

TIMELINE

MVP design: 2 months; MVP build: 9 months

Background

Background

EvenUp is a legal AI company that specializes in personal injury law. The company was expanding from AI-powered document generationt tool into an end-to-end case management platform for personal injury law firms.

The problem

The problem

Each matter can contain hundreds of fields that require manual data entry, creating repetitive work for law firm staff and leaving critical case information incomplete.

The solution

The solution

AI Autofill automatically populates and updates matter fields using information already available across the case. Admins configure where and how AI autofill will run, while law firm staff get the information they need without repetitive data entry.

SOLUTION 1. CROSS-FIRMS AND SPECIFIC FIRM CONFIGURATIONS

SOLUTION 2. EVENUP DEFAULT AND CUSTOMIZED PROMPTS

SOLUTION 3. HOW AUTOFILL SHOWS UP IN MATTERS WITH CITATION DRAWER

Define the MVP

Define the MVP

When I joined, months of exploration had already happened, but the team was struggling to converge on a shippable product. I helped turn that ambiguity into a focused MVP by focusing on the three more important questions.

QUESTION 1: WHICH FIELDS FOR MVP?

QUESTION 2: HOW WILL AUTOFILL BE TRIGGERED?

QUESTION 3. WHEN HUMAN REVIEW REQUIRED?

Design the ideal for review experience

To define the review experience for MVP, I started with the ideal and explored many different directions before we aligned on the "companion" option being the ideal. Knowing that the "in line card" option though viable for MVP, is a toss of engineering work, we eventually decided to deprioritize review experience for MVP.

REVIEW UX OPTION 1: REVIEW MODE

REVIEW UX OPTION 2: IN LINE CARD

REVIEW UX OPTION 3: COMPANION

Paint vision for AI Autofill as a product

As we are building towards the MVP, I collaborated with the VP of Product on AI Autofill vision to align the company on the future of the feature and enable a smoother transition between the original and new product owners. In the vision, we highlighted 3 main improvements we want to invest in.

Inform what's next based on MVP

After we released some aspects of the MVP, we have been hearing relatively mellow response from the feature. I took the initiative to understand the customer reaction with data dive. Then I conducted ~6 user interviews and help inform the priorities for Q3 planning incorporating the AI Autofill vision exercise.

INSIGHT: NOTES AS A SOURCE AS BASELINE FOR TRUST

We learned that notes often contain more up-to-date info than documents. Therefore, supporting notes as a data source to populate fields ensures AI Autofill is always producing the right value, which is fundamental to user trust.

INFORM PRIORITIES FROM AUTOFILL VISION

With the user feedback, we made a more informed decision around the Q3 priorities of the 3 investment areas from AI Autofill vision: with expand coverage and flexiblity being top priority, and transparency and user control being the lowest.

Impact

Impact

Since launching the first release in May, as of end of August, we already filled 68,174 fields for the 2 initial customers for the brand new EvenUp Case Management product called System of Work, saving 151 hours for these firms if assuming each field takes 8 seconds to fill.

AI Autofill also became a major selling point in closing one of our largest System of Work customers recently.

Next project: audit log