All work

Underwrite Assist

AI-assisted Commercial Underwriting

User Research Systems Thinking Agentic Design Interdisciplinary Collaboration UX/UI Design

Full Case-study

Context

Underwriting is the process of evaluating a client's risk profile and financial needs to match them with the right credit solutions.

Primary personas are the users of Underwrite Assist and secondary personas are directly effected by the outcomes of underwriting.

Primary Personas

RM

Relationship Manager

Engages with the client, gathers financial information and communicates loan terms.

FS

Financing Specialist

Does in-depth analysis of the clients historical and current financial state

Secondary Personas

Adjudicator

Performs an independent review of all credit applications & works with FS to understand their analysis.

Client

Commercial client seeking solutions for its business needs.

The Underwriting Journey

Underwriting happens for every client at least once a year to refresh their risk profile. It also happens when a client has new credit needs or when a new client comes on board.

Underwriting
01

Portfolio Monitoring

Actors
RM FS

GoalProactively identify client engagement touch points like upcoming annual reviews, covenant breaches, or interim requests.

02

Client Outreach

Actors
RM FS Client

GoalContact the client, understand their needs and gather documentation to begin financial risk assessment.

03

Financial Risk Analysis

Actors
FS RM

GoalAnalyze client's current and historical financial status, previous credit applications and adjudication commentary.

04

Deal Structuring

Actors
RM FS Client Specialist Teams

GoalStrategically match the client's needs and risk profile with appropriate credit products and loan terms.

05

Credit Application Submission

Actors
FS RM Adjudicator

GoalCreate a credit application and document the outcomes of steps 03 and 04 to convey the deal terms and financial rationale to the adjudicator.

06

Credit Adjudication & Decision

Actors
Adjudicator FS RM Client VPs

GoalAdjudicator conducts an independent analysis credit application and works with the FS to understand their anlaysis.

07

Documentation & Funding

Actors
RM Client Disbursement Teams

GoalFacilitate legal document signing and confirm final loan disbursement.

RM = Relationship Manager FS = Financing Specialist

My Role

Discovery,  Design and iteration

User Research

Understanding Underwriting

Mapped the underwriting journey with 12 relationship managers and financing specialists across industries and client sizes, covering their goals, frustrations and toolkit.

Understanding the system

Uncovered what happens before and after underwriting, which other personas it affects, and how needs differ for small, mid-market and large clients.

multidisciplinary ideation

Led workshops with ML research, engineering, product and business stakeholders to turn research findings into feasible, goal-driven, business-aligned features.

Design & Iteration

Tested concepts with 12 users to validate if Underwrite Assist met their goals, eased their frustrations, and where tasks broke down before release to 200 pilot users.

Product Management

Owning what gets built

Defined and communicated research-informed MVP features, prioritizing them with an effort/value matrix. Kept research, engineering and design aligned on that scope through to release.

Design debt

Documented and successfully resolved all design debt after MVP release to ensure any UX gaps were addressed before the next release.

Building partnerships

Commercial Design Team

Worked with the commercial credit team to understand how the toolkit is evolving and where Underwrite Assist fits alongside the new tools in development.

Business stakeholders

Worked with business stakeholders to prioritize features for the pilot release, prepare users with AI education and send out surveys to measure KPIs and user sentiment.

AI Education, marketing and user Feedback

User feedback

Ran the pilot feedback survey with the business and successfully advocated for fixing the major issues after release.

Marketing

Created demo videos shown at Commercial Banking and Gen AI townhalls, MVP launch calls and senior leadership meetings.

AI education

Acted as the link between business stakeholders and ML researchers, translating AI limitations and best practices for getting the best output from the agents into user guides and onboarding presentations.

Impact

Underwrite Assist launched as a pilot with 200 users and is now scaling 10x to 2,000 users. A survey of pilot users revealed the following impact:

2hours

Saved on average per credit application submission

86%

Rated Underwrite Assist “very trustworthy”

96%

rated the user experience as “intuitive and seamless”