Leveraging Fidelity's Program Impacts

I leveraged Fidelity's AI ability and designed a B2B sales visualization dashboard for Fidelity program's impacts from concept to development, integrating LLMs and generative models, which streamlined Fidelity's clients' decision making process.
duration
4 months | 2023 | Contract
Tools
Figma, Midjourney, Chat GPT4, Adobe Illustrator
My Role
UX Designer and Project Design Lead
Teams
SCADpro Design Studio
2 VDs  |  4 UXRs  |  1 Project Manager
4 UX Designers
Fidelity Investment
4 Account managers  |  1 Product Manager
1 Lead Developer

◦ Fidelity's Product

'Student Debt' program - provided for enterprise and small businesses

A tax-benefit program for businesses to sign on to, supporting their employees in paying off student debt.

It is often considered as company benefit for attracting top talent by offering student debt contributions through Fidelity's account. Businesses can also benefit from tax exemptions.

◦ Key Issues for the Sales Team

Struggle to elicit client's empathy with sales pitch deck

The sales team has struggled to elicit empathy from clients to help them understand the potential impact of enrolling in Fidelity's program when presenting their sales pitch decks.

Through research, we
identified limitations within the sales pitch deck as it focuses on industry-wide metrics rather than specific, niche data tailored to the company.  

Due to being under an NDA, I can only present the sales pitch deck during interviews.

◦ My Role

UX designer and design lead

I am primarily responsible for translating the sales pitch deck data into more effective visualization solutions. In other areas of the project design process, I guided our design decisions by focusing on the problems we are solving, analyzing user testing and stakeholder feedback, clearly identifying our user groups and their behavior, prioritizing key challenges, and defining concrete goals.

◦ Research Insights

Clients struggle to see the value of the program

Fidelity's clients, the decision-makers (HR Managers / CEO / CFO / The Board) struggle to understand how the programs would benefit the company's performance beyond the tax benefits presented in the sales pitch deck.
Sales deck lacks personalized data

Evaluating the program's value
takes several months because it involves extensive discussions with the board and conducting surveys to assess employees' interest in using student debt repayment as an employee benefit.

This prolonged evaluation period is primarily due to
the sales deck relying on general market data, which lacks the specific, tailored information clients need to make well-informed decisions.
Who were we designing for? Fidelity's sales personnel and their clients.

Our goal was to transform the sales deck insights through data visualization and help the sales team communicate the impact of the student debt program effectively, whilst leveraging Fidelity's AI technology.

◦ Solution Highlights

Transforming sales data: from generic industry metrics to tailored enrollment insights

For Fidelity’s clients: HR managers, CEOs, and CFOs — this analytics solution provides comprehensive enrollment data for their organization’s participation in Fidelity’s student debt program. It also streamlines board-level decision-making around enrollment.

For Fidelity sales team: the sales team can utilizes this analytics for sales presentations and strategy adjustments. 

AI Chat Bot as Sales Representative

Integrated generative AI technology into our tool to emulate the responsiveness of a real salesperson which leverages Fidelity's scalability and captures insights from client interactions to continually improve the Fidelity student debt program.
Problem
Decision makers struggle to understand the challenges faced by employees burdened with student debt.
Solution
A personalized scenario powered by generative AI, through Midjourney API, enables users to visualize their workforce and explore the impact data as they apply different contributions the program.

◦ Dashboard Iterations Challenge

Optimizing visual affordances in data visualization - how to digest data in ease

Through our research, we identified key data points, such as 'total program cost,' 'net ROI,' and 'tax exemption,' along with their correlated data groups. We then integrated these insights, translating the data from the sales deck into our analytics dashboards.

Initially, we designed various chart types, such as line and bar charts, and iterated on them based on feedback from both Fidelity and the users.

Final Iteration based on data prioritization and user testing

◦ Challenge in scaling the AI technology

Evaluate the sales process and understand AI capabilities

The main goal of our project is to enable decision-makers to fully understand the program's impact at ease, which will help the sales team in speeding up and boosting deal closures. We discovered that generative AI is the most optimal to reach our goal, as it can represent real sales representative to analyze data and summarize insights.

3 months constraint

Based on the generative AI capabilities, we proposed chat gpt4 API implementation to the tool. Given the time constraint, We prioritized essential functionality, ensuring the design included key features like recommended prompt chips for each user-entered question.

◦ Challenge in designing for data transparency

Empowering Decision-Makers with Data Transparency

The decision-makers are Fidelity's clients, including HR managers, CEOs, and CFOs.

Our research insights revealed that by
providing data transparency how the student program impacts on both client's company performance and employees' performance, can expedite the decision-making process during board discussions. To achieve this, we have added "the settings" for sales personnel or clients to validate their data and algorithms.

Initial design - Diagramming or mapping visualization

Engineering constraints

The developer from the AI Incubator team pointed out that the initial design - the click-and-drag functionality on the canvas - was too complicated and time-consuming to build and asked opting for a more simpler design approach.

Our final design is to have a drop down component for each data point for analyst the evaluate the algorithm and ability to view the database.

Simplified drop down table for each data set

Be proactive early on

To improve the design efficiency, early check-ins with developers are crucial, especially given our time constraints. Being proactive not only can help with the alignments but also streamline the design process.