AI Enabled Data Intake Redesign

A UX Case Study

My Role : Ideation & Design

Teams Involved : Product & PEC Teams

Timeline & Status : 1.5 Months | Iterative

Product Overview & Problem

Secondary Research

Foliosure is a portfolio management product enabling private market investors make more informed decisions by offering real-time insights through a distilled view of portfolio performance.
Private Equity (PE) Business Unit reported 18% delays in reports produced for clients (limited partner investors) using Foliosure. Foliosure was also losing out to competitors during client demos.

Quick summary

Context

Private equity firms were using Foliosure to track their portfolio performance. Foliosure's driving force is a cohort of financial analysts who manually format data into a proprietary excel template before uploading it on the tool.
This data then gets consumed through metrics, charts, logs or downloaded as customised reports.
As Foliosure scaled, manual data preparation became a bottleneck, delaying client reporting and weakening product market differentiation.

"In Private Equity time is

everything, it can cost you

clients"

Gradient

Outcome

Secondary Research

18% Reduction
In time spent on manual tasks owing to AI assisted workflow

Secondary Research

$200K Revenue
Generated with new client onboarding

Product Overview & Problem

Foliosure is a portfolio management product enabling private market investors make more informed decisions by offering real-time insights through a distilled view of portfolio performance.
Private Equity (PE) Business Unit reported 18% delays in reports produced for clients (limited partner investors) using Foliosure. Foliosure was also losing out to competitors during client demos.

Foliosure is a portfolio management product enabling private market investors make more informed decisions by offering real-time insights through a distilled view of portfolio performance.
Private Equity (PE) Business Unit reported 18% delays in reports produced for clients (limited partner investors) using Foliosure. Foliosure was also losing out to competitors during client demos..

Quick summary

Outcome

18% Reduction
In time spent on manual tasks owing to AI assisted workflow

18% Reduction
In time spent on manual tasks owing to AI assisted workflow

$200K Revenue
Generated with new client onboarding

$200K Revenue
Generated with new client onboarding

Context

Private equity firms were using Foliosure to track their portfolio performance. Foliosure's driving force is a cohort of financial analysts who manually format data into a proprietary excel template before uploading it on the tool.
This data then gets consumed through metrics, charts, logs or downloaded as customised reports.
As Foliosure scaled, manual data preparation became a bottleneck, delaying client reporting and weakening product market differentiation.

Private equity firms were using Foliosure to track their portfolio performance. Foliosure's driving force is a cohort of financial analysts who manually format data into a proprietary excel template before uploading it on the tool.
This data then gets consumed through metrics, charts, logs or downloaded as customised reports.
As Foliosure scaled, manual data preparation became a bottleneck, delaying client reporting and weakening product market differentiation.

"In Private Equity time is everything, it can

cost you clients"

"In Private Equity time is everything, it can

cost you clients"

Users & Jobs Involved

Analyst | Template Creator
Art that reveals how technology frames reality

1. Collects financial data from multiple client sources
2. Formats it into a propietary excel format.
3. Uploads it on tool to be validated by approver

Branding is the profound manifestation of the human spirit," says designer and Branding is the profound manifestation of the human spirit," says designer and Branding is the profound manifestation of the human spirit,"

Constant Context switching

Owns Data Collection

Approver | Data Reviewer
Art that reveals how technology frames reality

1. Reviews & makes corrections if needed to uploaded data.
2. Downloads reports

Branding is the profound manifestation of the human spirit," says designer and Branding is the profound manifestation of the human spirit," says designer and Branding is the profound manifestation of the human spirit,"

Focused workflows

Owns Data Audits

Exploratory Research

Secondary Research

Market research & a competitive audit of 16 products across 9 parameters revealed that 75% of products already automated data collection and 69% offered AI capabilities.

Secondary Research

Market research & a competitive audit of 16 products across 9 parameters revealed that 75% of products already automated data collection and 69% offered AI capabilities.

Secondary Research

Market research & a competitive audit of 16 products across 9 parameters revealed that 75% of products already automated data collection and 69% offered AI capabilities.

Primary Research

14 Hypotheses Tested Under 3 Categories

14 Hypotheses Under 3 Categories

6 Core Journey Tasks

8 Users

Trust matters more than automation

Recommendations needed to be reviewable & editable before they could be trusted.

Workflow was the real problem

Context-switching & repetitive validation were symptoms of process design rather than user behaviour

Transparency is essential for adoption

Confidence could increase when data lineage and system reasoning were visible

How might we streamline the journey from financial documents to investor ready insights?

Secondary Research

Which led to our problem statement

Which led to our problem statement

How might we streamline the journey from financial documents to investor ready insights?

How might we streamline the journey from financial documents to investor ready insights?

Design Direction

Design Principle

Continuity

Analysts should be able to complete the process within a single, uninterrupted workflow.
Design Principle

Efficiency

AI should reduce manual effort while keeping analysts in control of decisions and outcomes.
AI Enablement

Assistance

AI should support decision-making rather than replace analyst judgment.
AI Enablement

Recoverability

AI failures or low-confidence outputs should never prevent workflow completion.
Constraint

Architecture

The solution needed to operate within Foliosure's existing backend structures and data model
Constraint

Compliance

AI-assisted outputs still needed to meet existing validation, audit, and data quality requirements

Before Workflow Improvements

Product Modernisation

In parallel with the workflow redesign, the product received a visual refresh and navigation overhaul to improve discoverability, create a more cohesive experience, and better align with industry standards.

Product Modernisation
In parallel with the workflow redesign, the product received a visual refresh and navigation overhaul to improve discoverability, create a more cohesive experience, and better align with industry standards.


While the structure improved visibility, it introduced friction when analysts needed to revisit earlier stages, making the workflow feel restrictive rather than fl

Entry Points Considered

Concept 01 | Conversational AI Entry Point
The home screen was explored as an AI-first entry point where analysts could retrieve financial statements and portfolio data through natural language prompts.

Trade-off

While promising from an experience perspective, the concept depended on technical capabilities that could not be reliably supported within the existing architecture and project timeline.

Concept 01 | Conversational AI Entry Point
The home screen was explored as an AI-first entry point where analysts could retrieve financial statements and portfolio data through natural language prompts.


Trade-off

While promising from an experience perspective, the concept depended on technical capabilities that could not be reliably supported within the existing architecture and project timeline.While the structure improved visibility, it introduced friction when analysts needed to revisit earlier stages, making the workflow feel restrictive rather than flexible.

"Ask for data, don't search for it."

Concept 02 | Enhanced Bulk Upload
An alternative direction focused on refining the existing Bulk Upload module rather than restructuring the workflow.

Trade-off

Research showed that upload was only one step in a much larger process. Optimising uploads would reduce friction locally but would not solve the broader challenges around validation, mapping, and reporting.

Concept 02 | Enhanced Bulk Upload
An alternative direction focused on refining the existing Bulk Upload module rather than restructuring the workflow.


Trade-off

Research showed that upload was only one step in a much larger process. Optimising uploads would reduce friction locally but would not solve the broader challenges around validation, mapping, and reporting.While the structure improved visibility, it introduced friction when analysts needed to revisit earlier stages, making the workflow feel restrictive rather than flexible.

Concept 03 | Dedicated Data Ingestion Workspace
The workflow was reframed as a dedicated Data Ingestion module that brought collection, extraction, validation, and submission into one continuous experience.

Outcome

This direction best aligned with user needs around continuity, ownership, and transparency, and became the foundation of the final solution.

Concept 03 | Dedicated Data Ingestion Workspace
The workflow was reframed as a dedicated Data Ingestion module that brought collection, extraction, validation, and submission into one continuous experience.


Outcome

This direction best aligned with user needs around continuity, ownership, and transparency, and became the foundation of the final solution.While the structure improved visibility, it introduced friction when analysts needed to revisit earlier stages, making the workflow feel restrictive rather than flexible.

"A single home for the entire intake lifecycle"

Concepts Considered

Concept 01 | Stepper-Based Flow
The workflow was broken into clearly defined stages to help analysts understand their progress and reduce uncertainty during a complex ingestion process. Each step was completed independently before progressing to the next.


Trade-off

While the structure improved visibility, it introduced friction when analysts needed to revisit earlier stages, making the workflow feel restrictive rather than flexible.

Concept 01 | Stepper-Based Flow
The workflow was broken into clearly defined stages to help analysts understand their progress and reduce uncertainty during a complex ingestion process. Each step was completed independently before progressing to the next.


Trade-off

While the structure improved visibility, it introduced friction when analysts needed to revisit earlier stages, making the workflow feel restrictive rather than flexible.

While the structure improved visibility, it introduced friction when analysts needed to revisit earlier stages, making the workflow feel restrictive rather than flexible.

"Transparency through sequential stages"

Concept 02 | Modular Ingestion Hub
This concept consolidated the ingestion process into a single workspace where analysts could manage requests, track statuses, and move between stages without leaving the workflow.


Trade-off

While less prescriptive than the stepper-based approach, persistent breadcrumbs (Data Ingestion → Fetch File → Data Extraction) provided orientation without restricting navigation, resulting in a more flexible experience.

"Continuity through a unified workspace"

Final Direction | Modular Centralised Ingestion Hub
Testing showed that analysts valued flexibility over guided progression.

Why it won?

  1. Reduced perceived efforts

  2. Breadcrumbs provided sufficient orientation

  3. Simplified source selection and reduced manual work

Design Decision 01 : Streamlining File Selection


Design Decision 02 : Supporting Workflow

Continuity

Concept 02 | Modular Ingestion Hub
This concept consolidated the ingestion process into a single workspace where analysts could manage requests, track statuses, and move between stages without leaving the workflow.


Trade-off

While less prescriptive than the stepper-based approach, persistent breadcrumbs provided orientation without restricting navigation, resulting in a more flexible experience.

Concept 02 | Modular Ingestion Hub
This concept consolidated the ingestion process into a single workspace where analysts could manage requests, track statuses, and move between stages without leaving the workflow.


Trade-off

While less prescriptive than the stepper-based approach, persistent breadcrumbs (Data Ingestion → Fetch File → Data Extraction) provided orientation without restricting navigation, resulting in a more flexible experience.

While less prescriptive than the stepper-based approach, persistent breadcrumbs (Data Ingestion → Fetch File → Data Extraction) provided orientation without restricting navigation, resulting in a more flexible experience.

"Continuity through a unified workspace"

Final Direction | Modular Centralised Ingestion Hub
Testing showed that analysts valued flexibility over guided progression.

Why it won?

  1. Reduced perceived efforts

  2. Breadcrumbs provided sufficient orientation

  3. Simplified source selection and reduced manual work

Design Decision 01 : Streamlining File Selection


Design Decision 02 : Supporting Workflow

Continuity

AI Enablement Through Workflow

Old Workflow
Manual Data Collection from client sources -> Manual Reconciliation & Compilation -> Manual Ingestion In Foliosure -> Streamlined Data




New Workflow

Shape

AI

Google Cloud icon
Bloomberg icon
Icon
Dropbox symbol
Outlook icon

Streamlined Data

Shape

AI

Google Cloud icon
Bloomberg icon
Icon
Dropbox symbol
Outlook icon

Streamlined Data

Old Workflow
Manual Data Collection from client sources -> Manual Reconciliation & Compilation -> Manual Ingestion In Foliosure -> Streamlined Data




New Workflow

Shape

AI

Google Cloud icon
Bloomberg icon
Icon
Dropbox symbol
Outlook icon

Streamlined Data

Final Direction | Modular Centralised Ingestion Hub
Testing showed that analysts valued flexibility over guided progression.

Why it won?

  1. Reduced perceived efforts

  2. Breadcrumbs provided sufficient orientation

  3. Simplified source selection and reduced manual work

AI-Assisted Identification

AI identifies and maps KPIs from uploaded financial documents, replacing manual setup with a pre-populated workspace

Intelligent Financial Spreader

AI extracts and maps KPI values, reducing manual data entry and validation effort.

Designing the Financial Spreader

Concept 01 | Full Extraction, Analyst Reviews
AI generated a completed mapping for analyst approval.


Trade Off

Outputs lacked source visibility, making them difficult to trust and verify.

Concept 01 | Full Extraction, Analyst Reviews
AI generated a completed mapping for analyst approval.


Trade Off
Outputs lacked source visibility, making them difficult to trust and verify.

Outputs lacked source visibility, making them difficult to trust and verify.

Concept 02 | Side By Side Extraction With Source Reference
AI-generated values appeared alongside the source document, with direct links back to the originating line item.

Trade Off

Increased information density, but aligned closely with how analysts already worked.

Final Concept | Side-by-Side Extraction with Source Reference
Analysts consistently preferred the side-by-side model. Direct traceability transformed AI extraction from a black box into a transparent and verifiable workflow.

Concept 02 | Side By Side Extraction With Source Reference
AI-generated values appeared alongside the source document, with direct links back to the originating line item.

Trade Off

Increased information density, but aligned closely with how analysts already worked.

Final Concept | Side-by-Side Extraction with Source Reference
Analysts consistently preferred the side-by-side model. Direct traceability transformed AI extraction from a black box into a transparent and verifiable workflow.

Happy Path
An illustration of analyst-approver workflow

Product Transformation

The redesign transformed data ingestion from a manual upload process into a continuous, AI-assisted workflow—reducing effort, improving trust, and creating a stronger foundation for future AI capabilities.

Concept 01 | Stepper-Based Flow
The workflow was broken into clearly defined stages to help analysts understand their progress and reduce uncertainty during a complex ingestion process. Each step was completed independently before progressing to the next.


Trade-off

While the structure improved visibility, it introduced friction when analysts needed to revisit earlier stages, making the workflow feel restrictive rather than flexible.

While the structure improved visibility, it introduced friction when analysts needed to revisit earlier stages, making the workflow feel restrictive rather than flexible.

Get in touch

Email : anushkapandeycontact@gmail.com

WIP, Always

Get in touch

Email : anushkapandeycontact@gmail.com

WIP, Always