Work
Work
About
About
Automating business reports with AI-generated insights
I led the design of an AI-generated reporting experience that transforms raw data into structured, actionable insights — reducing analysis time and enabling faster decision-making across devices.
Lead Product Designer
Email Report
2024
Problem
Customer Success Managers (CSMs) at CloudKitchens use InsightGPT to create customer-facing reports with strong results—80% open rates and an 8% GMV uplift. But the manual process doesn’t scale, limiting access to a wider audience.
On the customer side, report format is also text-heavy and hard to scan, insights were buried in raw data, actions are not highlighted enough, creating a poor reading experience we needed to improve.
Manual process of creating business report
Step 1
Step 2
Step 3
Step 4
Step 5
Solution
Design a fully automated system that delivers bi-weekly GPT-powered reports directly to licensees-unlocking exponential value for both internal teams and end-customers.
Discovery
We ran several working sessions with the CSM team to understand current workflows and define the core requirements for automation. Together, we reframed the format to be Critical Data→ AI Comparison/Summary →Actionable insights.
Also, we defined the Minimum Viable Product (MVP) for the automated report, which includes:
- Business Summary
- Top Items
- Customer Ratings & Reviews
- Active Promotions
Design solution
MVP data module
Redefined data modules: Business Summary, Top Items, Customer Ratings & Reviews, Active Promotions
Reframed module structure: Critical Data→ AI Comparison/Summary →Actionable insights.
Email report
As a first milestone, we replaced the existing CSM-generated email reports with automated, AI-generated reports delivered directly to customers.
🖥️ Desktop
📕 Tablet
📱 Mobile
We need to ensure AI-generated reports match the quality and actionability of those crafted by Customer Success Managers (CSMs). Building trust in automation required us to focus not only on content accuracy, but also on insight relevance and clarity.
Feedback-driven loop: customer feedback collection
From the beginning, we aligned that AI-generated insights may not exactly match the quality or tone of CSM-authored reports. We focused on building a feedback-driven loop.
Customers can rate each section and each AI-generated insight, and that feedback is sent back to InsightsGPT to support continuous learning and improvement over time.
Feedback-driven loop: CSM feedback collection
Customer Success Managers can also evaluate and provide ratings for a report through the internal dashboard.
Next steps
Looking ahead, as the AI receives more feedback and better understands user needs, future iterations will shift the weight of the report toward AI-driven, actionable insights presented upfront.
Simplifying visuals for clarity
Remove overly granular information and charts
Simplified business summary
Simplified rating
Simplified promotions
New session: lost sales
The vision ahead
Looking ahead, as the AI receives more feedback and better understands user needs, future iterations will shift the weight of the report toward AI-driven, actionable insights presented upfront.
Other Projects
More Projects
See Project
→
See Project
→
Improving Inventory Accuracy With a Scan-First, Location-Based Workflow
See Project
→
See Project
→
Improve Market Analysis Tool for AE Adoption
See Project
→
See Project
→
Create AI-Driven Menu Optimization Experience
See Project
→
See Project
→
Redesign CloudKitchens Customer Onboarding Experience
See Project
→
See Project
→
Work
Work
About
About
Automating business reports with AI-generated insights
I led the design of an AI-generated reporting experience that transforms raw data into structured, actionable insights — reducing analysis time and enabling faster decision-making across devices.
Lead Product Designer
Email Report
2024
Problem
Customer Success Managers (CSMs) at CloudKitchens use InsightGPT to create customer-facing reports with strong results—80% open rates and an 8% GMV uplift. But the manual process doesn’t scale, limiting access to a wider audience.
On the customer side, report format is also text-heavy and hard to scan, insights were buried in raw data, actions are not highlighted enough, creating a poor reading experience we needed to improve.
Manual process of creating business report
Step 1
Step 2
Step 3
Step 4
Step 5
Solution
Design a fully automated system that delivers bi-weekly GPT-powered reports directly to licensees-unlocking exponential value for both internal teams and end-customers.
Discovery
We ran several working sessions with the CSM team to understand current workflows and define the core requirements for automation. Together, we reframed the format to be Critical Data→ AI Comparison/Summary →Actionable insights.
Also, we defined the Minimum Viable Product (MVP) for the automated report, which includes:
- Business Summary
- Top Items
- Customer Ratings & Reviews
- Active Promotions
Design solution
MVP data module
Redefined data modules: Business Summary, Top Items, Customer Ratings & Reviews, Active Promotions
Reframed module structure: Critical Data→ AI Comparison/Summary →Actionable insights.
Email report
As a first milestone, we replaced the existing CSM-generated email reports with automated, AI-generated reports delivered directly to customers.
🖥️ Desktop
📕 Tablet
📱 Mobile
We need to ensure AI-generated reports match the quality and actionability of those crafted by Customer Success Managers (CSMs). Building trust in automation required us to focus not only on content accuracy, but also on insight relevance and clarity.
Feedback-driven loop: customer feedback collection
From the beginning, we aligned that AI-generated insights may not exactly match the quality or tone of CSM-authored reports. We focused on building a feedback-driven loop.
Customers can rate each section and each AI-generated insight, and that feedback is sent back to InsightsGPT to support continuous learning and improvement over time.
Feedback-driven loop: CSM feedback collection
Customer Success Managers can also evaluate and provide ratings for a report through the internal dashboard.
Next steps
Looking ahead, as the AI receives more feedback and better understands user needs, future iterations will shift the weight of the report toward AI-driven, actionable insights presented upfront.
Simplifying visuals for clarity
Remove overly granular information and charts
Simplified business summary
Simplified rating
Simplified promotions
New session: lost sales
The vision ahead
Looking ahead, as the AI receives more feedback and better understands user needs, future iterations will shift the weight of the report toward AI-driven, actionable insights presented upfront.
Other Projects
More Projects
See Project
→
See Project
→
Improving Inventory Accuracy With a Scan-First, Location-Based Workflow
See Project
→
See Project
→
Improve Market Analysis Tool for AE Adoption
See Project
→
See Project
→
Create AI-Driven Menu Optimization Experience
See Project
→
See Project
→
Redesign CloudKitchens Customer Onboarding Experience
See Project
→
See Project
→
Work
Work
About
About
Automating business reports with AI-generated insights
I led the design of an AI-generated reporting experience that transforms raw data into structured, actionable insights — reducing analysis time and enabling faster decision-making across devices.
Lead Product Designer
Email Report
2024
Problem
Customer Success Managers (CSMs) at CloudKitchens use InsightGPT to create customer-facing reports with strong results—80% open rates and an 8% GMV uplift. But the manual process doesn’t scale, limiting access to a wider audience.
On the customer side, report format is also text-heavy and hard to scan, insights were buried in raw data, actions are not highlighted enough, creating a poor reading experience we needed to improve.
Manual process of creating business report
Step 1
Step 2
Step 3
Step 4
Step 5
Solution
Design a fully automated system that delivers bi-weekly GPT-powered reports directly to licensees-unlocking exponential value for both internal teams and end-customers.
Discovery
We ran several working sessions with the CSM team to understand current workflows and define the core requirements for automation. Together, we reframed the format to be Critical Data→ AI Comparison/Summary →Actionable insights.
Also, we defined the Minimum Viable Product (MVP) for the automated report, which includes:
- Business Summary
- Top Items
- Customer Ratings & Reviews
- Active Promotions
Design solution
MVP data module
Redefined data modules: Business Summary, Top Items, Customer Ratings & Reviews, Active Promotions
Reframed module structure: Critical Data→ AI Comparison/Summary →Actionable insights.
Email report
As a first milestone, we replaced the existing CSM-generated email reports with automated, AI-generated reports delivered directly to customers.
🖥️ Desktop
📕 Tablet
📱 Mobile
We need to ensure AI-generated reports match the quality and actionability of those crafted by Customer Success Managers (CSMs). Building trust in automation required us to focus not only on content accuracy, but also on insight relevance and clarity.
Feedback-driven loop: customer feedback collection
From the beginning, we aligned that AI-generated insights may not exactly match the quality or tone of CSM-authored reports. We focused on building a feedback-driven loop.
Customers can rate each section and each AI-generated insight, and that feedback is sent back to InsightsGPT to support continuous learning and improvement over time.
Feedback-driven loop: CSM feedback collection
Customer Success Managers can also evaluate and provide ratings for a report through the internal dashboard.
Next steps
Looking ahead, as the AI receives more feedback and better understands user needs, future iterations will shift the weight of the report toward AI-driven, actionable insights presented upfront.
Simplifying visuals for clarity
Remove overly granular information and charts
Simplified business summary
Simplified rating
Simplified promotions
New session: lost sales
The vision ahead
Looking ahead, as the AI receives more feedback and better understands user needs, future iterations will shift the weight of the report toward AI-driven, actionable insights presented upfront.
Other Projects
More Projects
See Project
→
See Project
→
Improving Inventory Accuracy With a Scan-First, Location-Based Workflow
See Project
→
See Project
→
Improve Market Analysis Tool for AE Adoption
See Project
→
See Project
→
Create AI-Driven Menu Optimization Experience
See Project
→
See Project
→
Redesign CloudKitchens Customer Onboarding Experience
See Project
→
See Project
→
Serenity: From Mental Health Chatbot to Clinical Platform
See Project
→
See Project
→
Modular TV Feed Redesign for Facility Operations
See Project
→
See Project
→
More Projects
See Project
→
See Project
→
