AI Data Assistant
AI powered data cleanup copilot that scans your event data, flags ambiguities, recommends fixes, and automates cleanup so teams can trust their metrics and focus on insights.
Year :
2024
Company :
Amplitude
Role :
Product Design Lead
Project Duration :
8 weeks to Beta launch
Problem :
Customers struggled to trust their event data, which meant they hesitated to trust Amplitude’s charts. Our platform included a a data management area, but it lived inside dense tables with unclear names, missing descriptions, and no obvious starting point. Most teams avoided it, so data quality decayed over time and undercut the value of the entire product.
Connect to Content
Add layers or components to swipe between.
SOLUTION :
I led design for Data Assistant, a focused space that turns telemetry into a single data health score plus a short, prioritized list of fixes. Under the hood, we looked at things like usage, volume, naming, and descriptions; in the UI, users saw simple batches such as “add descriptions to your top events” with one click approvals. Targeted nudges pulled them into the flow at the right moments so data cleanup became a lightweight habit instead of an overwhelming cleanup task.
IMPACT :
Teams using Data Assistant showed clear double digit lifts in weekly data cleanup actions (aka governance) and meaningful improvements in data health scores compared to orgs that did not adopt it. Customers reported more confidence in their charts, and the pattern of “review and approve suggestions” became a reusable foundation for other AI driven clean up flows across the product.
+41%
Weekly Governing Orgs
+16%
Data Health Score
More Projects
Let's connect!
AI Data Assistant
AI powered data cleanup copilot that scans your event data, flags ambiguities, recommends fixes, and automates cleanup so teams can trust their metrics and focus on insights.
Year :
2024
Company :
Amplitude
Role :
Product Design Lead
Project Duration :
8 weeks to Beta launch
Problem :
Customers struggled to trust their event data, which meant they hesitated to trust Amplitude’s charts. Our platform included a a data management area, but it lived inside dense tables with unclear names, missing descriptions, and no obvious starting point. Most teams avoided it, so data quality decayed over time and undercut the value of the entire product.
Connect to Content
Add layers or components to swipe between.
SOLUTION :
I led design for Data Assistant, a focused space that turns telemetry into a single data health score plus a short, prioritized list of fixes. Under the hood, we looked at things like usage, volume, naming, and descriptions; in the UI, users saw simple batches such as “add descriptions to your top events” with one click approvals. Targeted nudges pulled them into the flow at the right moments so data cleanup became a lightweight habit instead of an overwhelming cleanup task.
IMPACT :
Teams using Data Assistant showed clear double digit lifts in weekly data cleanup actions (aka governance) and meaningful improvements in data health scores compared to orgs that did not adopt it. Customers reported more confidence in their charts, and the pattern of “review and approve suggestions” became a reusable foundation for other AI driven clean up flows across the product.
+41%
Weekly Governing Orgs
+16%
Data Health Score
More Projects
Let's connect!
AI Data Assistant
AI powered data cleanup copilot that scans your event data, flags ambiguities, recommends fixes, and automates cleanup so teams can trust their metrics and focus on insights.
Year :
2024
Company :
Amplitude
Role :
Product Design Lead
Project Duration :
8 weeks to Beta launch
Problem :
Customers struggled to trust their event data, which meant they hesitated to trust Amplitude’s charts. Our platform included a a data management area, but it lived inside dense tables with unclear names, missing descriptions, and no obvious starting point. Most teams avoided it, so data quality decayed over time and undercut the value of the entire product.
Connect to Content
Add layers or components to swipe between.
SOLUTION :
I led design for Data Assistant, a focused space that turns telemetry into a single data health score plus a short, prioritized list of fixes. Under the hood, we looked at things like usage, volume, naming, and descriptions; in the UI, users saw simple batches such as “add descriptions to your top events” with one click approvals. Targeted nudges pulled them into the flow at the right moments so data cleanup became a lightweight habit instead of an overwhelming cleanup task.
IMPACT :
Teams using Data Assistant showed clear double digit lifts in weekly data cleanup actions (aka governance) and meaningful improvements in data health scores compared to orgs that did not adopt it. Customers reported more confidence in their charts, and the pattern of “review and approve suggestions” became a reusable foundation for other AI driven clean up flows across the product.
+41%
Weekly Governing Orgs
+16%
Data Health Score
More Projects
Let's connect!