By Matt Bouton, BI Manager, Amazon Shipping and Amy Marvin, Sr Technical Product Manager, Amazon Quick
Amazon Shipping’s Quick Chat Use Case
At Amazon Shipping, our BI team equips the portfolio owners and sales leaders responsible for a vast commercial shipper network with the data they need to manage growth, volume commitments, and revenue attainment.
To serve that audience, we built comprehensive Amazon Quick dashboards with multiple tabs, curated KPIs, and layered filter controls. Many views default to a rolling 25-week window so portfolio owners can monitor recent performance across their managed cohort.
That structure works well for standard reporting. But as our scale and use cases grew, so did the complexity of the questions.
Our teams don’t just ask:
How many packages did we induct today vs. yesterday?
They also ask:
-
How many new shippers onboarded YTD vs through the same time last year? -
How do package volumes trend for Mid-Market vs. Enterprise over the past two quarters? -
What percentage of shippers missed their volume commitments during peak holiday season? -
Have any of my shippers encountered pickup or delivery issues lately?
These questions often require switching owners, expanding time ranges, comparing cohorts, or applying performance thresholds. Previously, that meant manually adjusting multiple filters or navigating across sheets.
For a global sales organization, that friction adds up. Our stakeholders needed a faster way to express intent — that’s where chat changed the experience.
Moving from Dashboard Configuration to Intent Expression
We’ve been early adopters of Amazon Quick chat, and over time we’ve seen it handle increasingly nuanced analytical intent. For example, on our “Recent Launches” dashboard, the current filter state might reflect Owner: Ariel with a 25-week trailing window applied. If another portfolio owner—say, Anh—wanted to analyze her own portfolio across all historical data, she would previously need to manually adjust multiple filters (Image 1).
Image 1: Amazon Shipping dashboard open to the “Recent Launches” sheet filtered to Owner: Ariel, Cohort: Mid Market and ENT, and Trailing Weeks Since Launch: 25.
Now she can simply ask, I want to understand my portfolio of shippers (Owner: Anh) for all time (Image 2).
Image 2: Chat response to user asking about their shipper portfolio.
Behind the scenes, chat overrides the current owner filter and expands the temporal constraint beyond the 25-week window to “all time.”
From there, Anh can immediately go deeper:
-
When did performance begin to decline for Shipper X? -
Shippers are meeting volume goals but missing revenue attainment. What’s driving? -
How can I drive attainment to at least 70%?
Instead of stopping at a metric, she moves directly to action—identifying at-risk accounts, prioritizing outreach, and deciding how to manage her portfolio. That’s the difference: chat shortens the path from question to decision.
How This Changes Our BI Strategy
The impact isn’t just on end users—it’s changed how we think about dashboard design. Previously, we invested heavily in anticipating every possible analytical path. That often meant creating additional sheets, building specialized visuals for edge cases, and pre-building common comparison views (for example YTD vs. prior year or peak season analysis). As usage scaled, that approach became harder to sustain.
Now, instead of increasing dashboard surface area, we focus on strengthening the semantic foundation that powers both dashboards and chat.
Concretely, that means:
-
Standardizing chart titles and filter control names to reduce ambiguity
-
Ensuring metric definitions (via calculated fields) consistently encode the correct business rules
-
Adding agent custom instructions to guide how analytical questions should be handled
When we improve a metric definition or clarify a dimension, both dashboards and chat benefit automatically. That leverage is powerful. And because Quick Suite scales seamlessly in the cloud, our dashboards and chat workflows scale with our organization—without additional infrastructure overhead. As new feature improvements roll out, our users benefit immediately.
In practice, this means less time maintaining dashboard permutations, less time responding to one-off analysis requests, and more time strengthening data definitions and analytical clarity. We’ve shifted from pre-building every possible view to strengthening the semantic layer that defines how our data is understood.
How Other Amazon Teams Are Using Chat
We’re not alone in seeing this impact.
An Amazon Loss Prevention Specialist shared:
My Assistant saved me 20–30 minutes of investigation prep when identifying the top five high-value concessions drivers for Fulfillment Center X. Just two simple prompts delivered second- and third-order investigative data instantly, providing actionable insights exactly when I needed them.
An Amazon Station Manager noted:
Chat often prompts me to explore angles I hadn’t initially considered. I find the follow-up question suggestions particularly helpful for conducting more thorough investigations.
For cross-country workforce analysis, an Amazon Principal HR Program Manager explained:
Dynamic filters were immediately useful due to the multiple cross-country analyses I conduct. My prompt ‘Can you tell me for 2025 at a monthly level the values for Aus, Mex, Ind and Japan,’ now works without changing filters.
And an Amazon Operations Manager added:
This feature is huge. I am constantly switching between stations, weeks, and process paths.
Across these teams, the pattern is consistent: Chat reduces configuration overhead, interprets analytical intent, and accelerates decision-making—whether the question spans fulfillment centers, countries, time horizons, or performance tiers.
Getting Started
If you want to explore what Quick Chat can do for your team and stakeholders, get started at https://aws.amazon.com/quick/. Join the Community to find answers to your questions, learning resources, and events in your area https://community.amazonquicksight.com