Carriyo Intelligence Opens in Closed Beta: Revenue at Risk and CX Score
In September we wrote about the two reports at the centre of Carriyo Intelligence, Revenue at Risk and CX Score. This week Carriyo Intelligence opens as a closed beta for a selected group of users across some of our key accounts, so this seemed like a good moment to go through what those users will actually be looking at, and why we think it only works because of where the data comes from.
On Wednesday we wrote about what retailers should be measuring in returns beyond the return rate, and a lot of that piece applies here as well, because returns show up in both of the main reports.
What Carriyo Intelligence is
Carriyo Intelligence is the reporting layer of The Intelligent Commerce Platform. It works from the order, shipment, return, notification and customer feedback data that Carriyo already records as part of running the operation, so there is nothing new to collect and no separate data project before the reports mean something.
The two reports we lead with are Revenue at Risk and CX Score. Around them sits a library of real-time and historical reports, including an Executive Overview of orders, shipments, returns and fulfillment on one screen, Delivery Date Intelligence, which scores the customer promise, the AI estimated delivery date and the carrier's ETA against the actual delivery, and Order Lifecycle, which shows where the time goes between the order and the delivery and whether a delay sits in your own operation or with the carrier. Most of the reports open with a short AI-written summary of the numbers.
Revenue at Risk
Revenue at Risk is a daily number for the money tied up in deliveries that went wrong. At the close of each day it values the orders that are exposed and adds them up, in five groups:
- Delayed shipments, which are orders past their delivery deadline with the parcel still on its way.
- Failed deliveries, where a delivery attempt failed that day and the parcel has not been resolved yet.
- Stuck fulfillment, which covers orders with a fulfillment order that has been waiting for more than 24 hours.
- Pending returns, meaning open return requests, valued at the amount approved for refund.
- Churn risk, which is an estimate based on the recent spend of customers who left negative feedback in the last 30 days. It is clearly labelled as an estimate rather than money in play.
The report shows the exposure that is still open, a daily trend, and how that exposure compares with what you shipped over the previous 30 days. Every risk row opens the matching list of shipments, orders or returns in the Carriyo dashboard, so the person reading the number can go straight to the parcels behind it. There is also a table of the customers who left negative feedback, with their recent spend and their CX score. For accounts that use Carriyo for shipping only, the report still works, and it measures stuck fulfillment and pending returns from the shipments.
The point of putting it in money is that it is a number a finance lead and an operations lead can both look at and agree on, which is not always true of an on-time percentage.
CX Score
CX Score gives every delivered parcel a score from 0 to 100, based on what actually happened to it. It looks at delivery against the date you promised the customer (on time or early scores best, and each day late costs points), whether the parcel was delivered on the first attempt, whether the notifications reached the customer, how a linked return went, and the customer's own star rating where they left one. Each parcel is only scored on the signals that apply to it, so a parcel with no return and no rating is not penalised for either.
From there you get an average over the last 30 days, a daily trend and the spread across score bands. There is a score for each of your carriers next to its on-time rate and customer rating, a list of the lowest-scoring parcels with each one a click away from the shipment, and a check of the score against what customers actually said in their ratings, so you can see whether the score and the customers agree. The carrier comparisons are based on your own shipments.
Why it sits on Carriyo and not in a BI tool
We get asked fairly often why a merchant would not simply build this themselves, with their own warehouse and a BI tool, and especially now that AI makes building software much quicker. It is a fair question, and our honest answer is that building the reports themselves is usually not where the difficulty is.
The hard part is the data underneath. On Carriyo, orders, shipments, returns, notifications and customer feedback are recorded in one consistent model, and every carrier's tracking updates are translated into the same set of statuses, across a network of 140+ carriers. That means "late" is defined once, against the promise you actually made, and a failed delivery means the same thing whichever carrier it happened with. A reporting layer built in-house usually has to solve that normalization first, carrier by carrier, and it starts with no history.
That normalization reflects know-how built up across 100+ brands, $5B in GMV shipped and deliveries to 200 countries, and it has taken us a number of years to get right. Each customer's reports are built only on its own data, but the data model they sit on has been shaped by all of that experience, and we think that is what makes the numbers trustworthy enough to act on, whether the one acting is a person or an agent.
For retail groups, the Intelligence reports carry a merchant filter, so a group can read one brand on its own or all of the brands in its Carriyo account together.
Three ways to use it
Like everything on Carriyo, Intelligence works in three modes. Your team reads the reports in the dashboard and opens the shipments, orders or returns behind a number when they need to act on it. Automation rules and webhooks work from the same shipment data as it changes, so your own systems can hear about an event without anyone watching a screen. And AI agents connect through Carriyo's native MCP layer and work from the same normalized data your team sees. We expect most customers to use a mix of the three.
See it on your own data
If you would like to see what Revenue at Risk and CX Score would look like on your own data, you can read more on the Carriyo Intelligence page or talk to us, and we will walk you through the reports and scope what they would look like for your operation.
Carriyo is The Intelligent Commerce Platform, from checkout to doorstep.
---
Sources
1. Carriyo. "Carriyo Intelligence: Revenue at Risk and CX Score Are Built, and Open to Customers in Q4 2026" (Carriyo blog, September 2026). https://carriyo.com/blog/carriyo-intelligence-revenue-risk-cx-score-2026-09-17/ 2. Carriyo. "What Retailers Should Measure in Returns Before Peak, Beyond the Return Rate" (Carriyo blog, October 2026). https://carriyo.com/blog/retailers-measure-returns-peak-beyond-return-2026-10-13/ 3. Carriyo. Carriyo Intelligence platform page. https://carriyo.com/platform/intelligence/ 4. Carriyo. Carrier network. https://carriyo.com/carriers 5. Carriyo. MCP Server documentation. https://carriyo.com/docs/mcp/ 6. Carriyo. "How Carriyo's MCP Server Turns AI Agents Into Shipping Operators" (Carriyo blog, May 2026). https://carriyo.com/blog/carriyos-mcp-server-turns-ai-2026-05-23/