Illustrative scenario. This is not a customer story: the company, its data and its numbers are invented to show how the product is used.
The situation
A regional courier runs six depots and delivers for about forty business customers. Every Monday the operations manager wants one number, on-time delivery by carrier for last week, and every month each large customer asks for its own delivery performance.
The data already exists. Shipments, delivery attempts and status codes live in the courier's delivery-management database, a SQL Server inside the head office. Getting a report out of it means asking the one person who writes SQL, and waiting.
Step 1: connect the data
The team installs the gateway on a Windows machine in the head office and adds the SQL Server as a datasource. No inbound firewall port is opened, and the database password stays inside the network.
A team whose data only arrives as spreadsheet exports would upload the Excel file instead.
Step 2: ask the first question
In the workspace, the manager types:
What was on-time delivery by carrier last week?
The answer comes back as a bar chart, with the query behind it available to check. The follow-ups are just as short:
Which depots had the most failed first attempts this month, and what were the reasons?
Show average cost per shipment by zone for the last 12 weeks as a line chart.
Step 3: turn the answers into a report
Each useful chart goes onto a report. Auto Report drafts the rest: a KPI row, a trend by week and a table of late shipments. A report filter for customer and one for week let the same report answer "how did we do for this customer?" without a second copy.
Step 4: send it every morning
A subscription emails the report to depot managers at 07:00 each day and to leadership every Monday. Nobody has to remember to run it.
Step 5: give each customer their own view
For the large accounts, the courier embeds the report in its customer portal, filtered to that customer, so the monthly "how did you do for us?" request no longer lands on anyone's desk.
What to check before you trust the numbers
- One name per carrier. "DHL", "D.H.L." and "DHL Express" count as three carriers until they are cleaned up. A business definition can say which values mean the same thing.
- What "on time" means. Agree whether it is measured against the promised date or the promised time window, and write it down as a definition so every answer uses the same rule.
- Time zones. Delivery timestamps recorded in UTC can move a late-evening delivery into the next day.
Next steps
See AI reporting for courier and logistics teams for the reports teams usually build first, or build your first report free.