Post-purchase used to be one thing. You sent a shipping confirmation, and if something went wrong, someone called you.
That is not what our retail and eCommerce customers run today. They run delay notifications and failed-attempt messages and pickup reminders, across six markets in four languages, for three brands, with different carriers in each region and a returns policy that changes when the season does. Every year there is more of it, and more of it that customers notice when it’s wrong.
The retail teams running it have not grown at the same rate. Same number of people, same hours, considerably more things to be on top of.
That gap is what we have been working on.
parcelLab Copilot already answered the why
We shipped parcelLab Copilot last year to answer questions. Why did returns spike in the UK. Why is this order stuck. Why does this number look wrong. It reads your live account and tells you what it found.
Customers use it constantly, because it is the fastest way to an answer. A CX lead at Caraway Home put it better than we could:
“It’s like your data scientist bestie… parcelLab just has so much data, especially carrier data, and I think people underestimate how difficult it can be to have access to that data”
But answering is half the job.
The other half was still yours to figure out
parcelLab Copilot could tell you exactly what was wrong. Then it handed the work back.
Think about what happened next. You worked out what should change. You opened the right tool. You found the setting, which meant knowing where it lived. You made the change, tested it, published it, and checked it had not broken something else.
Seven steps, and six of them happened nowhere near the conversation where you worked out what to do.
That is hours and sometimes a week of someone’s attention. And here is the part that bothered us the most: a two percent improvement cannot justify a week of anyone’s time, so it waits. Next quarter. After peak. When we have capacity.
It never happens. Not because it was not worth doing, but because the cost of doing it was higher than the value of doing it.
So we asked a different question. What if it took twenty minutes?
parcelLab Copilot now does the work
From today, Copilot’s new capabilities are available to every parcelLab user. Ask it something about your account, and when you are ready, ask it to do something. Describe the ask and it reads your setup, works out everything the change touches, and prepares it. You review exactly what it plans to do. You approve. It runs.
A CX manager at a US retail brand wanted to know how often deliveries arrived on the promised date. She knew the data was in parcelLab. She did not know how to get a straight answer out of it.
“So I went to Copilot and got a really clear answer. I prompted Copilot to create a visual for me, along with a table showing how often the actual delivery is on time, one, two or three days early, or one, two or three days late. So I was able to package up this really thorough response to send up to my boss.” Then she asked what to do about it.
“Copilot produced this really nice phased approach of four steps that we can take, in order of lowest risk and biggest impact, to try and reduce that promise date, to get it more accurate, closer to the actual delivery date that the customer would experience.”
Then she asked it to go ahead.
“I got the green light for phase one, and then Copilot actually took those steps for me. Copilot tested the change and provided me with the test results prior to asking for my approval to make the change. I said, go ahead, let’s make this change, and it’s live right now.”
Question, analysis, recommendation, approval, change. One conversation. That is the whole thing.
Five kinds of actions
parcelLab Copilot investigates, creates, configures, improves and troubleshoots. In practice that means a delay notification for orders with no carrier movement in 48 hours. A return window extended to 31 January for holiday orders. A line removed from your Delivered email across nine templates, three brands and six markets, translations included. Return labels that stopped generating in Austria, diagnosed and fixed.
Real requests, described the way a person would ask a colleague.
Why we made it ask permission
Much of the industry is racing to make AI more independent. We deliberately explored another route.
parcelLab Copilot never acts on its own. It shows you exactly what it plans to change and everything that change will affect, before anything happens. New messages and journeys arrive as drafts. It works only within the access you already have. Every action is logged with who approved it and when.
This was not caution bolted on at the end. Giving software the ability to change a production configuration is a serious thing to do, and the review step is what makes the capability usable rather than frightening. One customer described it as the thing that makes the whole model work:
What early access customers are saying
We opened early access in September. Since then teams have used Copilot to audit their returns setups, stand up messaging for new markets, fix broken label routing and get ready for peak holiday season.
Jennifer Villalba, Director of Customer Experience
32 Degrees
“I wanted it to look just like my other emails that I have set up right now… It was literally done in like two seconds. With Copilot, I created three emails, which took me literally about 10 minutes to do… I feel like that’s something that would take a marketing team probably like a week. It just empowers me more.”
Nancy Gurd, Director CX
Caraway Home
“One of the use cases that I used Copilot for was to look at all the different couriers that we currently have in motion and understand which ones are performing well, which ones might be getting more exception rates. Normally, that would take me about two to three hours just to get the data sets and manipulate it together. But with Copilot, I was able to just prompt it and put it all in one easy place to read and understand and have almost a conversation with it.”
Ashley H., Senior CX Manager
Caraway Home
“I’m already using Copilot to create new flows. I’m using it to help prepare for Q4… Being able to identify what right now is at a more reasonable volume and what those smaller gaps are, I can fill them before we see them really peak at scale come Q4… Instead of waiting for the fire to happen, I can already create messaging. And that’s what I’m really excited for.”
Taylor Johnson, Director of CX
Nathan James
“Being able to utilise Copilot to essentially set that kind of thing up in a day — what would normally have taken a week or more — is game-changing… I can probably get it done so much quicker, which is great because it frees up my time.”
Caela Castillo, Director of CX
JAXXON
“It can edit things for you. It’s creating new things. But it gives you the power to say yes or no… It’s putting things into draft mode… We have so many ideas, but here, if you just take the time to go through those steps in Copilot, it already creates something for you that I can decide if I want to do now or if I need to get anyone else’s approval.”
We don’t think the finished post-purchase experience exists. You do not complete it. You keep improving it, market by market, message by message, as the business changes underneath you. So, ask Copilot and let it get to work.
AI
