Can Your 3PL Do This? Connecting Customer Service AI Directly to Fulfilment Data
Most businesses looking at AI for customer service quickly run into the same problem. The AI might be perfectly capable of having a sensible conversation with a customer, but unless it has access to the systems containing the answer, it can only do so much.
Delivery queries are a good example. A customer asks, “Where is my order?”, and traditionally somebody in customer service has to identify the order, open a warehouse or courier system, check what has happened and then translate that information back into a useful response. It is a straightforward task, but when you multiply it across hundreds or thousands of orders it consumes a surprising amount of customer-service resource.
We thought there should be a better way of doing it, particularly as more of our clients start experimenting with AI-based customer service.
So we have built one.
Giving AI Access to the Answer, Not Just the Conversation
Move Fresh has developed a service that allows an AI customer-service application to query fulfilment and delivery information directly from our systems. Instead of the AI telling a customer that somebody will check their order, it can check the order itself and provide the answer during the conversation.
The connection uses Model Context Protocol, usually shortened to MCP. There is plenty of technical detail behind MCP, but from a brand’s perspective the useful part is relatively simple: it provides a standard way for AI applications to connect securely to external systems and use the information held within them.
In our case, that means a client’s AI can ask Move Fresh for the current position of an order and receive the information it needs to answer the customer. The service can report where an order is within the fulfilment process, provide expected delivery information and, once the parcel is moving, use available courier tracking information to give a more current estimate.
That changes the role of AI considerably. Rather than being a chatbot sitting in front of a knowledge base and answering generic questions, it becomes capable of resolving a genuine customer enquiry using live operational data.
“Where Is My Order?” Is More Complicated Than It Looks
Anyone who has worked in ecommerce customer service will know that a delivery query is rarely as simple as reading out a tracking status.
An order that has been travelling for three days might be perfectly on schedule if it was sent on a five-day service, while a next-day order in exactly the same position clearly needs attention. An order might still be going through the warehouse, it might have been packed but not collected, or the courier may already have provided a more precise delivery window.
For an AI service to be useful, it needs enough context to understand those distinctions rather than simply finding a status and repeating it.
That was an important part of how we approached the Move Fresh service. Once an order has been dispatched, the system retains the delivery deadline that applied when it left us and can compare this with the latest expected delivery information. This allows an AI assistant to distinguish between an order that is genuinely late and one that is simply taking the expected amount of time for the service selected.
It sounds like a small detail, but it matters. A customer does not need an apology and escalation because their five-day delivery is still moving normally on day three. Equally, somebody waiting several days for a next-day shipment deserves more than a generic message saying that the parcel is “in transit”.
Good customer service depends on understanding the difference.
Freeing People to Deal With the Queries That Actually Need People
We do not see this as AI replacing customer-service teams. The more useful opportunity is removing a significant volume of routine administration from them.
Checking the status of an order is necessary work, but it is rarely where an experienced customer-service person adds the most value. Their time is better spent resolving genuinely difficult issues, helping vulnerable customers, dealing with damaged or missing orders, arranging replacements or handling the situations where judgement and empathy actually matter.
If an AI assistant can resolve straightforward delivery queries immediately, at any time of day, the customer gets a quicker answer and the service team gets more capacity to deal properly with the exceptions.
There is another benefit for D2C brands because customer enquiries do not necessarily arrive conveniently between nine and five. Someone wondering where their food order is at 8pm should not have to wait until the following morning simply because finding the answer requires a person to log into another system. If the information is available, there is no particularly good reason why the customer should not be able to access it.
Keeping AI Access Controlled
Giving an AI service access to operational data obviously raises questions about security, so we designed the connection around a deliberately narrow set of capabilities.
Access is read-only and restricted to the individual brand’s orders. The service can look up fulfilment and delivery information, but it cannot create an order, amend one, cancel it, place it on hold or issue a refund. Personally identifying customer information is not exposed through the service.
Each client controls access using its own secure authentication and can disable or replace its access credentials if required.
That distinction between allowing an AI to read information and allowing it to take operational actions is important. There will undoubtedly be more opportunities for AI agents to carry out actions in ecommerce over time, but for delivery enquiries we think the sensible starting point is giving them accurate information while keeping control of operational changes elsewhere.
Why This Matters for D2C Brands
There has been a lot written about what AI might eventually do for ecommerce, but we are more interested in finding practical things it can do now.
For a growing D2C food, drink or health brand, delivery enquiries are a good example because they sit at the intersection of customer experience and operational cost. As order volume grows, the number of routine enquiries tends to grow with it, which traditionally means adding customer-service resource or accepting slower response times during busy periods.
Connecting customer-service AI directly to fulfilment information gives brands another option. Routine enquiries can be resolved automatically, customers get access to current information more quickly, and the brand’s own team can concentrate on cases that genuinely require intervention.
It also demonstrates something broader about the relationship between a brand and its 3PL. A modern fulfilment partner should not operate as a closed warehouse where information has to be retrieved manually every time somebody needs it. The warehouse should be connected to the rest of the brand’s technology, whether that is its ecommerce platform, reporting environment, marketplace software or increasingly its AI tools.
We have invested in our own development capability at Move Fresh for exactly that reason. We want the fulfilment operation to become an extension of the systems our clients already use, rather than another piece of technology their teams have to work around.
The Next Step in Connected Fulfilment
AI customer service will develop quickly over the next few years and we certainly do not pretend to know exactly where it will end up. What seems much clearer is that the useful applications will depend on access to reliable, structured data.
An AI assistant without operational information can have a very convincing conversation while still being unable to answer the customer’s question. Connect it securely to the right information and it becomes considerably more useful.
Our new AI customer-service integration is a relatively simple example of that principle, but we think it points towards where fulfilment technology is heading. The opportunity is not to add AI terminology to an existing warehouse operation; it is to make fulfilment data genuinely accessible to the tools brands want to use.
For us, that starts with one of the most common questions in ecommerce: “Where is my order?”
If your customer-service team is still manually checking fulfilment and delivery systems every time that question arrives, there is now another way of doing it.
Want to See How It Works?
Move Fresh clients can connect compatible AI customer-service applications directly to our fulfilment data using our new MCP service, with secure, read-only access restricted to their own orders.
If you are a D2C food, drink or health brand shipping more than 100 orders a day and would like to understand how this could work alongside your existing customer-service operation, talk to us.
Find out more about Move Fresh AI Customer Service or speak to our team about connecting your customer-service AI directly to fulfilment data.
Can Your 3PL Do This? Connecting Customer Service AI Directly to Fulfilment Data Read More »


