Designing a Last Mile Delivery Chatbot
Task
Reducing the burden on last mile delivery customer service by introducing an intuitive chatbot to assist customers with delivery queries.
Last mile deliveries can be painful at times. Many a times, users are left in the dark on the status of the deliveries which results in angry calls to customer support. To reduce those calls on customer support, a popular last mile delivery company approached the team to assist on coming up with a solution for customers to easily keep track on their packages.
Research & Collaboration to Building the Chatbot Flow
The project involved working with the development team also since the idea we wanted to propose required multiple inputs. We had an internal discussion on proposing the use of a Chatbot to the client. The chatbot would be built using Salesforce Einstein – the first of it’s kind AI for CRM.
The client loved the idea and asked us to explore the concept and come up with possible approaches for the chatbot. While the technical team worked on the feasibility and implementation side of things, the team and I were looped in to look into strategising the functionality and content direction of the chatbot.
I dug into some research on building an effective chatbot and also a little into last mile deliveries and brainstormed the multiple scenarios that could occur with the design team. We were in constant touch with the business to take in their inputs based on their experiences with customers and delivery partners.
Some of the scenarios we came up with were:
- Non availability of a tracking number to track a package
- Delivery issues with the package – in case of damaged parcels, delivery disputes or other problems
- Situations where customers refused delivery of packages
- In case if customers needed to change address or add some additional instructions to the delivery.
With some inputs from business and based on a few assumptions, the team and I worked with the development team to work on the chatbot flow. The flow went through iterations in multiple conversations with the client until the flow was approved.
Implementing, Testing and Demo-ing the Chatbot
With the chatbot flow approved, the development team set up the chatbot. With the implementation in progress, I helped with correcting and rewriting the language wherever it seemed necessary. We tested and demoed each of the scenarios of the chatbot flow for the client and it went the way it was expected it to go.
The final chatbot has been deployed and delivered to the client. The team and I are currently waiting in on learning how much this solution has worked for the client.
