Relay Assistant - Negotiations
In 2024, I designed the first AI powered negotiations platform in Amazon that supports middle mile carriers. This chatbot helps carriers negotiate rates for work available in their domicile, and book the negotiated rate in a few clicks.
(To maintain confidentiality, I have included visuals that did not launch, but give an idea about my skills and my role as a designer.)
Project Overview
Customers
Our customers are carriers- trucking companies and owner-operators approved to haul freight for Amazon through the Relay for Carrier platform, which connects carriers directly with Amazon’s load board and contracts without third-party intermediaries
Opportunity
In 2023 the product team identified an opportunity: a manual rate-negotiation system that relied on employees calling carriers to coordinate pickups. This process was inefficient, biased toward established carrier relationships, and impossible to scale.
Goals
The goal was to enable scaling (by not being limited by availability of employees available to call), lowering costs by offering work to many carriers simultaneously and increasing rate of acceptance.
Key Screens
Key Areas of Design Influence
Inputs to Design System
Products in middle mile work with a design system. I helped design the chat pattern for this design system through a study of chat patterns, design explorations and testing. While chat patterns have been around for a while, designing details that are specific to our customer was imperative.
Conversation Design through continuous monitoring
Through this project, I helped set the standards and form guidelines for our studio for conversation design. The team first launched a pilot with about 10 carriers to get feedback on conversation design. The biggest piece of feedback we received was around the length of the messages that Relay Assistant was delivering.
“It’s good only if the job is done in few messages. When the bot keeps repeating messages, that’s when it gets annoying”
That’s when I started developing tenets and frameworks for the team to use. I dug into some of the prompts and realized that tweaking those prompts, following the CARE framework and adding specific examples gave results.
Scaling beyond Negotiations
I Collaborated with product and engineering leaders across the organization to test how this chatbot could help automate other tasks for carriers. Through workshops with partners, roundtable discussions with carriers, mapping user flows, I was able to propose ideas and other areas of the product where we could automate tasks using Relay Assistant framework.
How did we measure success?
Reduction in manual intervention
“Previously we had to have loads adjustments made and assigned to us using phone & chime communications with ROC agents. If the ROC agent and carrier assistant are unable to connect in a timely manner, we did not get the load. Relay assistant is available 24x7, so it can alert us about critical spot loads round the clock, and we can book work quickly by ourselves””
Achieved 5x the conversion rate of the original goal.
(Exact numbers cannot be revealed due to confidentiality)
Improved cost of conversion
Achieved millions of dollars of annualized reduction
(Exact numbers cannot be revealed due to confidentiality)
Established a reusable pattern
The same pattern scaled to 1 other use case and is in pilot for 2 other use cases.