Daniel Fitzpatrick
← All work
Parade

GenAI that answers carrier inquiries

Freight brokers spend hours a day answering carrier inquiries that rarely convert. We believed AI could handle them faster and more accurately than a human, and turn dead-end emails into quotes.

Role
Product Designer
When
2020–22
GenAIConversational UXRapid Prototyping0→1
CoDriver responding to a carrier inquiry on a broker's behalf

A strong hypothesis

We hypothesized that CoDriver, Parade’s proprietary AI, could handle inbound carrier inquiries quicker and more accurately than human brokers, freeing their time for strategic work, and converting otherwise-dead inquiries into valuable new quotes on the platform.

Fast iteration and testing

I was asked to design something we could demo to prospects and learn from as fast as possible. Working directly with the CEO and a small team of engineers and data scientists, we iterated on an email plugin and eventually a Parade UI to show how it would all work.

The premise: a load inquiry arrives in an inbox we control. CoDriver parses the request, supplies formatted load information, feeds it into our load-matching model, and generates a reply that shares load details and asks for carrier information and a rate. Most of the work was conversational UX: testing scenarios and ensuring the AI responded the way an experienced freight broker would.

CoDriver filters and conversation status
Status monitoring: filter to the ~10% of conversations needing a human

Building trust

We built a page to monitor the status of these requests, making testing easier and giving brokers confidence in what the AI was doing on their behalf. Users could filter straight to conversations that required intervention (about 10% of the total).

The results

CoDriver shown as one of several quote sources
CoDriver became one more source of quotes feeding the platform

In less than two months the idea was in production and being tested by customers. Early users saw 4× more quotes and 3× more digital transactions than before.