Shipwell is on a mission to transform the shipping industry through technology—turning multi-leg journeys involving several parties and complex logistics into structured, predictable workflows that delight shippers around the world.
Success depends upon delivering visibility at scale. Doing so requires high-quality data—which is crucial to understanding where items are. But many traditional operators are still tracking trucks using whiteboards, paper and pencil, or even manual phone calls.
In the long term, IoT will be part of any solution. Sensors placed on each truck can report its location, but implementing a solution like this broadly across the shipping industry is a multi-year process. Using conversational AI—automatically contacting drivers on the road to ask them for location information—is a first step toward filling in the gaps. With better data in hand, Shipwell can see where shipments are at risk of running late, and deliver the information and workflows that customers need to respond accordingly.
When Jacob Gordon joined Shipwell’s data science team, Shipwell’s CEO Greg Price quickly set up a meeting with him and the Georgian team. An engineer by training, Price appreciated the potential of conversational AI for Shipwell’s business.
To set the right strategy for conversational AI, the team drew on an in-depth knowledge of shipping industry dynamics and a thorough understanding of their users. Georgian and Shipwell identified the key personas—truckers, dispatchers, shippers, and 3PL management—and the most significant conversational AI opportunities for each. After scoring and prioritizing these use cases by business impact and cost/complexity, the team quickly honed in on Shipwell’s most critical business task: making sure goods get to their destinations on time.
The team, which was composed of Max McLennan, Anna Kómár, Steve Chronister, Jeff Wilson, and Jacob, began a deep-dive on the problem. Voice emerged as a logical solution, as drivers aren’t always able to quickly answer a text or enter data into an app while driving. But a crucial decision was deciding what to ask them. To keep things simple, instead of asking, “How’s your trip going?” or “Are you running on time?” the team settled on a single, straightforward question: “What’s the nearest city?”
By comparing driver responses to geographic databases, Shipwell captured highly accurate responses without having to develop a complex library of intents and responses. This built trust with drivers—a crucial step in setting the stage for future, richer interactions.
What’s next for Shipwell? While an obvious solution might be to increase the number of calls and conversational touchpoints, the team plans to focus on enriching the existing touchpoint to provide additional value to the driver. For instance, conversational AI could inform drivers of upcoming places to stop or make suggestions on how best to time breaks to minimize traffic time.
What advice would Jacob give to other companies looking to make their first foray into conversational AI?
“Even if you think your first project is obvious, take a step back and assess all your potential opportunities. And pay close attention to the business value, not just how to execute a specific feature.” Jacob says it’s also important to start small: “NLP is an advanced technology, but one of the highest impact things you can do is constrain the responses to keep things simpler. Ask for something that’s not unnecessarily subjective, and make sure the results are readily automatable downstream”.
To see how this type of information provides valuable insights for shippers, take a look at Shipwell’s interactive dwell time map. The map allows shippers to benchmark facility buffer times, estimate delivery windows and compare wait times across facilities on their favorite lanes.
Are you ready to dive into conversational AI? Check out Georgian’s conversational AI resources or get in touch with me.