Addressing Corporate Travel Leakage with Machine Learning

Just prior to starting Signal Scout, Jeremy worked with Complai aka. Shep Travel, an enterprise travel technology company that use data pipeline automation and also trained machine learning models from within a browser extension.

Their product, a duty-of-care platform with a client side browser extension, monitors employee travel intentions as they browse and interact with travel websites – Expedia, Air Canada, Hilton for example. The aim is to shift travel bookings away from non-partners websites onto websites with whom Shep’s client has negotiated rates. By implementing enterprise scale data capture within the Shep browser extension using Snowplow we were able to gather enough of the right website interaction data to train machine learning models to actually recognize travel intent across nearly all travel websites. This was a strategic slam dunk for Shep as it moved them instantly from coding for individual websites to leveraging Snowplow data, AWS’s Sagemaker API gatway and lamba to have trained models decide when where and how an employee was intending to travel and of course what website they intended to purchase from. This major shift lead to a series of ground breaking machine learning integrated features that attracted and excited new clients which lead to Shep getting acquired by Flight Center Australia in 2022.

One feature empowered by this shift was the integration of contextually relevant real time duty of care information across thousands of travel websites without Shep’s developers ever needing to program for them.

Read more about how Shep used machine learning to innovate through the pandemic into acquisition.

-Jeremy Brooks