Using AI to improve relocation outcomes

Using AI to improve relocation outcomes
Few employers truly understand the power and limitations of AI at the onset, but as long as there’s a solid business objective in mind, they can work towards it through a process of discovery. Credit: metamorworks/Shutterstock

Whether you call it predictive analytics, machine learning, or artificial intelligence (AI), these technologies are all about gaining actionable insights to improve outcomes — no matter what business you’re in.

Few employers truly understand the power and limitations of AI at the onset, but as long as there’s a solid business objective in mind, they can work towards it through a process of discovery. 

AI can be defined through a lens of practicality, as a way to automatically learn and codify patterns by looking at huge amounts of interrelated historical data. These patterns can explain why things happened in the past, what is likely to happen in the future and what actions we can take to affect future outcomes.

A good question to start with is: How can AI be leveraged to help solve a real business problem in a measurable way?

In the relocation business, it’s about giving clients maximum value from their mobility program and giving relocating employees an experience that minimizes the disruption to their job, family and life. 

Every relocation is associated with dozens of interconnected data points related to issues such as costs, timing and locations, resulting in a mind-boggling combination of attributes and, therefore, potential causal relationships (meaning which combination of variables was most responsible for a particular outcome).

Before AI, decision support or business intelligence (BI) systems would present complex visualizations about what happened in the past, and the onus was on us to make the right decision based on that complex information. 

Now, AI can tell us what is likely to happen in a very specific context. In relocation, it can predict things like: the total move cost, the probability that a transferee will request a policy exception, or a particular family’s total move duration.

The AI can start forecasting these issues even before the relocation process has started, and it can refine its predictions as the process evolves. But AI offers much more than just predictions — in its most useful form, it can offer explanations and generate context-specific recommendations. 

If old-school BI is a printed road map, then AI-powered recommendations are like driving using GPS, with live traffic updates and turn-by-turn directions. In this analogy, predictions are like knowing whether or not a person will be late in reaching her destination. 

It’s certainly useful, but not as useful as knowing how to avoid traffic so that driver can actually make it there on time.

In a similar regard, while predicting the total duration of a move might be useful, it’s not nearly as useful as knowing how to streamline the relocation so an employee is productive at his new job sooner. 

A client-facing predictive product can sift through and analyze data from past relocations to provide mobility stakeholders with real-time decision support for expenditures, exceptions, authorization volumes, move duration, and total move cost — resulting in better planning and cost management.

This tool can provide employers with a 12-month forward-looking predictive forecast of their mobility program’s key performance indicators (KPI): total authorizations, total expenditures, and total exception costs. 

In addition, at the employee level, it can predict employee move duration and identify those employees who are at a high risk of exceeding move cost estimates and requesting policy exceptions.

The AI constantly analyzes incoming data and updates its predictions everyday as the move progresses. Big picture KPI and fine-grained employee level predictions have helped mobility managers shift from a reactive to proactive posture. 

It’s about exploring AI to look at ways to embed it into the mobility process, providing intelligent and transparent advice to guide decision-making and deliver the best possible outcome for employers and employees.

That means, for example, building analytic applications to enable the following scenarios:  

• A new hire moving to an unfamiliar place receives relocation advice based on what has led to positive outcomes for previous moves that are similar.

• A rising executive embarking on his first long-term expat assignment with his young family learns up front how long things will take, what to expect, and which services and benefits are most critical to settling in quickly.

• A global mobility director designing a new policy receives service and benefit recommendations to minimize exceptions, optimize cost control, and reduce employee churn during a strategic group move.

Keni Patel is head of data science at Cartus in New York. For more information, visit

Latest stories