AI presents a whole new set of ethical challenges — and many questions about them. Join us to hear our thoughts on ethical machine learning, removing biases and privacy, and share your own questions and concerns.
If your organization is ready for AI but not sure where to start, don’t miss this chance to get expert advice from our lead data scientist on how to avoid common pitfalls and the questions you and your team should be asking about AI.
https://www.ironsidegroup.com/wp-content/uploads/2020/06/data-scientist-bg.png7681200Michael LeBarronhttps://www.ironsidegroup.com/wp-content/uploads/2018/03/logo-with-words.pngMichael LeBarron2020-04-21 11:09:492020-06-04 09:38:37Getting Started with AI
AI Jump-Start at Cartus to Create a Competitive Edge
Having moved over 2.3 MM people for work in the last 30 years alone, Cartus is the largest global relocation management company (RMC) in the world by volume. For more than 60 years, Cartus has helped relocating employees and their families find their way to new homes, new communities, and new experiences. They do it through an unwavering focus on listening and delivering services, solutions, and expertise that respond to their needs and the needs of their client companies around the world.
Cartus had a strong understanding of business intelligence and data warehousing. They had a structured data environment with some rigor around data governance and business processes for how data was managed during the life cycle.
They were looking to build a data science program that would provide increased customer value, leverage their wealth of data, and help drive them forward and transform from a traditional RMC to a “data-first,” “technology-first” organization within their industry.
SOLUTION: AI JUMP-START
Ironside’s data scientists, data engineers and Experience Design team engaged with Cartus on an “AI Jump-Start” to implement predictive analytics capabilities to enhance their relocation products for their corporate clients with decision support. Ultimately, this would allow Cartus to enable the sales and client services organizations to make their products even more attractive to their customers.
We applied the 5 phases on the AI Jump-Start to ensure that the end result was tailored to Cartus’ needs:
Provide AI Education
Educated both business and technology leadership at Cartus to develop a process to define key elements of strategy and consensus building
Established a partnership between key stakeholders across the organization – not just technical but also people who had a deep understanding of the products, business and processes and how they worked
Secured executive buy-in early in the process by meeting with the CEO to understand Cartus’ objectives and their plan for investment in data science to transform their organization and how they go to market
Establish AI Governance
Identified and established a Steering Committee made up of key stakeholders
Chartered the program and created a forum for communication with channels and process
Established a cadence for communicating results and prioritizing activities through the group of stakeholders
Develop AI Use Cases
Talked to process experts and corporate clients across the organization
Heard from the personas that were important to them in terms of the products they were trying to build and learned what their pain points were
Applied design thinking to the process and envisioned what a decision support product would look like to support the objectives of the individuals
Built a portfolio of data science use cases
Pilot the AI Solution
Ran a series of rapid experiments using DataRobot on AWS Platform to determine where there were gaps in the data and whether a particular use case would be feasible or not
Managed risk very early in the process so Cartus didn’t over-invest in any one direction before knowing if there was efficacy in that use case
Continuously communicated back to the organization to keep people up to speed on progress and results
Pivoted and documented what was learned when met with a roadblock/obstacle
Provided recommendations around how the organization might achieve goals through increased data governance, data quality, data modeling and data collection activities
Deploy Technology for Speed & Scale
Used the use cases and business processes to gain understanding about the user experience and how it all cohesively came together in a product
Built an environment where data scientists can get access to highly granular, detailed information in a raw state, which is critical for doing machine learning, early feature engineering, and the data prep process
Implemented a data wrangling toolset that is accessible to a variety of different skill levels within the data science team or for those using data science methods to solve problems
Leverage Corporate Data Lake on AWS on S3 to be able to access data from variety of sources
Leverage AWS VPC to securely access OnPrem, the AWS Data Lake, and within other Cloud Platforms
Set up DataRobot to manage the lifecycle of a machine learning model
Expressed results in a way that made sense, was easy to navigate and was contextual to the analysis that we were doing
Provided full transparency so they would understand the lineage of any decision or advice
Put model management monitoring in place after deployment to monitor and learn in order to lower the cost of ownership and operating cost of the machine learning environment on an ongoing basis
KEYS TO SUCCESS
For this initiative, Ironside partnered with DataRobot on AWS EC2 to be the catalyst for a data science practice within Cartus: their Leading Edge Analytics Practice, or LEAP. This partnership enabled the non-data scientist knowledge workers to contribute and made the life cycle of the machine learning model easily manageable within a few clicks of the entire platform.
Cartus is a lean team that was trying to build this capability out within the organization in an intelligent and effective manner.
As a consulting organization, we expect certain hiccups along the way and are capable of processing and dealing with any obstacles. When things get busy, we can bring in the right amount of resources to address the issues, and when it slows down, we can ramp down. We encourage our clients to be smart in terms of how they spend their money very early on in the process. This was advantageous to Cartus – we were working with them to build a long-term, sustaining program.
By working collaboratively with Ironside throughout this entire process, Cartus was able to see many benefits, including:
Reduce experimentation cost and risk by reducing cycle times
Increase model effectiveness through greater accuracy
Enable the organization to up-skill existing talent
Put statistical methods in the hands of BI analysts so that more accurate and unbiased business analysis can be done than what is possible with visual methods alone
Cartus wanted to be smart, run lean, and create a great amount of leverage in their data science capability. Cartus did not have any data scientists on staff when we started the process, but they did have operations researchers, industrial engineers and data analysts/domain experts that knew a lot about their business, the data and the processes. We felt that DataRobot on AWS was the perfect platform to be the catalyst for a data science practice within Cartus.
Cartus has used DataRobot to help their corporate clients forecast volume (using time series), predict individual move time and expenses, and even be able to determine when a relocating employee is at risk for exceeding the budget or experiencing a delay in the process.
ENABLE AI AT YOUR ORGANIZATION
Ironside’s AI Jump-Start is an accelerator offering comprised of targeted consulting and a pre-integrated suite of tools that that support the full data governance, management and data science life cycle, with DataRobot being an integral link in that value chain.
The platform is an answer to a recurring problem that we’ve observed in our clients who are new to data science, or perhaps more used to doing traditional data warehousing and BI. They have sometimes struggled to implement a program and architect an environment that is optimized for machine learning and the high-speed agile iterations that are more germane to data science.
We’re seeking to remove this friction from the process and more rapidly create value with machine learning by simplifying this critical early phase of their journey.
Through our experiences we’ve formulated a point-of-view on how data science should work in practice and how to develop use cases that generate real value. We also have a core belief that if we seek to enable more non-data scientists to answer questions in a data science way, using statistical or machine learning methods over visual methods alone, that they and their organizations would be much smarter, and mature faster in their use of data for decision making.
DataRobot is a manifestation of our core values. Our partnership is a combination of platform as well as the compatibility of our approach and people.
https://www.ironsidegroup.com/wp-content/uploads/2019/03/bigstock-Group-Of-Young-Business-People-255902965.jpg18692800Michael LeBarronhttps://www.ironsidegroup.com/wp-content/uploads/2018/03/logo-with-words.pngMichael LeBarron2019-05-20 14:00:502021-04-16 15:29:30AI Jump-Start at Cartus to Create a Competitive Edge
Ironside helps companies translate business goals and challenges into technology solutions that enable insightful analysis, data-driven decision making and continued success. We help you structure, integrate and augment your data, while transforming your analytic environment and improving governance.