Without data, the journey to AI is like a trek through the desert, sans compass
James Fisher & Sons had hearty ambitions to build predictive maintenance capabilities for its customers' subsea cables -- but lacked the right data to do so. In a creative pivot, the IBM Data Science and AI Elite team delivered more than what the heritage engineering company bartered for -- with an entire roadmap for their data science strategy.
The IBM Data Science and AI Elite Team consists of a mix of machine learning experts, data engineers, visualization experts, data storytellers and optimization savants. We represent 20 nationalities and 23 spoken languages. Within our European, Middle East and Africa team, which I lead, women represent well over the industry average of 15 percent.
The intense focus on inclusivity and diversity in our two-year old initiative is not by accident. It’s by design. With such a dynamic, diverse team in place, we turbo charge the AI journey for clients with a combination of a world-class platform for data and AI – Cloud Pak for Data– including its built-in Watson Studio and Watson OpenScale. We couple this with a well-honed methodology that can make a client successful with a proof of concept in six to 12 weeks.
Steeped in an agile approach and design thinking, we enmesh ourselves with clients to drive results regardless of whether they’re just starting out or if they have a center of excellence in place. What unifies all successful engagements is having the right data set to work with at the beginning. But even if your data might not be perfect, or plentiful, we’ve proven that even the smallest sample can be a small but powerful starting point.
Take James Fisher & Sons, a heritage engineering company steeped in UK maritime culture that supports offshore wind farms and shipping companies among others worldwide. When we started the engagement, the first challenge was to identify a quick-win use case where we could show quick value for James Fisher’s growing energy renewables business area.
They were looking for a way to service offshore wind farms with predictive maintenance capabilities, with the goal of detecting and predicting anomalies within a high voltage grid in order reduce operational costs.
While James Fisher has massive amounts of data coming from their partners, we soon discovered that the data we required for our use case wasn’t as available as initially thought. With only a 12-week engagement ahead of us, we had to avoid losing the client’s trust and turn the ship around with the limited data we had.
Rethinking the engagement, we decided to build a project from the ground-up from two different perspectives. The first perspective started with a focus on where most business value can be achieved in the mid- to long-term, and we worked our way backwards to establish a data strategy to support those business objectives. The second perspective started with the data easily available today, and we worked forward to deliver near-term use cases as initial success stories, while keeping the longer term goals in mind. The result was an entire data science strategy roadmap for the company that helped them understand the data they needed to collect, how they could use Watson Studio, and the models they could implement in order to achieve their long-term goals.
What James Fisher ended up with was a long-term strategy for predictive maintenance. At the same time, using our data storytelling skills we were able to create two very concrete proofs-of-concept around anomaly detection of underwater cables, as well as understanding usage of large equipment, such as diggers, to flag incorrect use or abuse, thus increasing the lifetimes of these expensive equipment.
Key to business buy-in was a custom data visualization dashboard that combined the business view and the operational views of the data, including the state of its subsea cables.
With these methods and tools, James Fisher will have access to information about all the data coming from all their assets and fields of business and this data comes to life in a visualization that alerts them to any asset anomalies. That was something they had never been able to see before.
Our ability to adapt to James Fisher’s lack of sufficient data for the initial use case was critical. Being able to provide them a visualization of their data and the benefit they could achieve with those data in place, engendered trust and helped inform the longer-term data strategy. Having a new, powerful way of analyzing subsea cables through data science has unleashed a desire to explore AI in other parts of the business, such as knowledge transfer and safety procedures.
Even without a compass, we were able to help them blaze a path forward, by mapping out the stars.
Unlock the value in your organization with AI. Explore the Journey to AI, a prescriptive approach that helps you modernize, collect, organize, analyze and infuse all your data.
Find out how the IBM Data Science and AI Elite team can help you build out your AI strategy.
Learn the secrets to a successful AI engagement. Get Agile AI: A Practical Guide to Building AI Applications and Teams.