Étude de cas SoftwareOne

DNV GL Maritime

DNV GL Maritime is the leading provider of risk management and quality assurance services to the maritime, oil and gas, and power and renewables industries.

The company is also a global leader in certifying management systems of companies across all types of industries, including healthcare, food and beverage, automotive, and aerospace.

The challenge

While exploring various emerging technologies as part of its digitalization journey, DNV GL Maritime saw the potential of automation enabled by AI and machine learning.

However, the company was puzzled about how to get started with implementing these advanced techniques and needed guidance to ensure value realization early on.

The challenges identified included the overwhelming workload at the help desk, which presented an opportunity for productivity enhancement. With 80,00 support cases per year, the staff found it difficult to keep their promise to respond promptly and struggled to find enough time to manually route each email to the right expert.

Another area for improvement pinpointed was the planning of field surveys, which, if not done right, could cost the inspectors significant time and manual effort.

Client: DNV GL Maritime
Industrie: Transports
Plateforme: Azure Cloud
Services: Cloud services
Pays: Norvège

The solution

SoftwareOne (formerly Crayon) started out by helping DNV GL Maritime stake out a clear path toward realizing business value from machine learning.

The support cases were tackled by training machine learning algorithms to process and interpret texts using text classification models on 250,000 archived emails.

Eventually, the algorithms were able to accurately route 82% of customer support messages to the right department (the same accuracy rate as a human). This reduced the manual workload by up to 70% and saved 2–3 FTEs per year.

The inefficiencies in the scheduling of surveys were addressed with an advanced regression machine learning model and a matching algorithm for vessels, survey plans, and inspections.

By matching inspector and vessel locations and predicting the amount of time required, a new smart survey booking tool was created. This added tremendous value for both DNV GL Maritime and its customers and significantly cut down on split and remote surveys.

Your team did a tremendous job mentoring our domain experts in understanding the vision and concept of what machine learning can do. Your mentoring and skills transfer was vital for us to understand the potential of AI and to communicate that potential to our organization.

Stian Barkbu Nilsen

Head of Information and Production Systems, DNV-GL