Skip to main content

Improve Operations by Employing AI and SAP HANA on Power Systems

Learn how Oxford Cancer Biomarkers, Wells Fargo and Ryerson streamlined processes by deploying SAP HANA and AI on POWER.

Blue background with red and orange illustration of factory, healthcare and finance.

Image by Fernando Volken Togni

Different goals, different tools. Makes sense, given no two businesses have the same requirements, even if they exist within closely aligned industries.

This mindset keeps everyone on their toes and looking for novel ideas arising from new or existing technologies to improve operations. For example, artificial intelligence (AI) and SAP HANA on Power Systems* can be deployed for any number of purposes for a varied list of disciplines.

Massive AI Throughput

Research institute Oxford Cancer Biomarkers (OCB), for instance, is using IBM PowerAI Vision to more quickly diagnose and devise treatments for cancer patients. 

Lifestyle choices can reduce the risk of cancer, the second-leading cause of death worldwide, but those who contract the disease, have better outcomes when it's detected early on. As a result, identifying genetic predispositions or early signs of cancer is a priority in oncology research.

Established in 2012, OCB’s primary mission is to discover and develop biomarkers (quantifiable biological parameters that provide insight into a patient’s clinical state) to advance personalized care within oncology, with a focus on colorectal cancer and targeted treatment programs, including cases of recurrence.

With that goal in mind, OCB wanted to move away from a one-size-fits-all approach to cancer care by giving clinicians as much information about a patient’s condition as possible. The ultimate goal is to develop screening tests that discover these insights more quickly and cost-effectively and less invasively than existing methods.

Working with IBM Business Partner Meridian IT, OCB began the development of image-analysis algorithms to enhance its existing ColoProg platform. Using IBM Power Systems* AC922 servers hosted by Meridian IT, OCB is using deep learning models to augment its proprietary DNA ploidy (a measure of the DNA content within tumor cells) and stroma content (non-malignant cells that can provide an extracellular matrix on which tumor cells can grow) assays to classify resected tissue samples. The organization combines these biomarkers to stratify patients into groups at low, intermediate and high risk of colorectal cancer recurrence, enabling better clinical decision-making. 

Designed for image classification, PowerAI Vision running on Power Systems AC922 has allowed Meridian IT to quickly build the models OCB needed to reach its goals. The team took advantage of the massive throughput capability offered by the IBM Power Systems AC922 servers, which are equipped with NVIDIA with NVLink GPUs to expedite model training, bringing ColoProg enhancements to fruition sooner than expected. Previously, it took months to years to build and test new models. Now, it can be done in hours.

Thanks to these improvements, OCB is providing insight that supports better informed clinical decisions following cancer surgery, improving patient outcomes. And by minimizing overtreatment with chemotherapy, this AI-based solution is also reducing patient distress and the cost of care. Based on its success with the enhanced ColoProg tool, the organization is now exploring how to extend the solution to breast and prostate cancers.

Using AI for Risk Mitigation

Due to several high-profile financial issues in recent years, every bank in the U.S. is likely to have a chief risk position, as is the case with Wells Fargo. But this job comes with high occupational hazards, because risk models can go wrong, resulting in large financial impacts for banks.  

They must file timely financial reports, and monitor and prevent fraud and financial crimes. Banks also need to make decisions in terms of credit risks and manage market-risk issues.  The list of these risks is seemingly endless. 

Regulators also want to take a look at every AI-based risk model—no small feat, given organizations such as Wells Fargo can have thousands of them. To avoid regulatory issues, every model must be carefully managed, from inception to operations to change control.

As with every other bank, Wells Fargo operates at a rapid, near-real time pace, making risk monitoring and detection a daunting task. It’s also dealing with very large amounts of data from, for example, individual account levels, which may include mounds of historical data.

To keep up with this, Wells Fargo applies IBM Watson* Machine Learning Accelerator and Power Systems AC922 to check for credit risk, fraud and compliance, and to monitor email using natural language processing. Computationally intensive business processing such as a derivative pricing also require machine learning in a now-based environment that requires fast movement in terms of data from CPU to GPU. All of this requires sophistication in both hardware and software to be manageable, which Wells Fargo found in its Watson Machine Learning Accelerator and Power Systems AC922 solution.

Customer Service With SAP HANA

The manufacturing industry also has particular requirements and goals. For example, Ryerson, a metals processor and distributor, wanted to distinguish itself from competitors by improving customer service functionality. Unfortunately, its supporting IT infrastructure needed an upgrade to deliver on this objective.

This upgrade was critical, because the global metals processing and distribution industry is highly fragmented, with many businesses jostling for marginal gains. Any competitive edge a company can achieve—no matter how seemingly slight—through improving efficiency or providing value-add services can result in big-time benefits.

Speedy response times to customer queries is a major focus for Ryerson and differentiates it as a customer service leader. As a result, a core aspect of the company’s IT department’s work is to enable the rest of the company, especially its sales teams, to work efficiently by providing excellent technical delivery at all times.

To help improve its customer-service deliverables, Ryerson wanted to develop an e-commerce site, add self-service functionality, and provide round-the-clock support for its customers. However, its version of SAP ECC was reaching end of life, which made integrating additional functionality into the system extremely challenging. After looking at possible solutions that would help it reach its e-commerce goals, Ryerson selected SAP Business Suite powered by SAP HANA to provide the best strategic
route forward.

To optimize the power and performance of the new SAP solutions, Ryerson next chose to implement two IBM Power Systems E880C servers running the SUSE Linux Enterprise Server for SAP applications OS. IBM demonstrated to Ryerson that it could run SAP Business Suite powered by SAP HANA with just two servers, compared to another platform that would have required 14 servers to deliver the same capabilities. By using the advanced virtualization capabilities of IBM PowerVM* virtualization software, Ryerson can easily scale capacity as required to meet its rigorous reliability standards.

To enable the E880C deployment, Ryerson worked with Meridian IT to design and deploy the infrastructure, implementing the Power Systems servers and creating the virtualization layer. Ryerson also selected IBM Systems Lab Services to help implement and configure SAP Business Suite and migrate from the company’s former database solution to SAP HANA. This team—Ryerson, Meridian IT and IBM—completed the project in around eight months, on time and on budget.

Ryerson now has the agility and raw system performance to help it realize its customer-service goals. Its sales teams are completing sales processes for ready-to-ship products 14% faster, while the new solution has cut the time needed to finalize custom orders by about 7%. Ryerson also now completes batch processing tasks more quickly, with day-end and week-end closing 28% and 10% faster, respectively, on average. With these essential processes now completing more rapidly, sales staff are free to spend more time on direct client engagement and other value-add tasks.

Future Goals

In some cases, off-the-shelf solutions are perfectly suitable to meeting operational goals. But increasingly, organizations are looking at new bespoke methods to improve operations. These can include AI or user-specific tools such as SAP HANA.

Thankfully, IBM provides a platform in Power Systems that offers many benefits, including raw performance, AI acceleration, robust AI support tools, simple-to-implement virtualization environments, server consolidation and scale-as-needed capabilities. All of these features help clients reach their goals—and leave room to accommodate future goals.

IBM Systems Webinar Icon

View upcoming and on-demand (IBM Z, IBM i, AIX, Power Systems) webinars.
Register now →