Blogs

When selecting a conversational AI platform for use in a regulated industry, remember that cutting corners can have disastrous consequences. When virtual assistants aren’t developed and deployed with an eye toward regulatory compliance, customer experience will suffer. Even worse, your organization risks running afoul of the many ever-shifting regulatory frameworks in which they operate.

HIPAA, GDPR, and SOC2 are just a few of the common frameworks that may constrain the behavior of your virtual assistant, the data it uses and how that data is stored and protected. You’ll want to be sure that your chatbots are designed and deployed with an understanding of all relevant regulations: those that apply to your industry and the regions in which you and your customers operate.

For example, consider GDPR. This framework was created to harmonize data protection law across the EU, but it imposes strict rules on organizations hosting and processing data anywhere in the world. Violations are costly – in 2020, the EU collected more than $191 million in penalties for GDPR violations, up 40% from the prior year. Virtual assistants need to provide users a simple way to access, review and download electronic copies of their data. Users should be able to delete their information if they want. You’ll need to evaluate what information your organization is allowed to store in the form of chat logs, and what you can share with third parties. Your organization must also take reasonable care to protect customer data against a network breach.

IBM has invested heavily in making Watson Assistant GDPR-ready, so that our customers can achieve compliance easily and with a high degree of control, enabling them to be ready to respond when regulations inevitably change. Our conversational AI platform allows you to easily opt out of log data use before data is collected or created, and to label and delete existing data that doesn’t conform to GDPR rules.

Watson Assistant is well-suited to organizations in industries that must comply with strict regulations on how customer data is used, such as healthcare and finance. Let’s look at a few real-world examples of how this can play out.

Finance

Watson was crucial to financial institutions adapting to COVID restrictions while adhering to industry regulation. When restrictions hit Europe in 2020, call volume at some banks increased by a factor of 20 to 30. On NatWest’s service lines, customers were struggling to get through to a representative to ask about access to services and the financial implications of the pandemic. NatWest’s AI team sprang into action with their virtual assistant “Cora,” which was built using Watson. Cora fielded customer service inquiries, transforming customer experiences and providing support to agents.

In Australia, leading bank Westpac also improved contact center efficiency with conversational AI. Their bot fielded over 215,000 chats, resolving over 70% of customer inquiries without having to engage an agent.

Auto loan provider GM Financial built a chatbot with Watson Assistant called “Nanci,” which answered 60% of customer interactions. The company also employed text analytics software to listen and learn from their customer interactions. “IBM’s commitment to cybersecurity and regulatory compliance allows us to rest easy, knowing that our customer data is in good hands,” said Bob Beatty, Executive Vice President, Chief Experience Officer, GM Financial.

Healthcare

In healthcare, organizations need to be especially aware of how they’re collecting, storing, and using customer data. In the U.S., for example, virtual assistants operating in the healthcare space must follow all HIPAA requirements, such as in-transit and at-rest encryption, strong passwords, training for employees, and more. Watson Assistant is designed to make implementation of these capabilities straightforward.

The University of Arkansas for Medical Sciences deployed a virtual agent to field incoming COVID-related queries about testing, symptoms, and other resources. Average registration time has been reduced by 50%. And the Andhra Pradesh National Health Mission portal implemented a virtual agent to help residents get quick answers to their COVID questions. This bot speaks English, Telugu, and Hindi. The virtual agent is deployed in a web browser and is built to safeguard user privacy.

Over the next few years, we can expect regulations to grow even more complex, and their consequences more severe. Watson Assistant helps the above organizations stay compliant while adapting to — and thriving under — new challenges. See what it can do for you.

When the COVID-19 pandemic triggered lockdowns in 2020, grocery stores were one of the few businesses that experienced a significant jump in demand. With restaurants closed, consumers hit grocery stores in droves to stock up on ingredients to prepare meals at home.

United States supermarket operator Albertsons Companies was among those that experienced a significant spike in business during the health crisis. In March and April of 2020 alone, sales were up 34% compared to the previous year.

Greater demand meant Albertsons needed to expand its workforce and modernize its IT infrastructure. Fortunately, a few years prior, the company initiated a plan to move its highly customized on-premises human resource applications to the cloud. The move would give the $70 billion grocery giant a fast and cost-effective way to scale payroll for its nearly 300,000 employees.

Founded in 1939, Albertsons is the second-largest grocery chain in North America, with 2,252 stores. The corporation operates multiple subsidiaries, including Safeway, Vons, Tom Thumb, Jewel, Pavilions, Shaw’s, Star Market, Haggen, Andronico’s, and ACME.

As competition in the U.S. grocery business grows and margins shrink, stores are under increasing pressure to provide a stellar customer experience. That means they must be more agile and adaptable. Toward that end, in 2019, Albertsons Companies launched a corporate mandate to move out of its corporate data centers and modernize its technology platform with a cloud-first strategy. As part of that plan, Albertsons selected Oracle Fusion Cloud ERP and Oracle Fusion Cloud HCM, which will provide platforms for continued modernization and productivity.

Albertsons’ enterprise transformation to Oracle Cloud HCM and ERP would take about 36 months and require several implementation waves. In the meantime, the company sought an interim solution to move its legacy PeopleSoft Payroll application to the cloud quickly. Because the existing software was nearing end-of-life, the company wanted to reduce the time it took to complete its weekly payroll runs, while scaling the business as it continued to grow. Albertsons also needed to confidently adhere to a tight weekly payroll cycle for about 265,000 employees.

After evaluating several options, Albertsons decided the best path forward was to transition its existing PeopleSoft Payroll environment to Oracle Cloud Infrastructure (OCI), including moving its Oracle database to Oracle Exadata Database Service.

“The PeopleSoft expertise Oracle has and the benefits and support we would receive from the Oracle Advanced Customer Services team to help us migrate PeopleSoft to OCI was a large factor in our decision,” says Albertsons Solutions Architect Sambuddha Ghosh.

A unique virtual cloud network topology

To make its plan viable, Albertsons’ PeopleSoft Payroll application was moved to a unique multiple virtual cloud network (VCN) topology on OCI, managed by the Oracle Advanced Customer Services team.

“Working with the Oracle Advanced Customer Services team has been awesome,” says Ghosh. “Oracle helped us understand how we could successfully migrate everything that was either end of life or close to end of life. We came up with a timeline for the migration and the team provided us with updates every step of the way, which helped us plan the project effectively.”

Albertsons has five VCNs. One (VCN00) is a transit VCN for external users, who can access Albertsons network on OCI from the internet and through a Web Application Firewall (WAF). From there, external users are sent to an Oracle public Load Balancer. Once inside the grocer’s public subnet, external users are routed to a WebLogic/ForgeRock Identity Gateway, which validates their single sign-on (SSO) and authorizes their access policies in an Azure Active Directory. This external transit VCN is peered to a second VCN (VCN02) for production, both of which are located in Oracle’s San Jose, California, cloud region. This production VCN is then peered to a third VCN, located in a disaster recovery and non-production cloud region in Ashburn, Virginia.

In a fourth VCN (VCN01), Albertsons internal users also connect from a site-to-site VPN, but are then moved to a private load balancer. From Albertsons’ private subnet, internal users are routed to the WebLogic/ForgeRock Identity Gateway, which validates their SSO and authorizes their access policies in an Azure Active Directory. This transit VCN is also peered to the production VCN in San Jose, which is then peered to the disaster recovery and non-production environment in Ashburn.

The fifth VCN (VCN04) is also a transit VCN and is used by OMCS personnel. These users actively manage and support the company’s OCI tenancy, and access the VCN through a site-to-site VPN, before passing through a dynamic routing gateway.

In Albertsons’ production environment, PeopleSoft Payroll applications run on eight virtual machines, and all the payroll data is stored in an Oracle Database 19c, which runs on a quarter-rack Exadata Database Service, supporting two node Oracle Real Application Clusters (RAC) to maximize availability while accessing shared storage. In the secondary region in Ashburn, two additional Exadata Database Service quarter racks support both non-production and disaster recovery environments. Oracle Data Guard replicates the database and the non-database components to disaster recovery (DR). The DR environment is deployed with the identical configuration as production, with one exception: The DR database is located on one of the Exadata Database Service servers, but shares this database with a collection of non-production environments. Albertsons chose Oracle Enterprise Database Service to support its third-party software applications, with these databases running on virtual machines.

Improving speed and flexibility

Since moving PeopleSoft to OCI in May 2021, Albertsons produces its weekly payroll 50% faster than when it ran those jobs on-premises.

“Because we run payroll weekly, it’s critical to make sure all the different divisions and groups are paid within a proper timeframe,” says Ghosh. “The Exadata migration helped us immensely on that front. We have seen a huge gain in performance and the overall time it takes to run payroll has improved significantly.”

In addition to improving the efficiency and core horsepower of its payroll processes, Albertsons can now scale OCI at predetermined times. When its non-production environments aren’t in high demand, they can be quickly powered down. When those resources are needed again, they quickly scale back up. This process allows Albertsons to optimize its OCI consumption and control costs.

Ultimately, moving its on-premises PeopleSoft Payroll workload to OCI provides Albertsons better flexibility so it can adapt quickly to sudden growth or other changes in the future. It also allows the company to focus on its longer-term migrations to Oracle Cloud HCM and ERP, providing a clear transition path to align with its new corporate cloud strategy. But perhaps mostly critically, the migration allows Albertsons to modernize its entire IT infrastructure without creating any disruptions in its payroll operations.

We are excited to announce the availability of Red Hat Enterprise Linux 9 (RHEL 9), the latest release of the world’s leading enterprise Linux platform. RHEL 9 provides a more flexible and stable foundation to support hybrid cloud innovation and a faster, more consistent experience for deploying applications and critical workloads across physical, virtual, private and public cloud and edge deployments.

What’s new?

RHEL 9 includes features and enhancements to help achieve long-term IT success by using a common, flexible foundation to support innovation and accelerate time to market.

Primary features and benefits

Here are a few highlights of what’s included in RHEL 9.

A new platform for developers today and in the future

Completing the migration to Python 3, version 3.9 will be the default Python for the life of RHEL 9. Python 3.9 brings several new enhancements, including timezone-aware timestamps, the recent string prefix, suffix methods and dictionary union operations to help developers modernize existing apps.

RHEL 9 is also built with GCC 11 and the latest versions of LLVM, Rust and Go compilers. RHEL 9 is based on glibc 2.34 for 10+ years of enterprise-class platform stability.

And finally, for the first time in RHEL, Link Time Optimization (LTO) will be enabled by default in userspace for deeper optimization of application code to help build smaller, more efficient executables.

Easy contribution path to future versions of RHEL

Organizations can now develop, test and contribute to a continuously-delivered distribution that tracks just ahead of RHEL. CentOS Stream, an upstream open source development platform, provides a seamless contribution path to the next minor release. RHEL 9 is the first RHEL major release built from CentOS Stream, and the RHEL 9 Beta was first available as CentOS Stream 9. All future RHEL 9 releases will be built from CentOS Stream.

Next-generation application streams

Building on the introduction of application streams and module packaging in RHEL 8, all packaging methods in RHEL 9 are incorporated into application streams, including modules, SCLs, Flatpacks and traditional RPMs, making them much easier to use.

Continuing commitment to multiple architecture support

Open source software gives users greater control over their digital future by preventing workloads from being locked into a specific vendor. RHEL extends this control beyond the source code by enabling diverse CPU architectures for users that need an evolving business environment. Whether you're running your workload on x86_64, aarch64, IBM POWER9, Power10, or IBM Z, we have you covered.

Container improvements

If you're building applications with universal base image (UBI) container images, you'll want to check out the RHEL 9 UBI images. The standard UBI image is available, as are micro, minimal and the init image. To get the entire experience, test the UBI images on a fully subscribed RHEL 9 container host, allowing you to pull additional RPMs from the RHEL 9 repositories.

RHEL for edge

RHEL 9 introduces automatic container updates and rollbacks, which expands the capacity to update container images automatically. Podman can now detect if an updated container fails to start and automatically roll the configuration back. Together with existing OS-level rollbacks, this provides new levels of reliability for applications.

Image Builder as-a-Service

Enhancements to Image Builder in RHEL 9 help organizations save time and drive system consistency at scale. With the new Image Builder as-a-Service, organizations can now build a standardized and optimized operating system image through our hosted service and deploy it to a cloud provider of choice.

Identity and security

New capabilities added to RHEL 9 help simplify how organizations manage security and compliance when deploying new systems or managing existing infrastructure. RHEL 9 now offers Integrity Measurement Architecture (IMA) to dynamically verify the integrity of the OS to detect if it has been compromised. RHEL 9 has also been enhanced to include digital signatures and hashes that help organizations detect rogue modifications across the infrastructure.

Automation and management

Organizations now have access to the enhanced performance metrics page in the RHEL 9 web console to help identify potential causes of high CPU, memory, disk and network resource usage spikes. In addition, customers can more easily export metrics to a Grafana server. Kernel live patch management is also available via the web console to significantly reduce the complexity of performing critical maintenance. The console also adds a simplified interface for applying kernel updates without using command line tooling.

Predictive analytics

Red Hat Insights now encompasses Resource Optimization, which enables right-sizing RHEL in the public cloud. Resource Optimization does this by evaluating performance metrics to identify workload utilization. Insights then provides visibility and recommendations for optimizing to a more suitable instance for the workload needs. Insights also adds Malware Detection, a security assessment that analyzes RHEL systems across the enterprise for known malware signatures and provides detailed visibility into the risk.