European Business Magazine (EBM) caught up with Ben Owen (Pictured), Cloud Director at Cloudera spearheading Public Cloud business throughout the EMEA region (Europe, Middle East, and Africa). With a strategic focus on championing the advantages of cloud technology through the lens of trusted data, Ben plays a pivotal role in harnessing the power of Generative AI.
Tell me about the recent study Cloudera published? We surveyed 850 IT decision-makers across the EMEA region, uncovering surprising insights about cloud which helps to lead to Generative AI adoption.
- Finding 1: 92% plan to increase their cloud data in 1-3 years, yet 76% are considering moving some data back on-premises.
- Finding 2: This dichotomy arises from concerns around data governance, fears around cloud lock-in, cybersecurity, performance, and costs. Interestingly, cost is the fifth most significant factor.
- Finding 3: The motives for cloud adoption include improved data accessibility, optimized storage and backup, cost reduction, faster innovation, and sustainability.
The two strategies of cloud migration and data repatriation complement each other. Migrating the right workloads to the cloud can be cost-effective, secure, and compliant, avoiding vendor lock-in. While at the same time taking advantage of the benefits of cloud such as flexibility, scalability, and agility. However, simply transferring existing applications to the cloud without optimization can be counterproductive due to the inefficiencies of legacy code not designed for modern cloud technologies. If unmanaged or, for example, not effectively secured, running in the cloud can actually be more costly and disruptive. So cloud first is not always the right answer, and organizations should instead focus on the workload first when determining where data sits. Having the capability to work both in the cloud (or multiple clouds) and on-premise with the same optimized tooling is critical to a long term data, Advanced Analytics, and Generative AI strategy.
What are the primary challenges for businesses trying to adopt Generative AI?
When I meet with our customers and partners, the two main challenges to Generative AI adoption that stick out are a lack of trusted data, and cost – both the operational control that comes with Generative AI, and the high costs of entry.
The Need for Trusted Data: Quality data is essential for driving impactful insight from Generative AI. Many enterprises are still in the early stages of establishing control, security, and governance over their data. And data is often isolated in silos across an organization, hindering the ability to make informed decisions without costly integration. Utilizing unmanaged or irrelevant data not only risks inaccurate results (or “hallucinations“) but also fails to deliver value from business queries. Leveraging quality data helps to reduce hallucinations, but more importantly provides the organizational context and information to add value. The main reason for trusted data comes down to the following three pillars; Relevance and Accuracy, Trust and Safety, and Risk and Compliance.
Implementing Generative AI in-house incurs substantial costs. Outsourcing this process can introduce risks and compliance issues. For perspective, consider the cost of cutting-edge GPUs like the Nvidia H100, essential for running large language models. A single unit costs around $30,000, and training a model with 175 billion parameters might require over 2000 of these GPUs, amounting to tens of millions in investment. Of course this is a more extreme example as training a model requires substantially more horsepower than running a pre-built one. However having the capacity to run a farm of GPU resources can be an expensive investment.
The solution lies in leveraging cloud services. Cloud providers, with their vast GPU resources, offer a flexible and secure environment for businesses to experiment and run Generative AI. This approach enables companies to scale their use of resources and only pay for what they use, without committing to large upfront investments. As a business grows, it may choose to move some operations back in-house for cost efficiency, but the key is maintaining an open strategy with the ability to adapt as needed.
How can customers best adopt Generative AI assuming they are starting from scratch, is there a roadmap?
In the journey to adopt Generative AI effectively, businesses can follow the BRIESO model, which stands for Build, Refine, Identify, Experiment, Scale, and Optimize.
Build: The first step should be to construct a modern data architecture and create a universal enterprise data mesh. This enables organizations to gain control and effectively manage their data, whether on-premises or in the cloud, establishing a unified ontology and strategy for mapping, securing, and ensuring compliance across all data silos. Choose tools that meet current needs and accommodate future growth, with a preference for open-source solutions for greater flexibility.
Refine: Next, refine and optimize data sets according to current business needs, while anticipating future requirements as accurately as possible. This is important as companies do not want to migrate useless data which will later increase costs.
Identify: Identify opportunities to leverage the cloud based technologies on specific workloads. Perform a workload analysis to determine where the most value can be derived. Connect data across locations, whether on-premises or in a multi-cloud environment, and consider potential use cases for development.
Experiment: Try out pre-built Generative AI frameworks offered by major cloud providers, like AWS’s Bedrock (Hugging Face), Azure’s OpenAI (ChatGPT), or GCP’s AI Platform (Vertex), among others. The key is not to commit to one platform prematurely; instead, find the one that aligns with your business needs. Ensure integration with trusted data, as the real value lies there, not just in the Generative AI model.
His responsibilities include fostering collaboration with industry giants such as Amazon Web Services, Microsoft Azure, and Google Cloud Platform, with the ultimate goal of optimizing clients’ investments in the cloud. Join us as we delve into Ben’s insights on the intersection of cloud innovation, trusted data, and the transformative potential of Generative AI for businesses across the EMEA landscape.
Scale and Optimize: Once a suitable platform is identified, pick one or two use cases to scale into a production model. Continuously optimize the process, whilst remaining mindful of the costs associated with GPU usage. As companies expand their Generative AI capabilities, look for opportunities for economical optimization. Flexibility in their chosen platform is crucial for long-term success.
You attended AWS Re:Invent in 2023. What were your key takeaways?
At AWS re:Invent 2023, the spotlight was unmistakably on Generative AI. Everywhere you looked, from the conference center to the Las Vegas Strip, Generative AI was a pervasive theme. However, it became clear that despite the widespread buzz, actual deployment of Generative AI in production remains limited. Many Independent Software Vendors (ISVs) were discussing future plans, showcasing prototypes, or conducting Proof of Concept (POC) projects. In contrast, Cloudera stands out as we already have clients successfully running Generative AI at scale, handling millions of transactions daily in real-world applications. This signifies a shift from mere potential to tangible, operational implementations of GenAI. The industry, as a whole, needs to pivot from the hype surrounding Generative AI to solving concrete business challenges. Another significant highlight from the event was the acknowledgment
My professional journey encompasses diverse leadership experiences, from executive roles at SAP to the launch of a startup SaaS product, and an immersive exploration of the hyperscale domain during my tenure at AWS. Currently, at Cloudera, I find myself at the forefront of the upcoming IT transformation, with a distinct emphasis on harnessing trusted data to propel Generative AI. Despite my American roots, I’ve been based in Europe for nearly 19 years, including a period where I commuted to the Middle East for a year or two.”
Think of it like this: 25 exabytes is a really staggering amount of data. It’s as if you take every single cell in a person’s body – about 33 trillion cells – and then multiply that by the number of people living in a big city like Frankfurt or Amsterdam.
This aids clients in reducing costs and risks, enhancing productive of Cloudera in AWS’s announcement of the new Amazon S3 Express One Zone service. This is testimony to our relationship with AWS, symbolizing the synergy and alignment between our teams. The S3 Express service enhances performance for our clients managing large data sets in the cloud. This announcement by AWS, coupled with our strategic collaboration agreement signed earlier in the year, is a testament to our growing partnerships not just with AWS but across the hyperscale landscape, including Microsoft Azure and Google Cloud Platform. This collaboration is the cornerstone of our go-to-market strategy and product development in the coming years, demonstrating the strength and depth of our industry alliances.
How is your company approaching generative AI?
Cloudera’s approach to Generative AI mirrors the role of merchants “selling the shovels” during the historic gold rushes: providing the crucial tools and knowledge needed for success. As a pioneer in the open source data movement, Cloudera’s ethos revolves around enabling data access everywhere and integration anywhere. This has given us a head start on the Generative AI revolution as we truly believe with your own trusted data enables trust in your Generative AI. This approach ensures future-proof solutions, allowing customers the flexibility to switch processing engines, AI models, or cloud environments without losing control of their data.
The focus areas for Cloudera in the realm of Generative AI include addressing scale, costs, and security – key concerns for our customers.
- Scale: In the vast landscape of data, size and scope are paramount. Cloudera adeptly manages over 25 exabytes of data, equivalent to more than 25,000 peta bytes, showcasing our capability to handle data at an extraordinary scale. We have more data under management than many hosting providers.
- Costs: Balancing cost-efficiency with the need for innovation is a common challenge. Cloudera’s integrated data services range from streaming to data cleansing, enabling clients to concentrate on business outcomes rather than the intricacies of technology.
- Security: In today’s data-driven world, security is a top priority. Cloudera stands out by offering comprehensive security and control over a wide array of data types, both on-premises and in AWS environments, through our open data lake house model.
At its core, Cloudera empowers organizations to transform their data into trusted enterprise AI. This aids clients in reducing costs and risks, enhancing productivity, and expediting business performance. With Cloudera, companies can trust their data and the AI systems they deploy, paving the way for reliable and innovative solutions.
How are European companies adopting the cloud overall?
From what we see and according to the EC Statistics agency more than 42.5% of EU companies adopted some form of Cloud in 2023. This is up more than 4% since 2023 which is a large increase. For comparison American companies have just a slightly higher rate of adoption of 48%. So, the gap is closing rapidly. Our own data also shows a much larger percentage here in EMEA, as 92% plan to migrate some data to the public cloud within the next three years. Our customers typically lean towards the large enterprises as they have large data estates to manage, but regardless the growth is outstanding.
What are some example business outcomes your customers are getting from their data on Cloudera?
Cloudera’s customers are truly out in front of AI. We don’t just talk about the Generative AI – our customers are deploying it daily!
We also have a number of other benefits:
Speed of Innovation – Enabling Developers: One of our financial services customers has boosted developer productivity by 20% by standardized code quality leading to faster innovation.
Productivity Improvements that Save Lives: In healthcare, we have customers like BeTheMatch (USA), which is able to speed up the complex process of searching for suitable bone marrow donors when trawling a database of more than 41 million records, with a large amount of computations. Cloudera on AWS streamlined the process and has sped up the search results by a factor of 10 times. This is impacting and saving lives and will only keep improving.
Providing a Foundation for Innovation: In the public sector, Norway’s IFE (Institute for Energy Technology) uses Cloudera to underpin its efforts as part of the My Digital City smart city program in Halden. This helps the population adapt to climate change and create new jobs by becoming a zero-emission society. IFE also uses Cloudera to support a data factory aligned with the EU’s Digital Single Market strategy to provide AI services for Norway’s public sector and SMEs. This enables a ‘test before invest’ infrastructure in the cloud.
Keeping our Roads Safe: Manufacturing company, Continental in Germany, is using Cloudera in the public cloud for ContiConnect. Continental can now manage its streaming engine, database, and related technologies from one central point, and has seen improved interaction between services, alongside greater ease of use and security. Providing real-time insights through Cloudera’s data analytics capabilities has helped the fleet customers of Continental reduce breakdowns by 83% and achieve average monthly fuel savings of £1.646, with out of hours calls reduced by 95%. Continental also aims to also enable additional data driven, AI related use cases within the solution.
Helping Telcos prevent Outages: Fastweb Telecommunications in Italy wanted to enhance its service by answering business requests in real time and to leverage ML and AI to anticipate potential network disruptions. These two goals need to process large amounts of data in streaming, without network disruptions, while keeping costs low.
Leading the Way in Medical Drug Discovery: Large pharmaceutical companies like Abbvie in the USA are bringing life-saving drugs to market faster; with massive improvements in productivity in R&D to improve quality of life and health outcomes.
Building strong customer relationships: OCBC Bank in Singapore successfully seized the opportunity to personalize banking experiences. The bank’s team built Next Best Conversation, a centralized platform that uses Generative AI, LLM and ML to analyze real-time contextual data from customer conversations related to sales, service and more.
The bank increased its revenue considerably and nearly doubled its campaign conversion rates. This was achieved by using the data to identify the most relevant information for each customer and curate personalized experiences across communication channels. With ML models, OCBC Bank could also send over 100 different personalized nudges on its mobile banking app. This notifies customers about financial opportunities including eligibility for a new credit card or loans—achieving up to 50% click through rates.
Banking customers also enjoyed faster and more efficient transactions. For instance, OCBC’s Banks data science teams developed chatbots that handled 10% of customer interactions on the website. The bank also used ML to predict the potential time to failure of several bank systems, ensuring that IT teams could take preemptive actions to keep data centers always up and running. The ML models have also helped the bank to reduce risk of losing sensitive customer data, such as financial details, and avoid costly regulatory fines from downtime. As part of its future strategy, the bank plans to make AI/ML available to more of its systems and users. Cloudera provides a strong and flexible foundation to help OCBC Bank integrate AI at scale throughout the organization and drive more customer innovation and operational efficiency.
Future – How does your team plan to use these services in the future to serve your customers? How do you foresee Generative AI evolving?
Cloudera is a leader in the top industries in the world like Financial Services, Telcos, Automotive, Energy, and Lif eSciences. We’re just scratching the surface of what we will build using Cloudera and Generative AI Cloud services. We have a roadmap of cloud integrations in our development pipeline, and are working with AWS, Microsoft Azure and Google Cloud Platform on advanced integrations.
