Hitachi Vantara Exchange New York: 10 takeaways on generative AI, transformation, data management
Hitachi Vantara Exchange New York: 10 takeaways on generative AI, transformation, data management
Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology, digital transformation, and emerging trends shaping the future of business. He oversees research-driven news, analysis, interviews, and event coverage designed to help technology buyers and vendors navigate complex markets with clarity and context. ...
Read more
Commercial generative AI use cases are promising, but CXOs at the Hitachi Vantara Exchange in New York note there's a lot of work ahead--data management, privacy and training models--to scale.
Here's a look at some of the key takeaways from Hitachi Vantara Exchange in New York.
Data management and architecture are the base of generative AI efforts. Mark Katz, CTO of Financial Services at Hitachi Vantara, said the explosion of data, most of it unstructured, means that businesses have to be able to manage data, protect personal identifiable information with segmentation and track lineage. "It's not enough to simply store the data and retrieve it," said Katz. "You need a set of data management tools across enormously complex cloud and hybrid cloud environments."
Debika Bhattacharya, Chief Product Officer, Verizon Business, said enterprises will have to harmonize data to get the most out of generative AI. Bhattacharya said Verizon Business customers are focusing on harmonizing data that's housed in separate towers.
Bharti Patel, SVP, Product Engineering, Hitachi Vantara, said a semantic data plane can be used to bridge these data silos. "The reality is that the data is not going to be static. You have to do it in a metadata way, so the processing is closer to the data," said Patel. "How do you feed LLMs with only the data that makes sense?"
Also see: How Generative AI Has Supercharged the Future of Work | Generative AI articles | Why you need a Chief AI Officer | Software development becomes generative AI's flagship use case | Enterprises seeing savings, productivity gains from generative AI | Work in a generative AI world will need critical, creative thinking
Cyber resilience and privacy are being driven by regulations globally. Part of the data management layer has to be security, so you know exactly what's happening to protect data and recover it at scale, said Katz.
Enterprise use cases for generative AI are just starting. "The thing to keep in mind with generative AI is that it's really early days for commercial use," said Katz. Virtual assistants are a primary use case and consumers are already used to it. Artificial intelligence and machine learning have been used in the enterprise for a while, but generative AI use cases are more experimental.
Katz said credit card companies are an area to watch for generative AI use cases. "More advanced models are coming for new use cases," he said. Indeed, Mastercard launched a generative AI initiative to battle fraud. Jeb Horton, SVP, Global Services, Hitachi Vantara, said enterprises have "varying degrees of embedding AI into what they do."
Bhattacharya said more advanced generative AI use cases are promising. She said personalized healthcare is a fascinating area "as long as the risks are studied and handled the right way."
Santhosh Keshavan, CIO of Voya Financial, said generative AI provides an opportunity to enable investors to make smart decisions from an early age to retirement.
Other use cases that are evolving with generative AI include security, sustainability, customer experience. CXOs noted that use cases for generative AI are almost endless as any process can benefit from a copilot.
Responsible AI frameworks are critical for enterprise adoption. Multiple CXOs at Hitachi Vantara Exchange New York noted that generative AI adoption will depend on trust and responsible use.
Katz said there needs to be traceability into how models were trained and on what datasets. Models should also be explainable.
- The immovable opacity of AI meets the irresistible force of privacy
- Trust In The Age of AI
- Return on Transformation Investments (RTI)
- Here's why generative AI disillusionment is brewing
- BT150 CXO zeitgeist: AI trust, AI pilots to projects
Smaller models will trump large language models (LLMs) in the enterprise. Katz said enterprises are likely to use models that are use case specific for a host of reasons, including lower compute costs, data safety and more predictive behaviors. In addition, limited training data sets, say a corporate knowledge base, will inherently minimize the risk of errors.
Patel added that one large language model to address multiple use cases isn't efficient. "You'll have fine-tuned smaller models that can delegate tasks when they don't know about certain things," said Patel.
"Train the model on corporate offerings, company financials and recommendations designed for a specific use case," said Katz. CXOs at Hitachi Vantara agreed that commercial use of generative AI will initially be very targeted.
Lan Guan, Chief AI Officer at Accenture, said "your propriety data is your biggest advantage” and success is more about the quality of data.
Traditional AI and machine learning technologies can be turbocharged and advanced with generative AI. "AI is not new, but the recent events have shined the light on AI-based applications that have been developed for years," said Bhattacharya. Verizon has been using machine learning and deep learning for networks for insights on the past and predicting the future. "Now we are tying traditional AI to generative AI to create new content," said Bhattacharya.
Keshavan said generative AI can take lessons from traditional AI and data sets and make companies more productive. Voya is looking to generative AI for efficiency gains as well as improving customer service.
Previous AI, data science and machine learning investments are building blocks for generative AI. Andrew Chin, Head of Investment Solutions and Sciences, AllianceBernstein, convinced his company to build out its data science office 7 years ago and today those investments apply to generative AI.
"The lowest hanging fruit was natural language processing so our analysts can gather data faster," he said. "Now we can apply techniques to read documents and make recommendations and summarize for analysts. Humans still have ownership for the final decision."
Take a crawl, walk, run approach with generative AI. Keshavan said his CIO role is part evangelizing generative AI with the business, aggregating use cases and then going from pilot to production.
Bhattacharya said Verizon Business is taking a cross-functional approach that includes security, product, IT and legal to make sure the right policies are in place. "The foundational building block of AI is data," she said. "We prioritize use cases within Verizon and from a product standpoint we look to embed generative AI for new experiences."
Efficiency and cost savings are the table stakes, but enterprises need to eye broader transformation. Accenture's Guan said, "cost efficiency is the table stakes for generative AI, but how do we transform ourselves?"
CXOs at Hitachi Vantara Exchange said one challenge is getting companies to think more in terms of generative AI transformation over productivity.
Dave Malik, Cisco Fellow and CTO, Customer Experience, said the broader transformation due to generative AI will come as more use cases are adopted. "Once there's trust in the system and adoption people will be willing to take more calculated risk with new use cases," said Malik.
Data to Decisions Digital Safety, Privacy & Cybersecurity Innovation & Product-led Growth Tech Optimization Future of Work Distillation Aftershots Next-Generation Customer Experience Big Data SaaS PaaS IaaS Cloud Digital Transformation Disruptive Technology Enterprise IT Enterprise Acceleration Enterprise Software Next Gen Apps IoT Blockchain CRM ERP CCaaS UCaaS Collaboration Enterprise Service Security Zero Trust AI GenerativeAI ML Machine Learning LLMs Agentic AI Analytics Automation Chief Information Officer Chief Technology Officer Chief Information Security Officer Chief Data Officer Chief Privacy Officer Chief Executive Officer Chief AI Officer Chief Analytics Officer Chief Product Officer
51% are evaluating or planning to evaluate open-source LLMs and 48% are planning to implement retrieval-augmented generation (RAG) with LLMs.
