As they say: Rubbish In, Rubbish Out — and where AI is concerned, getting your organization’s data primed for AI/ML is critical
Now that ChatGPT is making waves around the world, people are actually referring to non-generative AI as “boring AI”.
However, AI — whether in boring or generative form — is an important part of a digital world. As Nick Eayrs, Vice President, Field Engineering (Asia Pacific and Japan), Databricks explained to this DigiconAsia contributor, organizations will need a good blend of generative AI and “boring” AI to solve business challenges, optimize operations and enhance customer experience. The key to AI is quality data. There is no AI without data.
DigiconAsia: What does the launch of GPT-4 mean for the future of generative AI, and what would be the reasons why workers should embrace and not fear losing jobs to AI?
Nick Eayrs (NE): GPT-4 is one of many tools available. In fact, while many organizations are impressed with the interactivity and human-like chat ability of (the recent versions of) ChatGPT, they have concerns around data ownership (and its associated risks to hand over their proprietary data to third parties), privacy, vendor lock-in, and uncertainty around regulatory compliance.
Also, AI is meant to be a productivity enhancer: not to replace people. It can liberate workers to focus on the more interesting, creative and high-value aspects of their jobs.
We believe the rapid adoption of AI by businesses will result in greater overall business and economic growth. Like with most disruptive technologies, some jobs will be made redundant, but most will be enhanced, and many other jobs that we cannot even imagine today will become mainstream in the future.
DigiconAsia: How do organizations use a combination of ChatGPT and “boring AI” to further improve their operations?
NE: Many enterprises struggle with siloed infrastructure. They use disparate data platforms and tools that are often incompatible and difficult to integrate (e.g., data warehouses for BI and data lakes for AI). With data stored and managed in silos, it is hard for organizations to unlock value from their data.
This is where the data lakehouse comes in. This is an open and unified data management architecture that combines the best of data lakes (where you store data for AI) and data warehouses (where you analyze data for business intelligence). The lakehouse enables firms to do everything from AI to BI on a single platform.
As an example, a commercial bank can use the lakehouse to power key innovations in banking, including a modernized loan application process. Historically, people would submit everything on paper for manual evaluation, which could take weeks. With a lakehouse in place, a commercial bank can now offer instant loan approvals based on predictive analytics: For high-risk customers, the bank can have a digital system in place to determine risk level and the best collection strategies for each individual; for current bank customers, they can use existing data to predict whether they qualify without the customer submitting any documents.
AI powered data storage also allows banks to offer more personalized recommendations on investment strategies and new banking products based on customer behavior and how likely they are to engage and convert.
DigiconAsia: Are there specific industries that can use “boring AI”, or can this technology be used for all business sectors?
NE: Data and AI can help organizations make quick and actionable decisions to stay ahead of the curve.
Although many people assume that data and AI are only applicable to tech-focused companies, it has proven to be imperative for all organizations across industries, from healthcare, retail, energy, and utilities to financial services.
Here are some examples:
- In healthcare and life sciences, AI is used to advance drug discovery and deliver lifesaving medical aid to remote communities around the world.
- In retail, fashion platforms can use AI to offer customers personalized recommendations and enhance operational efficiency.
- In the energy and utilities space, Australian waste management companies use AI to deliver efficient waste and recycling services daily to millions of households and facilities.
- In the financial services sector, AI can be used to deliver superior personalized financial experiences and greater customer engagement.
DigiconAsia: Can you briefly outline the trends and outlook for the wider AI and ML space in the Asia Pacific region?
NE: AI is at the top agenda for all CIOs/CDOs, and those in the region are very keen on exploring how they can build their large language models without handing over their proprietary data to a third party.
According to IDC, a surge in spending on AI systems is expected among Asia Pacific businesses, from US$17.6bn in 2022 to US$32bn in 2025, as organizations seek to enhance employee productivity, streamline decision-making, and achieve other objectives. This significant potential investment in AI signifies that businesses and organizations across diverse industries in the region will increasingly incorporate AI and machine learning into their operations and client-facing services.
In a nutshell, AI is no longer just a fad or a good-to-have technology, but it has become a norm in the region and the world. How can businesses maximize the value of AI with sound data platforms?
- Using an open and unified data management architecture that combines the best of data lakes and data warehouses to do both AI and BI on a single platform, maximizing the value of AI
- Breaking down data silos and unifying disparate data platforms and tools
- With quality data in hand, thousands of customer-centric attributes can be stored and accessed as the single source of truth by data teams through the lakehouse so that they can easily collaborate to explore customer data, insights, attributes and customer lifetime value. The data analytics platform delivers these insights at scale, democratizing data through the rapid deployment of AI across their operations.
DigiconAsia thanks Nick for sharing his technology insights with readers.
