Rewired, 2nd Edition
by Eric Lamarre, Kate Smaje, Robert Levin, Alex Singla, Alexander Sukharevsky
SECTION FIVEEmbedding Data Everywhere
What it takes to make data easy to consume across the organization
In collaboration with Antonio Castro and Kayvaun Rowshankish1
In 2018, it took Asian bank DBS 15 to 18 months to develop and deploy AI models.
Data was stored in different areas across the business, which meant it took months just to locate the data and get access permissions. Even when teams did get access, discovering how the data was structured and assessing its quality turned into a project in and of itself. Eventually teams managed to build one-off data pipelines that couldn’t easily be reused by other teams.
That just wouldn’t do! So DBS embarked on a journey to develop a unified data platform, automate data access controls, and make data easy to discover with metadata and easy to consume through data products (data assets, in their words). And they aligned their data governance around the leadership of their different business domains.
And it paid off. By 2023, it took just two to three months to deploy AI models.2
Data is the single most important asset you have. No data, no value. But for most companies, data is more of a problem than an advantage.
In our experience, as much as 70% of the development efforts of an AI-based solution are composed of wrangling and harmonizing data. The journey always starts with figuring out which parts of your data estate most ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access