SA Water eyes data lake, enterprise-wide AI platform

South Australian water supplier SA Water is poised to expand the use of artificial intelligence across its operations, revealing plans for a data lake and enterprise-wide AI platform.

The government-owned corporation has approached the market to find potential solutions to fill gaps within its data and analytics environments after undertaking a stocktake last year.

The exercise followed several years of investment in a number of smart technology initiatives, including deploying sensors in water pipes that serve more than 1.7 million people across the state.

“During 2021, SA Water undertook a program to define cohesive strategies for the management of data and outline how the organisation will leverage artificial intelligence,” SA Water said in a request for information.

The program, apart from defining a reference architecture, “highlighted where SA Water is currently lacking capability” across AI, data and storage.

While SA Water already uses a range of AI tools, including within its Microsoft Azure tenancy, they are used “specifically for given business units rather than as enterprise-platforms”.

AI tools include Databricks, Azure Cognitive Services, TensorFlow, FICO and Darknet, while visualisation platforms include PowerBI, Shiny and Grafana.

Its data landscape, meanwhile, currently “revolves around an on-premises data warehouse architecture leveraging primarily Microsoft SQL Server”.

SA Water is now looking to adopt a hybrid AI operating model with a “central core capability”, according to the RFI.

The model, which is not finalised, is also expected to align “with the data management capability”, and be “technically managed and supported through [the] SA Water technology team”.

SA Water is currently “establishing a data and AI governance body to prioritise funding and an operating mode to standardise delivery and support of data and AI solutions”.

It said that it expected to “transition from on-premises tooling to cloud services” as part of its AI strategy, but that this was “not a mandated direction”.

At the same time, it is planning to modernise and expand its storage capability by optimising its use of on-premises infrastructure and increasing its use of cloud services.