AI helps LG Valuations raise bar on rates, land tax accuracy

Victoria-based LG Valuations has used Microsoft Azure cognitive services to build an artificial intelligence platform to analyse data and assess the value of properties which is promoting exceptional accuracy and speed.

Developed in association with Microsoft partner Advance Computing, the AI-infused property valuations platform is running alongside LG Valuation’s own proprietary valuation system and according to LG Valuations CEO Marcus Hann is helping enhance overall accuracy and ensuring valuations are free from bias.

“I think we’re quite staggered as to how quick and easy and accurate it could be,” said Hann. He said that the insights delivered by the AI platform were playing a key role in ensuring property owners were not being overcharged for rates or land tax.

LG Valuations is contracted across Tasmania and Victoria to provide statutory valuations which are used in the determination of land tax and rates. Modelling is complex and reliant on multiple data inputs.

One of the key benefits of the AI-infused data modelling platform is that it has revealed which data has most influence on the accuracy of predictions, allowing LG Valuations to focus its efforts on collecting the right data to feed into the Valorise valuations platform.

Variables such as building area and land area have great influence and LG Valuations has also been able to leverage other locality factors that people consider ahead of buying property, such as driving distance to railway stations in urban areas.

“We use the machine learning to inform us about errors with our models,” said Hann. “It is significantly increasing the confidence in the outputs we’re able to put forward for councils. We are making sure people aren’t being over charged and certainly the efficiency level it’s helping us achieve is dramatic.”

While LG Valuations has not replaced its own valuation model with the AI platform, it has compared the values predictions of each. In a trial the company found that the AI model was more accurate in 75 per cent of cases – which Hann attributed to the AI platform’s ability to interpret a vast array of data.

Chris Motton, director, Advance Computing said that Advance had worked with LG Valuations on the rewrite of its evaluation software package for Azure, before being called on last year to help the company build the AI infused valuation platform.

According to Motton; “Because the AI infused solution is now part of Valorise itself there are plans to further improve, adopt and rely on these models and functionality going forward. It’s already amplifying and improving their decisions.

“It’s also showcasing the talent that we have in regional Australia. This is a great example of using AI in local government, and the outcomes are fairer ultimately for the rate payer and I think everyone would appreciate that.”

The solution leverages Azure Machine Learning Service, Logic Apps and Application Insights. It took a week to implement and further insights, training, learning and improvements have occurred over the past six months.

Lee Hickin, CTO, Microsoft Australia, said; “The progress that LG Valuations has made with artificial intelligence is impressive. It has been able to identify which data sources are most important for accurate valuations, reduce the risk of property owners being overcharged for their rates or land tax, and enhance its own productivity.

“AI is no longer the province of large enterprise alone – it is having significant impact in organisations of all sizes and in all sectors right across Australia. Working with Microsoft partners such as Advance Computing, we are seeing more and more businesses like LG Valuations be able to extract genuine value from their data using AI infused technologies.”

While LG Valuations continues to run both its own Valorise platform and the AI solution Hann said; “It’s producing such incredibly good statistical outcomes, it’s certainly far and above anything that I’ve seen of any other automatic valuation models.”