Nuix, the Australian developer of investigative analytics and intelligent software, has announced the launch of a solution designed to transform complex unstructured data into high-quality, AI-ready information assets, Neo AI Data Curator.

The company's new offering aims to address what it describes as the most critical yet often overlooked prerequisite for successful generative AI implementation: quality data preparation.

Roland Slee, VP of Nuix Ventures, told IDM, "In dialogue with customers, it's become very clear that businesses everywhere are looking to leverage artificial intelligence and especially generative AI."

Slee explained that many organizations with early AI proof-of-concept projects have discovered that "in order to get the best from AI, you need to train large language models on corporate data," and that "the quality of outcome that you can get depends very greatly on the quality of that input."

“So, in order to get the best return on investment from AI focused projects, you need the first curate your data in order to ensure that you're training large language models on a set of information which represents the gold standard view of the enterprise. 

“That is a function which the NEO platform is able to support, and we've made some further investments in functionality and features to enable this and are now promoting our solution as Nuix Neo AI Data Curator.”

The core intellectual property of Nuix and the foundation for the company’s success has been its data engine which still sits at the heart of the NEO platform which also includes a range of other capabilities, some of which have been engineered in-house by Nuix and some the result of acquisitions.

The acquisition of Topos Labs in 2021 provided NLP technology that has been integrated into the NEO platform and is today providing Cognitive AI capability.  

Similarly, Nuix made an acquisition of a company called Rampiva in 2023 which provided a workflow engine which has been rebranded as Nuix Automate and is now part of the NEO platform.  

“In 2025, we're more looking to promote the business solutions that are available on the platform and progressively to expand that range of solutions,” said Slee.

"The way to think about Nuix today is that NEO is our platform for the processing of unstructured data and leveraging artificial intelligence," said

According to Nuix, unstructured data makes up over 90 percent of all enterprise information and is inherently difficult to manage. The company warns that even sophisticated AI systems will fail to deliver valuable outcomes if fed with inaccurate, irrelevant, or poor-quality data.

The Neo AI Data Curator builds on Nuix's 25 years of forensic-grade data analytics experience and handles the entire unstructured data processing workflow - from cleansing and normalization to vectorization, enrichment, prioritization, and redaction.

A case study highlighted in the company's fact sheet describes how a major legal sector client initially struggled with their AI implementation. The client attempted to build a document filtering system powered by a fine-tuned LLM but encountered significant challenges. After three months of unsuccessful experiments, they discovered that 90 percent of their dataset was irrelevant or detrimental to training the model.

After implementing Nuix Neo AI Data Curator, the client was able to rapidly identify data anomalies and processing errors, determine the fitness of their dataset for AI applications, and avoid time-consuming iterations of fine-tuning and data cleaning.

"Nuix's engine gives us the ability to very deeply inspect the content of unstructured data and to do that at scale with great efficiency.,” said Slee.

The solution features a five-step process that includes collection, transformation, enrichment, fortification, and promotion of data. The Nuix platform can connect to enterprise data across various storage systems, normalize over a thousand file types, and process terabytes of information in hours.

A notable capability of the Neo platform is its ability to analyse multi-layered documents. Slee provided an example where the system can process an email with a zip file attachment containing multiple documents, including embedded files, and still identify sensitive information like a driver's license number hidden several layers deep.

Nuix is positioning its Neo AI Data Curator as essential for both public and private organizations with complex or messy datasets that want their generative AI initiatives to deliver accurate, scalable, reliable, and explainable results.

As organizations continue to invest in AI technologies, the focus on data quality and preparation appears to be gaining recognition as a critical success factor, with Nuix quoting Oracle co-founder Larry Ellison: "Unless you get your data properly organized, you can't use AI. It becomes utterly useless."