To Get AI Right, Get the Data Right
There is no hotter topic than AI. In the CIO MarketPulse research survey, 96% of the participating business and IT decision-makers said their organization is using, testing, or planning to investigate AI technology.
“Data scientists play a critical role, but they can’t do anything without trustworthy data,” says Jackson. Generating clean, accurate, and timely data and subjecting it to AI algorithms that elicit actionable insights is fundamental to a successful modern business, she says.
So, what’s the most advantageous path to a successful AI implementation?
“Within an overall data governance strategy, implement enterprise content management,” says Healey. The idea, he explains, is to create a “single source of truth” from your data that can be used by multiple AI-powered apps.
Creating data that will produce good results with AI means finding, enriching, and tagging it with metadata. The data can then be subjected to ML algorithms with OpenText™ Aviator Analytics (IDOL), an AI platform for text analysis and data discovery.
A great use case is litigation, in which documents must be searched for specific words and references in response to legal discovery demands. “Aviator Analytics can narrow the search down to just a few documents,” says Jackson.
Litigation is only one use case. The proliferation of IoT data is opening new horizons for AI and ML. As IoT sensors have multiplied, raw processing power has increased, enabling manufacturers and healthcare providers alike to perform real-time analytics on IoT data from machines.
In healthcare, sensors on MRI machines monitor performance and relay information about the machines’ health to a management console. Running ML algorithms against sensor data means that organizations can predict when equipment maintenance is needed. This way preventive work can be scheduled to ensure that patients receive uninterrupted high-quality care.
Another example: Jaguar TCS Racing is feeding IoT sensor data from its electric racing vehicles to OpenText for analysis. “By leveraging data from lots of sensors, OpenText enables Jaguar TCS to take in information during a race, analyze it, and make adjustments so the cars can run at peak efficiency,” says Healey. And lest you think electric race cars are a one-off use case, he adds, “This might involve IoT, but it applies to every industry. Everybody should be treating their business as a high-performing vehicle.”
