“Companies are looking for the best use cases for GenAI, and there is a lot of experimentation at play right now,” Kristen Hanich, director of research at Parks Associates, a market research and consulting company specializing in consumer technology products, in Dallas, told TechNewsWorld.
She pointed out that one of the main challenges companies face is related to data structure and cleanliness, which are immensely important for the reliability and validity of general AI. Another key challenge is that certain use cases people might assume are low-hanging fruit for GenAI, like lease abstraction, may not be in practice, and that hallucinations can cause operational and legal issues, she added.
“Embedding GenAI to specific workflows has a lot of potential for the right use cases, but it does take a specific approach to designing systems — virtualized workflows that are well-mapped and well understood, carefully trained models, and such — to create the reliability and consistency that companies need,” Hanich said.
“For those using public AI models, there is also the risk that data may be leaked,” she added. “We have seen companies get around this by leveraging private models instead.”
From the article, "Corporate Real Estate AI Pilots Surge, ROI Still Elusive: Report" by John P. Mello Jr.
