Microsoft has acquired an artificial-intelligence startup, Semantic Machines, to bolster its efforts in “conversational AI” and potentially make its Cortana virtual assistant better at understanding natural language enquiries.

Microsoft didn’t disclose how much it paid for Semantic Machines but said it would open a new AI research center in Semantic Machines’ hometown of Berkeley, California.

Semantic Machines contains numerous AI veterans, and the acquisition gives Microsoft access to some formidable talent. Semantic’s chief technology officer, Larry Gillick, is a former chief speech scientist for Apple’s Siri. Microsoft also name checked the University of California at Berkeley professor Dan Klein and the Stanford University professor Percy Liang.

Several Semantic staffers also worked at Nuance, the voice-recognition company that once powered Siri. Semantic’s chief executive, Dan Roth, is a cofounder of Voice Signal Technologies, which Nuance acquired in 2007.

Microsoft says most AI doesn’t really understand much human communication, something anyone who has put anything more than the simplest question to Siri will already know. “Most of today’s bots and intelligent assistants respond to simple commands and queries, such as giving a weather report, playing a song or sharing a reminder, but aren’t able to understand meaning or carry on conversations,” Microsoft wrote in a company blog announcing the acquisition.

Semantic Machines’ team and technology could fix that by helping digital assistants like Cortana have a “natural dialogue.”

“We call this ‘conversational AI,'” Microsoft added.

Microsoft has been experimenting for some time in this space. Aside from Cortana, the company has dabbled in chatbots, though not always with great success. Its Chinese chatbot Xiaoice has millions of loyal followers who talk to it. But earlier chatbot experiments, such as Tay and Zo, spouted racist and anti-Microsoft messaging and had to be shut down.