As conversational artificial intelligence (AI) rapidly evolves, companies are racing to provide solutions that effectively bridge the gap between human and machine interactions.
openstream.ai's newly patented software aims to help businesses interact through its Enterprise Virtual Assistant (EVA) platform. The company says EVA enables the creation of AI avatars, virtual assistants, and voice agents that can have human-like conversations without the need for backend complexity, scripts, or hallucinations.
of Conversational AI The market is highly competitive, with players such as: IBM, microsoft, Google, Amazon and nuance communications We offer unique solutions. Many companies rely on generative AI (GenAI) for these interactions. prone to hallucinations Their ability to handle complex interactions is limited. However, Openstream's neurosymbolic approach combines the power of large-scale language models (LLMs) with the planning and reasoning of symbolic AI.
“Our scriptless approach to interaction management is different from other vendors in that we do not rely on designed interaction scripts. Raj Thumruri, CEO and founder of Openstream, said in an interview with PYMNTS. “Combining the ability to understand these actions, conditions, and entities in an interaction (the neural part using LLM) with the ability to reason about and plan for these (the symbolic part) gives us a very powerful ability. You can.”
More empathetic AI
Openstream recently patented an approach aimed at enhancing the human-like understanding and response of virtual agents and contributing to an improved user experience.
“By having a secondary means of expressing emotion, formal statements can be delivered with a smile, kindness, and compassion, which can help 'break the ice,'” he said. Ta.
Most companies manage conversations by looking for the purpose of what people are saying and fitting it into specific categories, Tumulli said.
“So if you say, 'I want to order a cheese pizza,' the 'dish' slot will have 'pizza' and the 'type' slot will have 'cheese,'” he said. “Then map these intents/slots to pre-scripted templated responses.”
Tumulli said brands can be perceived as more empathetic by responding with the right emotion, but responding with the right emotion requires models to know the user's current emotions and personality traits. He said it needs to be at least evaluated.
“When a user says, 'My car hit a tree,' the system recognizes that as a negative emotional state and first asks the user if everyone is okay and if anyone was hurt,” Tumulli said. . For example, the software might say: I hope no one gets hurt and you're okay,'' he explained, instead of jumping to a more measured response, “Do you want to file an insurance claim?''
Openstream is one of many companies working on conversational AI. Voice interaction is a seamless way for users to interact and even transact with her AI platform. According to PYMNTS Intelligence, Voice assistants are currently used by 86 million U.S. consumers each month. Almost one-third of U.S. millennials use voice assistants to pay their bills.
The integration of voice into daily work emphasizes its convenience. However, ease of use does not guarantee perfection. Challenges in the realm of voice AI go beyond just delivering natural conversations. Concerns about privacy, security, and ethical handling of voice data are salient issues that need to be addressed.
Keevan MohajarCEO and Co-Founder of Conversational Intelligence Platform sound houndwill be reflected in last year's interview Join PYMNTS to discuss the early days of voice AI. Users initially sought to engage in extensive, sci-fi-like interactions with their devices, but the technology was limited to basic functions such as playing music, setting timers, and updating weather information. This gap between expectations and reality has led to tempered expectations about the potential of voice AI.
Despite these hurdles, PYMNTS Intelligence shows significant interest in voice technology, with 63% of consumers saying they would use voice technology if it matched human capabilities. Additionally, 58% said they would choose voice technology over manual tasks because it was easier and more convenient, and 54% said they preferred voice technology for its speed compared to typing or touchscreens.
The future of conversational AI
Although the company doesn't predict trends as a matter of policy, Tumulli said, “We strive to achieve human-level understanding with human-like interactions.”
Openstream is focusing on several key areas to enhance the quality of conversation experiences and relationships that clients can build using their platform.
“We are constantly working to improve our AI avatars and ensure that they fully fidelity to human emotions and true personality,” Tumulli said. The company employs various techniques to avoid this problem. uncanny valley while emulating human expressiveness.
The company is also working to alleviate thorny issues such as: LLM that causes hallucinations.
“We are currently using LLM extensively, which can suppress hallucinations,” Tumulli said. “We are deploying our own model variants and fine-tuning our datasets to further enhance our ability to transform the chaotic human world into symbolic representations for planning and reasoning. .”
The company is developing tools to build interaction domains without coding. The goal is to democratize the creation of interaction domains using a neurosymbolic interaction manager.
“We want to enable any business user to design highly complex interaction domains using visual tools and incorporating existing business documents,” said Tumuluri.