kama.ai announces the introduction of two key industry-leading features to its emotional, conversational intelligence platform, kama DEI (Designed Experiential Intelligence®).
The Canadian Indigenous-owned conversational intelligence provider kama.ai recently presented at the Toronto Collision Conference on the importance of taking an Indigenous approach to AI using traditional teachings and values like Honesty, Respect, and Humility to guide the development of responsible AI.
Today, kama.ai announces the introduction of two key features that were demonstrated to an industry-leading technology research firm on August 17th. Coming later this year in Version 3, these new features continue kama’s progression of using conversational intelligence responsibly in enterprise applications. The two new capabilities of Draft Assistant and Advanced Inference are briefly described below.
Draft Assistant:
Generative AI using Large Language Models (LLMs) represent some of the latest advancements in AI, but they can’t always be trusted to stay ‘on brand’ and have been known to hallucinate responses with untruthful and biased information. kama’s new Retrieval Augmented Generation (RAG) Draft Assistant feature uses real-time search directed at targets like the open Internet, a specific URL, or even an enterprise repository to find the most relevant documents or text. It then auto-generates a draft to the enterprises’ kama DEI Knowledge Managers for review and editing to ensure responsible, authorized use after appropriate enterprise approval processes. The approved response and related information are then stored securely in the enterprise knowledge graph for accurate, unchanging, energy efficient, and immediate responses to user inquiries prioritized by kama DEI’s patented human value guidance capability.
Advanced Inference with Intelligent Conversational Triage:
Building upon kama’s patented human value based problem solving and prioritized information delivery attuned to the users’ values, kama DEI Version 3 will include Advanced Inference. “We can now address what the industry calls ‘Intentless Inquires’. These are cases like ‘I want to buy a house’ or ‘We just had a baby’ wherein there is no direct problem and therefore no ‘solutions’ or related information we can pass on,” Kaiyang Li, kama.ai’s Head of Product says. “With Advanced Inference, we are now able to understand when we have encountered one of these intentless inquiries and traverse our knowledge graph, prioritized by human values, to have a conversation with the user about what potential issues we can assist with. Examples of this with the report of a pedestrian auto collision, for example, could be ‘Has there been an injury?’, or ‘Is the driver of the car still on scene?’. This type of intelligent interactive triage in user conversations is something that Large Language Models cannot deliver at this time.”
Brian Ritchie, founder and CEO of kama.ai and a member of Chapleau Cree First Nation says, “We want to avail the use of the latest technologies in search and auto-generation, but we needed to employ these capabilities in ways that keep the enterprise safe. Our Human(s)-in-the-Loop knowledge curation process ensured responsible enterprise use previously, but it lacked the ability to provide good research and drafting tools for our enterprise client’s kama DEI Knowledge Managers and subject matter experts. With Draft Assistant, you get the best of both worlds: industry-leading Retrieval Augmented Generation combined with human-led enterprise governance for authorization and approval to mitigate any rights, bias, or authenticity issues.”
With its new Advanced Inference and conversational triage of unsolvable inquiries, kama is truly leading the industry. Ritchie commented “Using our knowledge graph and human value rating mechanism, we can detect particular contextual situations and then prioritize the underlying client concerns from a general statement like ‘We may be heading for a divorce’.”
Current LLMs may provide a single response with many possible ‘solutions’, but that is not the way humans interact says Ritchie “Our Advanced Inference feature uses our knowledge graph as a ‘human behavioural model’ to determine possible underlying situations, validating these possibilities in human value priority conversationally, with empathetic responses and helpful, supportive information offerings along the way. This is something that is just not available in the industry right now and we are very pleased to be introducing it.”.
The features are available for demonstration to key customers and prospects immediately and will be introduced to customers for general market use in kama DEI Version 3 later this fall.
For more such updates, follow us on Google News Martech News