kama.ai, a leader in responsible, conversational AI solutions, announces two new capabilities to launch in early 2025 as part of its Designed Experiential Intelligence® (kama DEI) platform. The new features, GenAI’s Sober Second Mind® and Knowledge Administrator, are designed to combine trusted, human-curated knowledge with generative AI responses, offering enterprises the perfect blend of accuracy, trust, and efficiency.
These advances expand kama.ai’s commitment to Responsible AI with tools that manage and deliver hybrid AI content seamlessly. Hybrid AI being the blend of large language model (LLM) technology with Knowledge Graph AI systems. With this blended Hybrid AI Agent, enterprises and AI Agent builders can now leverage kama.ai’s human-governed knowledge delivery alongside real-time LLM generative responses. This gives users the best of both technologies – dynamically creative responses from the LLM, and accurate, reliable, and brand-safe outputs from kama’s Knowledge Graph AI technology.
“Our customers demand solutions that balance accuracy and flexibility,” said Brian Ritchie, CEO and Founder of kama.ai. “These new capabilities let enterprises deliver trusted information along with the speed and creativity of generative AI. But it also lets them grow their curated and sanctioned content and processes – faster than ever before.”
GenAI’s Sober Second Mind® introduces two APIs to support Hybrid AI Agents. The kamaChat API delivers reliable, curated enterprise knowledge and alerts the Hybrid AI Agent when no pre-approved information exists. The latter allows a generative response to be provided in real-time. However, in this case the real-time generative response is governed by the LLM’s Retrieval Augmented Generation (RAG) guardrails.
A secondary Feedback API lets enterprises collect, validate, and integrate the generated response into their trusted knowledge base. This is where human-in-the-loop provides the highest level of trust. It lets the enterprise’s subject matter experts edit and approve the generated response for addition to the knowledge graph, continually enhancing the accuracy, relevance, and automation of future interactions.
Supporting these capabilities is kama.ai’s new Knowledge Administrator toolset. These tools give the enterprise human knowledge administrators greater control over RAG with its unstructured data. It integrates and leverages enterprise repositories like SharePoint, M-Files, and S3 Buckets. It allows organizations to curate specific document collections for use with kama DEI’s generative drafting capabilities. Generative drafting and human-in-the-loop review and approval ensures all responses remain accurate and grounded in enterprise-approved information.
“With these tools, enterprises can confidently deliver both trusted and approved information along with the most accurate generative responses minimizing the risks of AI hallucinations,” said Jacky Wang, Head of AI Development at kama.ai. “This combination of human oversight and AI efficiency means enterprises can optimize user expectations while minimizing risk.”
The new capabilities were developed in collaboration with kama.ai’s customers and partners. Tena Thiessen, Director of Operations at Procom, mentions, “As a customer of kama.ai’s trusted conversational knowledge delivery platform, we have been using their chatbot services to assist our internal employee Q&A and make our business rules easily searchable. We were happy to provide input into their new Knowledge Administration and Hybrid AI Agent designs. We look forward to testing these features in-house very soon to enhance our knowledge management capabilities further and improve internal communications within our company.”
For more such updates, follow us on Google News Martech News