Many organizations have complex business processes, extensive documentation, and information distributed across disparate domains. As a result, customers—including the public, a specific customer group, or internal customers—have a hard time finding the information they need despite it already existing.
Cornflower Chat solves the issue of hard-to-navigate information: it is an AI-powered expert for your business processes, tailored to your specific use case and under final compliance review for use in the State of Colorado pursuant to the State’s Statewide GenAI Policy. Cornflower Chat automatically ingests a list of SME-defined documents, such as webpages, Google Docs, and PDFs. The system then learns from this source information, custom FAQ documents, or other user-provided input. Then, Chat uses state-of-the-art LLMs to respond to business-related questions using text-based question and answer sessions. Tested guardrails are built into the system to ensure that chatbot output is appropriate and context-specific.
Cornflower Chat follows the State OIT Generative AI policy, including Generative AI Testing Guidance and risk- reduction procedures. The tool is tested on metrics including bias, accuracy, completeness, and correctness.
For more information about Cornflower Chat, Cornflower Modules, or Cornflower Labs products or services, contact info@cornflowerlabs.com. When you buy Cornflower Chat or Cornflower Modules, you’ll work directly with Cornflower Labs principals who understand the unique Colorado technical and compliance requirements. We take a multi-phase needfinding approach, involving SMEs throughout the engagement to ensure quality.
Cornflower Labs develops various products for public sector customers, including AI systems deemed high-risk in the State of Colorado, such as Cornflower Chat. Cornflower Labs manages risks associated with quality and bias, including algorithmic discrimination through a multi-pronged approach that includes: working with SMEs throughout the project lifecycle to understand the business case, engaging in automated testing to protect and monitor against risks of discrimination, and using the latest guardrail and other intervention technology to prevent discriminatory content, among other safeguards.