TLDR: Fisent Technologies has introduced a new confidence rating capability for its BizAI solution, designed to improve generative AI process automation. This feature provides businesses with deeper insights into AI-driven decisions, helping to identify outcomes that may require human review and minimize costly errors.
TORONTO – Fisent Technologies, a leader in Applied Generative AI (GenAI) Process Automation, announced on August 12, 2025, the integration of a new confidence rating capability into its Fisent BizAI solution. This enhancement aims to provide businesses with critical insights into the reliability of GenAI-driven automated processes, particularly those involving repetitive decision-making and review tasks.
Unlike conventional GenAI confidence metrics that typically focus on internal model certainty, Fisent’s innovative approach employs multiple techniques to assess confidence, prioritizing them based on specific customer use cases. A key methodology involves multi-model similarity: if multiple Large Language Models (LLMs) concur with the base model’s outcome, the confidence rating is high. Conversely, a lack of consensus among models results in a low confidence rating, signaling the need for additional measures, such as human intervention, to ensure accuracy and completeness.
This new capability is meticulously tuned to the specific process being automated, with the goal of minimizing the risk of costly false negatives without creating a productivity bottleneck from excessive false positives.
Adrian Murray, Founder and CEO of Fisent, highlighted the significance of this development, stating, “GenAI models generally lack a mechanism to indicate their own reliability. Our confidence rating capability solves this by providing a clear, intuitive way for businesses to identify what decisions should be triaged for human intervention. It’s about evaluating GenAI outcomes with the same rigor you’d apply to human work, but at a speed and scale not possible with traditional methods.”
Key features of the Fisent BizAI confidence rating capability include:
Multiple Confidence Assessment Methodologies: Different techniques are leveraged to assess confidence, tailored to a customer’s use case and priorities.
Multi-model Similarity: Compares the base model’s outcome against responses from multiple peer or superior LLMs to calculate a confidence rating, which is predictive of confidence, not just accuracy.
Predictive Confidence for Human Review: Helps users determine if a GenAI outcome requires human intervention, effectively identifying items in an automated workflow that need additional scrutiny.
Dynamic Weighting: Allows the confidence rating to be sensitized based on the criticality of the task.
Fisent’s new confidence rating capability for BizAI is now available and has demonstrated particular value in complex document processing scenarios, such as information identification and analysis within mortgage contracts, insurance claims, and legal documents, where the financial and operational impact of errors is substantial.
Also Read:
- AiSentr Secures Strategic Investment from 5I Ventures to Expand No-Code AI Solutions for Businesses
- Egnyte Unveils AI Agent Builder for Enhanced Productivity and Secure Automation
This advancement underscores Fisent’s commitment to enabling enterprises to harness the transformative power of Generative AI by ensuring trust and efficiency in AI-powered processes.
