- Insituate's Newsletter
- Posts
- Product Update #1
Product Update #1
Evaluate your copilot; Monitor Costs & performance;
Hello,
We’re excited to share our very first product update with you! Thank you for subscribing to the Insituate’s newsletter. You’re now part of a community that’s shaping the future of AI and innovation.
Product/ Media Updates
We're excited to announce significant enhancements to our Retrieval-Augmented Generation (RAG) evaluation capabilities on the Insituate platform. These updates will empower our clients to build more robust and effective AI solutions.
Here are the key features:
Cost Analytics Dashboard - Presenting the analytics dashboard of Insituate. Monitor your APIs in testing and production for credit usage and set alarms.
The Analytics Dashboard
• Comprehensive RAG Benchmark Suite - We've integrated a diverse set of RAG-specific benchmarks, including RAGAS, ARES, and RGB, to provide a holistic evaluation of your RAG systems. This suite covers critical aspects such as context relevance, answer faithfulness and information integration robustness.
Insituate’s Copilot Studio RAG Evals - Evaluate copilot responses on predetermined benchmarks
• Advanced Metrics for Retrieval and Generation - Our platform now supports sophisticated evaluation metrics for both retrieval and generation components. This includes MAP@K and MRR@K for retrieval accuracy, as well as ROUGE and BERTScore for measuring the quality of generated responses.
• LLM-as-Judge Evaluation - Leveraging the power of large language models, we've implemented an LLM-as-Judge feature for more nuanced and context-aware evaluations of RAG outputs. This allows for a more human-like assessment of response quality and relevance.
• Multi-faceted Performance Analysis - Our new evaluation framework examines various aspects of RAG performance, including relevance, faithfulness, correctness, and additional requirements like latency and diversity. This ensures a comprehensive understanding of your RAG system's capabilities.
• Customizable Evaluation Datasets - We now offer tools to create tailored evaluation datasets that better reflect real-world scenarios and domain-specific challenges, allowing for more accurate and relevant performance assessments. These updates significantly enhance Insituate's capabilities in RAG system evaluation, providing our clients with deeper insights and more reliable performance metrics. Stay tuned for more exciting features coming soon
Spotlights
We were honored to have Mr. Alok Tiwari, the CEO of Aptivaa & Cognext attend our Webinar and enlighten the audience about AI adoption in Finance.
His truly eye-opening insights helped us understand the future of AI in the financial world. We're excited to apply what we've learned and push forward with innovative ideas!
Technology
Check out our blog on DSPy
The DSP framework combines language and retrieval models to enable efficient in-context learning, automating complex, multi-step NLP tasks.
Need Help?
Reach out [email protected]
Get started here, or if you’d like to bring these features to your enterprise please get in touch.
— The Insituate Team