Intended audience: developers administrators
AO Platform: 4.5
AI/ML Enablement & Accuracy
What is the process for moving from raw enterprise data to AI-powered insights? How much configuration is required versus out-of-the-box capability?
All our ML recipes are fairly generic. With minimal data science knowledge, end users can configure and train ML models using the recipes on their data sets.
How do you handle model drift, data bias, or data quality issues in real-time analytics scenarios?
Our Quick Insights models are generated using the customer data and can be scheduled to be refreshed at periodic intervals. This would address both model and data drift. We use Bayesian modeling to address outliers and data bias. Since we are not delivering the models, customers are responsible for addressing any ethical or other such biases.
What level of explainability or transparency do you provide for AI-generated recommendations or predictions?
We strive to make the responses generated by the platform transparent and explainable. Users can tap into the platform's/LLM's chain of thought, pinpoint data sources, tables, and attributes used, or even inspect the SQL query generated by the platform.
Go to all FAQs.
Contact App Orchid | Disclaimer