---
title: Monitor generative AI with custom metrics
description: Monitor LLMs and generative AI solutions to measure alignment, return on investment, provide guardrails.

---

# Monitor generative AI with custom metrics {: #monitor-generative-ai-with-custom-metrics }

[Access this AI accelerator on GitHub <span style="vertical-align: sub">:material-arrow-right-circle:{.lg }</span>](https://github.com/datarobot-community/ai-accelerators/blob/main/generative_ai/Using%20Custom%20Metrics%20to%20effectively%20monitor%20Generative%20AI/Using%20Custom%20Metrics%20to%20effectively%20monitor%20Generative%20AI.ipynb){ .md-button }

While it gets easier to build generative AI solutions with each passing day, it is becoming evident that it is critical to effectively monitor these solutions to measure alignment and ROI, and to provide guardrails. Monitoring generative AI solutions or [LLMOps](https://www.pluralsight.com/resources/blog/data/what-is-llmops) is a multi-faceted endeavor and requires more than simple out-of-the-box metrics. Each business is unique and the solutions they build require customized monitoring metrics. This accelerator aims to illustrate how businesses can use DataRobot to effectively and holistically monitor generative AI solutions, using the metrics relevant to them.

This accelerator provides metrics that segment into themes to cover the monitoring requirements of a majority of businesses.
