I like that Monte Carlo helps make sure downstream consumers see updated, correct data and promptly alerts us when that isn’t the case. Its direct connection to Looker is quite helpful for knowing what feeds which dashboards, and I like the Slack alert format. Review collected by and hosted on G2.com.

I find Monte Carlo’s UI not immediately intuitive, and it’s a bit overwhelming at first glance. It doesn’t take very long to get accustomed to it, though. Review collected by and hosted on G2.com.

I like that Monte Carlo is a very hands-off platform. You can set up everything at the beginning and it can basically run itself, which solves a lot of the headaches of figuring out different thresholds for anomalies. It is smart enough to set up those rules itself. Additionally, it connects well with other services, like Fivetran and Snowflake, allowing my data to live where it does. Review collected by and hosted on G2.com.

I think the setup can be a little involved, and it's a lot of connections you have to make. One of my team members took a while to set it up. Review collected by and hosted on G2.com.

Monte Carlo has a great support team that has been willing to help us with questions and improvement requests. Their product has easy to use out-of-the-box tools like machine learning thresholds that we've found to be helpful as well. We are also experimenting with their agent observability tools which allow you to have better insight into what is really happening in agentic workflows. Review collected by and hosted on G2.com.

My biggest frustration with Monte Carlo is that there isn’t a coding wrapper (Python) I can use. Right now it’s limited to out-of-the-box functionality or SQL, so it’s difficult to implement more in-depth checks. Review collected by and hosted on G2.com.

I love how quickly Monte Carlo is adapting to a changing data world, especially with the rise of AI. They've worked closely with us to set up alerts on Snowflake agents, and I appreciate that they open up office hours for collaboration with their agent experts. It's a centralized location for monitoring our data and notifies us immediately if there's an issue. I really like the integration features, particularly with Slack. Monte Carlo also enables us to get a widescreen picture of how our agents are performing, highlighting areas for improvement. Their support team is easy to reach and quick to respond. Review collected by and hosted on G2.com.

I would love if you could tune models from Slack. It would be great if when you receive the alert you could open and tune the model from inside Slack instead of having to open up the Monte Carlo UI. There were some bumps getting access set up correctly. The error handling is a bit of a black box. You cannot get details on what is happening and why it's not working. Review collected by and hosted on G2.com.

I like Monte Carlo because it's a very complete tool. It provides everything in the platform from data quality alerts, volume, and schema monitoring, all the way to offering a summary of all the alerts within our tables. Additionally, you can set up alerts in different ways including automatic ones with volume freshness and schema monitoring. Plus, it's quite useful to be able to set up personalized alerts. Review collected by and hosted on G2.com.

There are two main issues I have with Monte Carlo. First, is the communication. Monte Carlo does a lot of changes, but we're not always aware of them. This can lead to us doing some work and then having to rework it because there has been a migration or changes in the project that we weren't aware of. The second issue is with their ML monitoring. They set thresholds for alerts based on machine learning, but it's not adjusting well. I can classify alerts as expected, but it doesn't adjust the threshold as much as needed, leading to a lot of false errors. I'm wondering about the point of having those ML thresholds. Review collected by and hosted on G2.com.

I use Monte Carlo for work to ensure our tables are correct and accurate. It helps us validate our tables/data in a timely manner automatically. The benefit is we save time, and it's easy to see an alert. I like that it is easy to use and navigate even for beginners. As someone who has not used a tool like Monte Carlo before and was running notebooks, now Monte Carlo really helps. Monte Carlo is the place where all our monitors sit, and we do not need to look anywhere else. We get timely alerts. The initial setup was pretty easy, just wait time for tables to be loaded. Review collected by and hosted on G2.com.

I think sometimes there are issues with the correct number of rows returned. Monte Carlo gets it wrong sometimes. Not very sure, we have also raised this with the Monte Carlo team. But sometimes it does not populate all alerted rows. Review collected by and hosted on G2.com.

I like the lineage feature in Monte Carlo because it allows me to track data back to its source and see where it's being fed into and from. This feature almost gives me a flow diagram of where data is going, making it easier to isolate various types of data flows. I also appreciate the nice, cushy UI that Monte Carlo offers, which helps me see what tables are feeding into each other or what came beforehand. Review collected by and hosted on G2.com.

I'm often dealing with alert fatigue due to false alarms with the SQL monitors in Monte Carlo. I'm constantly checking on things that either self-resolve or don't need input, which is a bit of a hassle. It's mostly about configuring and tweaking the monitors to reduce the number of unnecessary alerts. Review collected by and hosted on G2.com.

What I like most is how Monte Carlo shifts data quality from reactive debugging to proactive observability. Instead of waiting for broken dashboards or stakeholder complaints, we can detect anomalies at the data level early - especially on partitioned, time-series datasets where issues surface quickly.

A few things stand out in practice:

The UI is nice and clear.

Pricing is reasonable, but the difference between the Scale and Enterprise tiers isn't clear.

The Monte Carlo support is quick and helpful. Most of the issues were solved within a few hours.

Anomaly detection that actually works in real workflows. The ability to monitor freshness, volume, and distribution changes across domains (Product, Finance, Business) helps us catch issues before they propagate into decision-making.

Scalability via monitors-as-code. Integrating Monte Carlo into CI/CD (GitHub Actions, domain-specific repos) makes data quality reproducible, reviewable, and scalable across teams - not dependent on manual setup in UI.

Cross-domain visibility. In a setup like ours (Trino + S3/Glue + ClickHouse), having a single place to surface incidents across domains is critical. It creates a shared language between Data Office and other teams. However, our tech stack is unusual for the platform.

Clear ownership model enablement. Monte Carlo supports the model we aim for: domain teams own their data quality, while a central team provides governance, standards, and observability. Alerts become actionable because they can be routed to the right owners.

Fast incident investigation. Even without perfect lineage everywhere, the context MC provides (upstream/downstream signals, history) significantly reduces time to understand “what broke and when.”

Pragmatic flexibility. It works across different stages of maturity - from quick anomaly detection to more structured, SLA-driven data quality processes.

If I had to summarize in one line:

Monte Carlo is most valuable not as a tool, but as the backbone for building a scalable data quality operating model. Review collected by and hosted on G2.com.

What I dislike most is that while Monte Carlo is strong as an observability layer, it still requires quite a bit of surrounding infrastructure and process to make it truly effective at scale.

In our setup, lineage is not equally mature across all engines. For example, with Trino and ClickHouse the visibility is limited compared to Databricks and Snowflake, which makes root cause analysis less reliable and often requires manual investigation.

There is also some integration friction. Certain systems require workarounds, like older ClickHouse versions or limitations in Tableau access control, which adds operational overhead and slows down adoption across domains.

Trino support isn't native. We use Starburst to make an integration.

Ownership and alert routing are not fully solved within the platform. Alerts are generated well, but assigning clear responsibility and ensuring follow-up still depends on external processes and team structure. Stronger built-in ownership and escalation mechanisms would help. Also, we use YouTrack as a bug-tracking system, which makes things less trackable for us.

From a usability perspective, managing a large number of monitors becomes harder over time. It is not always easy to understand monitoring coverage or to manage monitors at scale without relying on external tooling or code-based workflows.

The pricing model is another challenge. The credit-based approach can be difficult to predict and plan for, especially when scaling across multiple domains and teams. It requires continuous optimization and careful usage tracking.

Finally, while anomaly detection is strong, the higher-level intelligence is still evolving. It would be valuable to have more actionable insights, such as clearer grouping of incidents or better support for identifying likely root causes.

Overall, Monte Carlo is very good at detecting that something is wrong, but scaling the operational side of data quality still requires additional effort outside the platform. Review collected by and hosted on G2.com.

I really appreciate how easy Monte Carlo is to use, which makes identifying what's wrong with the data straightforward. I like that it provides a quick way to configure default anomaly detection on data assets at scale. The initial setup was very easy, and we were able to start monitoring about 80% of our assets right away. It's also great that Monte Carlo integrates with tools like Looker and PagerDuty. Review collected by and hosted on G2.com.

The way monitors are defined and changed (the migration that happened recently) is a bit confusing. The distinction between built-in monitor and custom ones was a bit difficult to understand for some consumers. Also, the 'forced' training of data for anomaly detection is tricky, as a lot of users ask how to better train the data that Monte Carlo has to tweak the detection. There should be a way to configure the thresholds before the actual datasets get trained properly. Review collected by and hosted on G2.com.

the integration with Slack is very useful in monitoring our services, likewise the addition of AI agent is very helpful for problem solving and troubleshooting issues. Review collected by and hosted on G2.com.

the threshold setting options are somehow limited and do not always allow us to capture our use cases Review collected by and hosted on G2.com.