IBM watsonx.governance as a Service
IBM SoftwareReviews from AWS customer
0 AWS reviews
-
5 star0
-
4 star0
-
3 star0
-
2 star0
-
1 star0
External reviews
44 reviews
from
External reviews are not included in the AWS star rating for the product.
Watsonx Governance
What do you like best about the product?
I like how watsonx governance combines model risk, generative ai, and governance. For example, the ability to create a use case, track a model, associate with workspaces, and evaluate prompts all in a single area are great. Even better, but IBM recently launched a new feature called the Governance Console where you can track multiple models, enable and create controls, and evaluate risks.
Governance had Gen AI metrics like ROUGE and BLEU, as well as the ability for classification like F1, Accuracy, precision, recall, etc
Governance had Gen AI metrics like ROUGE and BLEU, as well as the ability for classification like F1, Accuracy, precision, recall, etc
What do you dislike about the product?
I don't like how an evaluation in Watsonx Governance overwrites the previous evaluation results if done on the same day. OpenScale will keep the history, but governance in the AI Factsheet does not. Same issue exists for the trend charts.
For example, the trend charts available in OpenScale have a time setting for hourly, daily, weekly, etc. But let's say i do 20 evaluations in a day. Then a week later I want so show how those 20 evaluations did, and the overall trend. If I change to hourly, then the graph is too sparse (it's been a week), but if I change to daily, then it's just a single dot (all 20 evals were on a same day). Wish OpenScale and Governance would fix this.
Then the ability to show the generated_text against the reference (ground truth) text. Yes, you can download the results to a CSV but there's two big missing features (1) the downloaded CSV is missing the reference text, it only has the generated text. So you have to come up with a way to match that record to its original source if you want to compare, and (2) you have to do this each time for each subset. Why not just have the ability to view that record in the software and avoid having to download the CSV? the text is truncated and there's no way to see it
For example, the trend charts available in OpenScale have a time setting for hourly, daily, weekly, etc. But let's say i do 20 evaluations in a day. Then a week later I want so show how those 20 evaluations did, and the overall trend. If I change to hourly, then the graph is too sparse (it's been a week), but if I change to daily, then it's just a single dot (all 20 evals were on a same day). Wish OpenScale and Governance would fix this.
Then the ability to show the generated_text against the reference (ground truth) text. Yes, you can download the results to a CSV but there's two big missing features (1) the downloaded CSV is missing the reference text, it only has the generated text. So you have to come up with a way to match that record to its original source if you want to compare, and (2) you have to do this each time for each subset. Why not just have the ability to view that record in the software and avoid having to download the CSV? the text is truncated and there's no way to see it
What problems is the product solving and how is that benefiting you?
Model Governance and Monitoring
Governance product to face the EU AI Act
What do you like best about the product?
The product had matured greatly over the past year and offers access to a wide range of personas. This is extremely valuable since preparing the documentation for the EU AI Act is a team effort.
It is well integrated with Watsonx AI.
It is well integrated with Watsonx AI.
What do you dislike about the product?
Pricing tag.
Integration with Git can be improved.
Integration with Git can be improved.
What problems is the product solving and how is that benefiting you?
Monitoring AI models at different stages (from dev, test, production) and support for the EU AI Act requirements
Review for IBM Watsonex.governance
What do you like best about the product?
Much needed framework for governing models applications and data workflows.
Responsible AI Framework
Automated Compliance
Responsible AI Framework
Automated Compliance
What do you dislike about the product?
Not much as of now. New to me .Cost can be reduced if feasible.
What problems is the product solving and how is that benefiting you?
Customized dashboards and model factsheet,simplifies documentation.
The tool is pretty solid for managing AI models and ensuring compliance,
What do you like best about the product?
It makes AI models more transparent and easier to explain to others
What do you dislike about the product?
Some features feel a bit complex to set up and need more guidance
What problems is the product solving and how is that benefiting you?
It helps ensure our AI models are compliant, explainable, and less biased, which makes it easier to build trust and meet regulatory requirements
Reliable platform for AI governance
What do you like best about the product?
the ability to manage and monitor models across their lifecycle in a structured way. The explainability and transparency features are especially useful for building trust with stakeholders, since you can clearly see how decisions are being made. The automated governance workflows save time on documentation and compliance checks, making audits much smoother and also the dashboard seems good with so many features.
What do you dislike about the product?
The initial setup can feel complex, especially for teams that are not already using IBM’s ecosystem. There’s a learning curve in configuring policies and integrating all models into the platform.
What problems is the product solving and how is that benefiting you?
It is helping us address the challenge of managing AI models at scale while staying compliant with internal policies and external regulations. Before, it was difficult to track model performance, bias, and drift across different teams, which slowed down decision-making and created audit risks. With Watsonx.governance, we now have a centralized system to monitor all models, generate explainability reports, and automate compliance documentation
Good tool for governance, bit room to improve - IBM watsonx.governance
What do you like best about the product?
It makes it easier to keep track of AI model risks and compliance without spending hours in spreadsheets. The dashboards are simple and I can see issues quick.
What do you dislike about the product?
Some parts feel a bit slow to load and integrations with other tools could be smoother. Takes little time to get used to navigation.
What problems is the product solving and how is that benefiting you?
It helps us keep AI models in check with compliance rules and track risks before they become big issues. Saves time and gives more confidence using AI in projects.
featuring high ratings, multiple categories, and numerous award-winning products.
What do you like best about the product?
Coverage across 343+ categories (AI, analytics, project management, IT services)
What do you dislike about the product?
mall and mid-sized companies sometimes find IBM products too “heavy” or costly compared to SaaS-native solutions
What problems is the product solving and how is that benefiting you?
It helps organizations detect and mitigate AI-related risks such as unfair bias, model drift, and erratic behavior—enhancing model performance and trustworthiness.
Made Automated Workflows Easy!
What do you like best about the product?
I like the automated workflows that can also be customizable.
What do you dislike about the product?
Not anything for now. I like every feature.
What problems is the product solving and how is that benefiting you?
It helps me to manage my AI Generative Model throughtout development to functioning.
Powerful AI Governance, but Complex to Master
What do you like best about the product?
IBM watsonx.governance helps companies manage AI risk, bias, and compliance (EU AI Act, NIST, etc.). It offers audit trails, bias/drift monitoring, and automated compliance workflows, and integrates with AWS, Azure, Google Vertex, SageMaker.
What do you dislike about the product?
Complex setup, steep learning curve, some users report latency and limited connectors
What problems is the product solving and how is that benefiting you?
Keeps our AI models compliant with regulations (EU AI Act, NIST, internal audits).
Detects bias, drift, and other risks before models go live.
Creates audit trails so we can prove transparency and accountability.
Automates documentation and compliance checks that were manual before.
Detects bias, drift, and other risks before models go live.
Creates audit trails so we can prove transparency and accountability.
Automates documentation and compliance checks that were manual before.
WatsonX Governance
What do you like best about the product?
The tool's proposal is good. The user experience as well. The content it aims to present.
What do you dislike about the product?
Difficult to use. Limited in connections. Limited in implementation methods. The SDK is complex to use. A lot of latency and connection breaks with OpenScale during tool usage. Not very debuggable, difficult to find out what's going wrong when an error occurs.
What problems is the product solving and how is that benefiting you?
Monitoring of models.
* Training and release/update cycles of the endpoint.
* Model training metrics.
* Number of models in use.
* Training and release/update cycles of the endpoint.
* Model training metrics.
* Number of models in use.
showing 1 - 10