I really enjoy the image and video models available in Vertex AI, which I think are the best and have no comparison due to their attention to detail. The platform helps automate daily tasks, like email sorting, sending, and replying in my tone, which is a significant convenience. I love how it generates faceless videos with great attention to detail. Using it with GCP cloud architecture is a game-changer for projects, especially when working with custom models and productionizing them effortlessly through the MLOps Pipeline. The collaboration with multiple tools to scale and productionize applications is quite unique, and integrating Vertex AI with tools like BigQuery enhances its power. Review collected by and hosted on G2.com.

I usually face issues with the pricing, and at many places, it feels overpriced or costly compared to competitors like Bedrock. I would suggest making billing easier to follow, so it's clear where the costs are coming from. It's currently calculated on multiple dimensions, which makes it super complex. Review collected by and hosted on G2.com.

I use Vertex AI to streamline and scale my machine learning projects efficiently. It simplifies the entire ML workflow by allowing me to manage everything from data preparation to model deployment on a single platform. I appreciate features like AutoML, managed training, and seamless integration with Google Cloud services, which make experimenting and scaling models much easier. Monitoring and versioning capabilities let me track model performance and improvements over time. I value how Vertex AI makes working on complex AI projects more organized, efficient, and reliable. The initial setup was really smooth, allowing me to dive in and start experimenting quickly. Review collected by and hosted on G2.com.

While I really like using Vertex AI, a few things could be better. The pricing can be a bit confusing, especially when running bigger experiments, so it’s hard to predict costs sometimes. Some of the advanced features also take time to learn, and I wished there were more clear, practical examples for beginners like me. While it works really well with Google Cloud, connecting it to other tools I use can feel a bit tricky and require extra effort. If the pricing were clearer, the tutorials more beginner-friendly, and integrations smoother, it would make the whole experience even better. Review collected by and hosted on G2.com.

I like how Vertex AI automatically trains the best model for me after I upload data. I enjoy that it helps me build chatbots using Vertex AI agent builders and trains models without needing to code, thanks to AutoML. It's convenient to use Vertex AI as a single interface, like a playstore for models, with Google models, open source models, and third-party options. I appreciate the ability to call different models like Gemini, Llama, or Claude using the same code, which saves me development time. It's also helpful that I receive just one bill for all models, avoiding the hassle of managing multiple vendor subscriptions. The initial setup was very easy, with model parameters being simple to configure. Review collected by and hosted on G2.com.

I don't like the newest features almost always launch in US regions first, so you may have to accept higher latency or wait for them to become available in the Mumbai region Review collected by and hosted on G2.com.

I appreciate Vertex AI for providing Colab Enterprise Notebooks that are easy to use and can be shared with my coworkers effortlessly. The feature that allows me to see which runtimes are active helps me keep track of my training runs efficiently. I also like that I can run model training in the background without any trouble. The convenience of accessing different files without using the terminal is another aspect I find beneficial. Review collected by and hosted on G2.com.

At times the notebooks would freeze when the output got too long, which resulted in crashing the whole browser session, this caused a lot of problems at certain occasions. A suggestion could be to automatically limit the verbose of the training run in the notebook so as to minimize this problem or at least display the last few lines of the verbose only. Also, we had some trouble with setting the right permissions for different users. Review collected by and hosted on G2.com.

I like that Vertex AI brings everything needed for machine learning into one place. Training, deploying, and monitoring models feels organized and well integrated with Google Cloud. The managed services save a lot of time and make scaling models much easier. Review collected by and hosted on G2.com.

It can be a bit overwhelming at first, especially understanding permissions, setup, and pricing. Some errors are hard to debug, and without careful monitoring, costs can grow faster than expected. Review collected by and hosted on G2.com.

We used a conversational agent (Dialogflow) with Vertex AI RAG datasets, and we were able to create an AI chatbot connected to our custom healthcare data. Review collected by and hosted on G2.com.

It does semantic search. so getting accurate data without missing any datapoints is hard. many times we were not getting full data. Review collected by and hosted on G2.com.

An AI agent is only as good as the data it can access. Vertex AI Agent Builder genuinely shines in its Retrieval-Augmented Generation (RAG) pipeline. Review collected by and hosted on G2.com.

Perhaps the biggest complaint about Agent Builder is its highly fragmented, usage-based pricing structure. Review collected by and hosted on G2.com.

It provides access to multiple third-party LLM models. The agent can connect to multiple databases, enterprise documents, and other API’s. It also offers solid security, compliance, and governance. Review collected by and hosted on G2.com.

It comes with a relatively high cost, along with additional operational overhead. Review collected by and hosted on G2.com.

What I like most about Vertex AI is how it unifies the entire machine learning workflow — from data preparation and training to deployment and monitoring. We’ve used it to streamline our ML pipeline, and the integration with BigQuery and Google Cloud Storage makes data handling incredibly efficient. The UI is intuitive, and it’s easy to move between no-code experimentation and full-scale custom model development. Review collected by and hosted on G2.com.

Some advanced configurations can be complex at first, especially for setting up custom training jobs or tuning hyperparameters. Pricing can also become high with frequent model retraining. Documentation is thorough but sometimes fragmented across different Google Cloud sections, which can slow down setup for new users. Review collected by and hosted on G2.com.

The seamless integration with BigQuery and the ability to deploy Gemini models instantly via the Model Garden makes it the fastest way to build enterprise-grade AI. Review collected by and hosted on G2.com.

The billing structure is incredibly confusing, and the costs for idle endpoints can spiral out of control if you aren't monitoring your quotas daily. Review collected by and hosted on G2.com.