Press "Enter" to skip to content
Credit: Freepik

IBM, AWS Deepen AI Collaboration to Help Enterprises Scale Agentic AI

IBM and AWS are expanding their AI partnership to help developers and enterprises accelerate adoption of agentic AI technologies through joint product integrations and governance frameworks.

“We have been collaborating with AWS across the AI stack for at least two to three years now,” said Vishwani Dua, IBM’s CTO for the AWS partnership. “We’re really helping to democratize AI for our customers together, and developers are a key part of that customer base.”

At the center of the collaboration is a new integration between IBM’s watsonx Orchestrate and Amazon Q Index. “Amazon Q Index is a feature of Amazon Q Business, which allows customers to build their own virtual agents,” Dua said. “What we’re doing is basically offering customers the best of both worlds – integrating Q Index directly with Orchestrate so customers can build an AI agent through Orchestrate and leverage the different data sources from Q Index from AWS.”

The companies are also working together in areas such as foundation models, AI governance and data connectivity. “We have partnered across the AI stack – around data, AI agents, foundation models and even AI governance,” Dua said. “We’re really looking at … how we can complement the technologies together to really give a best solution to the end user.”

One joint initiative involves integrating IBM’s Watsonx.governance product with Amazon SageMaker. “Customers using SageMaker now can leverage those models directly into Watsonx.governance and bring in the compliance and the model monitoring and the lifecycle of those models and solutions together,” Dua said.

IBM and AWS have worked with customers such as GigaSpaces to implement this joint governance solution. “They’ve leveraged this for model governance for the use case that they’ve implemented,” Dua said. “It helped with the monitoring and making sure that the models they were building through SageMaker weren’t drifting.”

Dua noted that many enterprises are now looking to scale and secure their AI agents. “The trend really now is how do you scale it, how do you properly secure it and govern it,” Dua said. “That’s where Orchestrate comes in. Orchestrate is a model router and a planner, which allows customers to use their in-house agents, even third-party agents, and bring it all together.”

IBM’s agentic AI tools also include prebuilt domain agents in HR, sales and procurement. “As part of Watsonx Orchestrate, we are offering to customers prebuilt domain agents in those three domain areas,” Dua said. “The prebuilt agents give you a really high leg up to get something going quickly and in production faster.”

Developers can also create their own agents using IBM’s Agent Development Kit, which offers a low-code or no-code option. “You can leverage a no-code or a low-code option to create your agent very quickly,” Dua said.

To reduce costs, Dua said enterprises are turning to smaller, fine-tuned models. “Using a smaller language model, making sure that it’s fine-tuned for your specific use case is the way to go to make sure that it’s cost effective,” Dua said. “Fine-tuning large models is very expensive. If it’s not giving you the right answer and you’re hitting it over and over, you’re adding to that cost.”

Dua said IBM and AWS are making life easier for developers in several ways, including putting over 70 IBM products on the AWS Marketplace. “Customers can come in, developers can come in, and some of them have free trials so they can quickly leverage one of the free trials and try these out,” Dua said.

Other integrations include IBM’s open Lakehouse product Watsonx.data with AWS data services such as Glue, RDS and Aurora, as well as IBM’s Granite family of foundation models now available on Amazon Bedrock and SageMaker.

While developers vary in their readiness to adopt generative and agentic AI, Dua said a Center of Excellence (CoE) can help companies shift. “That’s where a Center of Excellence comes into play,” Dua said. “It’s not just that one developer needs to learn. It’s really the shift of the company and it’s really the shift of the industry.”

Looking ahead, Dua said the partnership will continue to grow. “We’re going to continue to bring more products on Marketplace and continue around our AI stack,” she said. “There are enhancements we’re going to focus on for our Orchestrate agentic AI solution with Q Developer and Q Index.”

Author

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *