AWS Partner Network (APN) Blog

How AWS Partners are Driving Innovation and Efficiency with Amazon Bedrock and Amazon Q

By Chris Dally, WW Head, Generative AI & ML Specialized Partners – AWS
By Victor Rojo, WW Tech Lead, Generative AI & ML Specialized Partners – AWS

Want to get just the key points of this article? Try Article Summarizer on PartyRock.

In April, Amazon Web Services (AWS) unveiled a suite of groundbreaking features for Amazon Bedrock and Amazon Q, ushering in a new era of generative AI capabilities. These advancements have the potential to transform how businesses and developers build, scale, and deploy intelligent applications, unlocking unprecedented opportunities for innovation and efficiency.

AWS Partners responded to the new Amazon Bedrock and Amazon Q features with overwhelming excitement. On launch day, AWS briefed Generative AI Competency Partners on the features and asked about the groundbreaking projects they’ve collaborated on with customers. Find that briefing and other getting started info in the Generative AI Center of Excellence for AWS Partners (login required).

Partners shared use cases, like utilizing natural language search with Amazon Q to make information retrieval from various sources easier and automating tasks like data analysis, customer segmentation, and hiring with Amazon Bedrock AI models, freeing valuable time for more strategic initiatives. Partners also prioritize responsible and legal use of generative AI through setting up Guardrails policies. They emphasize how these new capabilities unlock transformative potential for secure and ethical AI adoption with mitigated risks. These features empower partners to confidently deliver robust generative AI solutions at scale for our joint customers.

Amazon Bedrock: Empowering enterprises with unparalleled flexibility and security

As the demand for generative AI solutions skyrockets, enterprises are seeking robust, flexible, and secure tools to develop intelligent applications. In response, AWS enhanced its flagship AI service, Amazon Bedrock, with a comprehensive suite of features that offer a user-friendly, scalable, and intuitive solution.

“At DoiT International, we partnered with mscope to establish the cloud infrastructure needed to train and deploy data models that could integrate multi-language, web-scraped data into a common language and classify the information using an algorithm.” said Jaret Chiles, Chief Services Officer at DoiT. “DoiT recommended mscope transition their LLM model from Mistral to Amazon Bedrock, allowing the customer to experiment at a lower cost while building and scaling their systems. As a result, mscope was able to optimize processing times and reduce costs, maximizing the value of their multi-language data and accelerating the time to market for their data-based solutions.”

Through the custom model import, customers can now bring their own custom models, trained outside of Amazon Bedrock, to the service. By leveraging their prior investments in model customization within the fully-managed Amazon Bedrock environment, enterprises can enhance their AI capabilities without starting from scratch. This feature supports popular open architectures, like Flan-T5, Llama, and Mistral, enabling seamless integration of customized models tailored to unique business requirements.

AWS has also streamlined the model selection process with the new model evaluation capability. By evaluating the outputs of different models against built-in or custom prompt datasets, users can confidently make informed decisions when choosing the best model for their use case, ensuring optimal performance and results.

“Amazon Bedrock has revolutionized how we handle AI workloads,” said Bakrudeen K, Head of AI/ML practice at ShellKode. “It has made everything from evaluating models to training and utilizing them much more efficient. For example, we worked with PreSkale to assist their AI-powered PreSales platform in achieving a remarkable 40% improvement in product gap creation. Additionally, all this is secured by the same robust frameworks that safeguard AWS’ critical services.”

Recognizing the paramount importance of safety and security, AWS released Guardrails for Amazon Bedrock, which is now generally available. This feature provides robust safeguards to prevent harmful content and manage sensitive information within generative AI applications, ensuring peace of mind for enterprises. Guardrails offers customizable safety filters, sensitive information detection and redaction, and word filters aligned with an organization’s responsible AI policies.

“At Innovative Solutions, we are using Guardrails to ensure our Tailwinds product adheres to the highest standards of safety, privacy, and ethical conduct,” said Travis Rehl, CTO of Innovative Solutions. “This enables our clients to confidently leverage generative AI capabilities while mitigating risks and maintaining full compliance with industry regulations and internal policies.”

Amazon Bedrock’s updated console streamlines the process for developers to start building with Agents. The simplified console allows for swift creation of new Lambda functions for action groups, leading to better performance through Sonnet and Haiku. Function definitions expedite schema creation, eliminating the need to conform to OpenAPI schema standards. The implementation of Return of Control enables efficient integration with backend services and the application of business logic to actions. Furthermore, robust support for CloudFormation ensures a seamless integration and deployment process, providing customers with reassurance of the product’s reliability.

Amazon Titan Foundation Models: Cost-effective, accurate text embeddings and secure, high-quality image generation for diverse industries. 

Amazon Titan Text Embeddings V2 is a new model that offers several improvements over its predecessor. Designed to significantly reduce storage and compute costs while maintaining high accuracy, the model achieves this by allowing more flexible embeddings, which can reduce overall storage needs by up to four times. Despite these reductions, it retains 97% of the accuracy for Retrieval-Augmented Generation (RAG) use cases, making it more efficient than other leading models in the market.

Titan Image Generator, now in general availability, is a valuable tool for customers in industries such as advertising, e-commerce, and media and entertainment. It enables the production of studio-quality images or the enhancement and editing of existing images at a low cost using natural language prompts. A notable feature of the Titan Image Generator is its application of an invisible watermark to all images it creates, helping identify AI-generated images and promoting the safe, secure, and transparent development of AI technology.

Amazon Q: Revolutionizing productivity and collaboration

In addition to Amazon Bedrock enhancements, AWS announced the general availability of Amazon Q, its most advanced generative AI-powered assistant. With Amazon Q, developers and employees can speed up their work and access internal data more efficiently.

Amazon Q Business provides AllCloud a solution to increase productivity by streamlining information access,” said Peter Nebel, CTO of AllCloud. “By leveraging Amazon Q’s natural language search capabilities, AllCloud can empower its personnel with a central hub to find answers to their questions across all their existing information sources. This drives efficiency and accuracy by eliminating the need for time-consuming searches across multiple platforms and ensuring all teams have access to the most up-to-date information. Amazon Q will significantly accelerate productivity, across all lines of business, allowing AllCloud’s teams to focus on delivering exceptional service to their clients.”

With Amazon Q Developer, users now have an AI partner that helps them do more than simply generate code. It can assist with testing, debugging, and multi-step planning to implement new code based on users’ requests, making it a valuable partner in software development and innovation.

Capgemini and their client, a global energy leader, leveraged Amazon Q over the course of an eight-week engagement to optimize vessel and overall fleet performance while upholding environmental standards. Capgemini recently partnered with them to harness the power of generative AI within their software development lifecycle.

“Our client expressed specific requirements for using generative AI to optimize the technology capabilities they offer to their maritime customers,” said Niels Thomsen, Vice President, Insights & Data with Capgemini. “Their requirements included improving code quality and performance through automated coding assistants, automating the creation of SDLC documentation, and automating the creation of test harnesses as well as software testing data.”

In addition to achieving the initial goals of optimizing the software development life cycle, the combined teams also planned innovative, new capabilities—such as enabling vessels to optimize fuel consumption based on a combination of historical IoT data as well as dynamic variables such as weather and sea roughness. “Our results are consistent with achievements we are seeing across multiple industries,” Thomsen continued. “Given the results we have achieved together in less than two months, we are excited to continue working with our client and AWS to bring additional improvements in both software performance and sustainable innovation for the maritime industry.”

Expanding on these achievements, Capgemini recently announced the availability of the RAISE framework—which enables clients to produce Generative AI solutions at scale on AWS. “Together with innovative products like Amazon Q, SageMaker, Amazon Bedrock and Amazon Kendra, RAISE on AWS enables us to not only create successful generative AI POCs, but to take those POCs into production quickly and with reduced risk,” said Thomsen.

Amazon Q Business offers a comprehensive solution for organizations of all sizes. It is a versatile tool that connects to various enterprise data repositories, catering to the diverse needs of employees across an organization. It can summarize data, analyze trends, and provide information about company policies, product details, business outcomes, and more. Additionally, Amazon Q supports custom plugins, enabling seamless integration with any third-party application, enhancing its adaptability and potential for customization.

“Amazon Q continues to garner interest from our customers with features such as context-aware simplified Q&A, and built-in connectors, ensuring a superior user experience to drive faster adoption,” said Sivaram Sivanarayanan, Product Leader at Tata Consultancy Services‘ (TCS) AI.Cloud business unit. “One of our Life Sciences customers leveraged Amazon Q to reimagine cloud operations, realizing early success with benefits such as a 20% reduction in toil and up to 15% efficiency gains in assimilating information for responding to (audit) queries. The simplified Q&A chat interface’s ease of interaction in querying underlying data and uncovering insights has been a key driver, elevating the overall user experience. Amazon Q has the potential to revolutionize the use of highly intuitive self-service interfaces for gaining valuable insights.”

AWS has also launched Amazon Q in Connect, which utilizes real-time speech analytics, natural language processing, and generative AI to enhance agent productivity and customer satisfaction. This feature improves upon Amazon Connect Wisdom by incorporating advanced large language model technology, enabling contact center agents to directly access Amazon Q in their Amazon Connect workspace, receiving automated recommendations, and searching for information conversationally.

Another exciting feature of Amazon Q is Amazon Q Apps, which empowers employees to turn their ideas into AI-powered applications that streamline tasks, boost productivity, and foster team collaboration.

Partner testimonials

“Rackspace’s customer, a financial services company, manages hundreds of thousands of support emails monthly, and it takes their support team hours to process each one and route them to the appropriate group for resolution,” said Nirmal Ranganathan, VP of AI at Rackspace Technology. “We implemented a four-step process using Amazon Bedrock models (Claude v3 LLM) to summarize the email and automate the support case management. This will reduce email-to-ticket duration from hours to minutes with an 80% higher ticket SLA achievement, increase the agility of support teams by recommending issues from the knowledge base, and provide cost savings due to task automation.”

“Caylent partnered with IdenX to deliver an Amazon Bedrock-powered solution to analyze incoming unstructured data and assign a ‘value’ score,” said Randall Hunt, VP of Cloud Strategy & Innovation at Caylent. “This work improved the efficiency of their manual data science efforts by 2X and allowed their teams to focus on the highest value data. It also jumpstarted the usage of generative AI within the company and empowered IdenX to bring Amazon Bedrock powered solutions to other work streams.”

Partners: Shape the future of generative AI with AWS

Get started today by exploring our cutting-edge services, including the newly launched Amazon Bedrock Studio—a collaborative web-based development environment that offers a seamless workspace for experimenting with foundation models and building innovative applications, providing you the opportunity to accelerate your AI development and enhance your customer solutions.

Also, visit the Generative AI Center of Excellence for AWS Partners (AWS Partner Central login required). Beyond a resource hub, the Generative AI CoE is a collaborative community that provides optimized learning paths, curated curricula, and forums, designed to collectively advance the applications of generative AI for our joint customers. Be sure to check out the Generative AI Partner Playbook, housed within the CoE.

Ready to take your business to new heights? Become an AWS Generative AI Competency Partner to gain access to the latest advancements and go-to-market resource from AWS. Partners looking to get their Generative AI offering validated through the AWS Competency Program must be validated or differentiated members of the Software or Services Path prior to applying. To apply, please review the Program Guide and access the application in AWS Partner Central.

AWS Generative AI Competency Partners have proven success in delivering tailored consulting services, foundation models, and applications that drive efficiency, ingenuity, and productivity across industries. This is an opportunity to revolutionize your business and shape the future of generative AI, unleashing your potential for innovation and creativity. Together, we can redefine what wasn’t possible.