Generative AI architecture patterns in production for SMBs (SMB304)

Here is a detailed summary of the video transcription, broken down into sections for better readability:

Evolution of Generative AI

  • The field of AI has been around for 50-70 years, but it took a major turn in late 2022 with the launch of ChatGPT.
  • This made AI accessible to millions of people, allowing them to be more productive and address pressing queries.
  • In 2023, AWS launched the Amazon Bedrock service to give customers access to a large array of foundation models to build generative AI applications.
  • 2023 was a year of exploration and experimentation, while 2024 focused on observability, security, and delivering real business value.
  • Now in 2025 and beyond, the focus is shifting towards scaling, increasing developer velocity, and reducing time-to-market.

Common Generative AI Adoption Patterns

The presenters have identified three common mindsets or strategies that customers are taking when adopting generative AI:

AI Pioneers

  • These are organizations that are heavily investing in building transformative, customer-facing use cases that can disrupt their industry.
  • Examples include a law firm building a large language model focused on the legal domain, an education tech company creating an immersive personalized learning platform, and a digital health insurance company offering hyper-personalized policy generation.

Adopters

  • These organizations are gradually building skills and releasing quick-to-market use cases to get business value, such as internal and customer-facing chatbots, image/video editing for e-commerce, and personalized marketing.

Observers

  • These are customers who are closely observing the space but not yet developing any generative AI applications, often due to regulatory or business priorities.

Common Architecture Patterns

The presenters then discussed three common architecture patterns they have seen customers use for generative AI applications:

Employee Productivity and Knowledge Management

  • This pattern uses Amazon Q Business, a generative AI-powered assistant, to consolidate knowledge from various sources and make it accessible to employees through channels like Slack.
  • Key considerations include flexible authentication, subscription tiers, observability, multimodal support, and evaluation frameworks.

Intelligent Document Processing

  • This pattern leverages Amazon Bedrock to power intelligent document processing use cases, such as a consulting firm's payment intelligence solution.
  • It discusses the "retrieve, augment, generate" (RAG) architecture, including techniques like caching, query translation/expansion, query routing, and re-ranking.

Image Generation and Modification

  • This pattern uses AWS services like S3, Lambda, and purpose-built accelerators like Inferentia to build scalable, serverless image generation and modification workflows.
  • Examples include a global developer of generative AI-powered image and video editing apps.

The Platform Mindset

As generative AI applications scale, the presenters recommend adopting a "platform mindset" to address common challenges like cost, developer velocity, and sharing best practices.

The key components of a generative AI application platform include:

  1. Data for AI: Providing data pipelines and knowledge bases to ingest, transform, and store data for AI.
  2. AI Core: Offering secure access to foundation models, prompt management, agents, and evaluation capabilities.
  3. Application Backends: Delivering general-purpose compute and storage for application-specific logic and workflows.
  4. Application Operations: Enabling observability, security, CI/CD pipelines, and business metrics reporting.
  5. Platform Control Plane: Handling onboarding, governance, developer assistance, and platform-level operations.

These components can be built using a variety of AWS services, with recommended architectural patterns and account structures.

How AWS Can Help

The presenters outlined several ways AWS can support customers in their generative AI journey, including:

  • Free discovery workshops to evaluate use cases
  • Design and insights workshops to assess data and architecture
  • "Demo squad" offerings to quickly prove business value
  • Enablement through immersion days and MVP support
  • Scaling support through AWS Professional Services and the partner ecosystem

Your Digital Journey deserves a great story.

Build one with us.

Cookies Icon

These cookies are used to collect information about how you interact with this website and allow us to remember you. We use this information to improve and customize your browsing experience, as well as for analytics.

If you decline, your information won’t be tracked when you visit this website. A single cookie will be used in your browser to remember your preference.

Talk to us