AWS: Bedrock integration
AI choice matters — and now you have it all. Our AWS Bedrock actor brings Amazon, Anthropic, Meta, and Mistral AI models together in one powerful actor, ending the 'one-size-fits-all' approach to AI-powered automation. No longer restricted to a single provider's capabilities, you can now select the perfect AI model for each specific task, whether you need Amazon's Nova for contract analysis, Claude's factual consistency for legal research, Llama 3's reasoning prowess, or Mistral's lightning-fast document processing.
Available models
Amazon’s Nova models are among the most advanced, designed for legal text analysis, contract automation, and compliance monitoring. With strong reasoning capabilities and efficient processing, they’re ideal for:
- Extracting key clauses from long contracts
- Summarising regulatory documents
- Automating legal risk assessments
Claude models are known for factual consistency and deep understanding, making them a great choice for:
- Drafting contracts with minimal human input
- Legal research, where accuracy is crucial
- Redlining documents, catching ambiguous clauses before they become a liability
Meta’s Llama 3 models are built for complex reasoning and structured text generation, making them well-suited for:
- Generating detailed legal summaries
- Supporting legal chatbots and contract automation
- Enhancing document review processes with AI-assisted insights
Mistral AI delivers cutting-edge, open-weight models designed for efficiency and performance, making them a strong choice for AI-powered contract analysis and legal automation. Known for their speed and adaptability, these models excel in:
- Automating legal document categorisation, quickly classifying contracts and compliance records
- Generating precise risk assessments, flagging potential issues in agreements with minimal manual intervention
- Enhancing workflow automation, ensuring that contract reviews and compliance processes remain fast and accurate
Model configuration
Fine-tune model behaviour with key parameters to ensure the right output for your use case.
Temperature
Controls randomness in responses.
- Lower values (e.g., 0.2) make AI more deterministic - great for legal and compliance workflows where accuracy is key.
- Higher values (e.g., 0.8) make AI more creative, useful for brainstorming contract clauses or summarising content in different styles.
Top P & Top K
Fine-tunes response diversity.
- Top P (nucleus sampling) defines how diverse the AI’s response can be by considering only the most probable words. When drafting formal documents, keeping Top P low ensures responses remain formal and structured.
- Top K sets a limit on the number of top words the AI can pick from - lower values make outputs more predictable.
Maximum Length
Prevents overly long outputs. Essential for controlling response size, especially when summarising contracts or generating reports.
Stop Sequences
Defines where the AI should stop generating text. Helps in workflows where responses need a specific format, such as structured contract clauses or predefined summaries.
Summary
This AWS Bedrock integration fundamentally changes how you'll approach AI in your business processes. The days of compromising or building complex workarounds to accommodate a single model's limitations are over. Now you can match each workflow to its ideal AI partner, fine-tune every parameter to your exact requirements, and build truly intelligent automation that adapts to your needs rather than forcing you to adapt to it. Your AI toolkit just expanded dramatically - the only question is which model you'll put to work first.