AWS re:Invent 2025 - A leader's guide to agentic AI (SNR201) | AWS Events

The speaker at reInvent discusses leading in the new era of high agency with AI, highlighting the shift from deterministic automation to managing intelligent autonomy at scale. The talk covers key leadership mental model shifts in governance, risk management, organizational structure, and culture, emphasizing the need for strategic intent, real-time visibility, fluid team structures, and a research-oriented mindset. The speaker uses accounts payable as an example of business process re-envisioning and explores the technical capabilities required for agentic systems, focusing on intelligence, context, and trust, and introduces Amazon Bedrock and AgentCore as tools to enable these capabilities.

Outlines

Part 1: Introduction, Definitions

Part 2: Leadership, Mental Models

Part 3: Business Process Re-envisioning

Part 4: Technical Capabilities: Intelligence

Part 5: Technical Capabilities: Context

Part 6: Technical Capabilities: Trust

Part 7: Implementation, Future Outlook

Sign in to continue reading, translating and more.

Continue

AWS re:Invent 2025 - A leader's guide to agentic AI (SNR201)

AWS Events

Part 1: Introduction, Definitions

Introduction to Leading in the Era of High Agency

Understanding AI Agents and Their Levels of Autonomy

Key Behaviors and Attributes of Agentic Systems

The Fundamental Shifts Driving the Rise of Agentic Systems

Part 2: Leadership, Mental Models

Real-World Applications and the Shift in Leadership

Harnessing Non-Determinism and Managing Agents at Scale

Four Key Areas of Leadership Shift: Governance

Four Key Areas of Leadership Shift: Risk Management

Four Key Areas of Leadership Shift: Organization Structure

Four Key Areas of Leadership Shift: Culture

Part 3: Business Process Re-envisioning

Re-envisioning Business Processes: Accounts Payable Example

Optimizing Cash Flow with Agentic Systems in Accounts Payable

Orchestrators in Action: Amazon's DeepFleet Example

Part 4: Technical Capabilities: Intelligence

Technical Capabilities for Agentic Systems: Intelligence, Context, and Trust

Agents Need Intelligence: Model Selection and Trade-offs

Intelligence in Action: Amazon Pharmacy and Amazon Ads Examples

Intelligence in Action: Prime Video and Syngenta Examples

Part 5: Technical Capabilities: Context

Agents Need Context: Beyond Data to Context Engineering

Technical Capabilities for Context: Relationships and Semantics

Technical Capabilities for Context: Memory

Technical Capabilities for Context: Content and Machine-Readable Files

Technical Capabilities for Context: Access to Tools and Knowledge as a Capability

Context in Action: Amazon's Rufus Shopping Assistant

Part 6: Technical Capabilities: Trust

Agents Need Trust: Setting Guardrails

Trust Through Automated Reasoning Checks

Trust in Action: Amazon Text and Compliance

Part 7: Implementation, Future Outlook

AgentCore: Building Blocks for Operating Agents in Production

AgentCore in Amazon Bedrock: Providing Primitives for Agentic Systems

Getting Started with Agents: Focus on Dynamic Tool Selection and Adaptability

Preparing Organizations with Skills and Training

AWS re:Invent 2025 - A leader's guide to agentic AI (SNR201)

AWS Events

Part 1: Introduction, Definitions

00:00Introduction to Leading in the Era of High Agency

Introduction to Leading in the Era of High Agency

01:31Understanding AI Agents and Their Levels of Autonomy

Understanding AI Agents and Their Levels of Autonomy

03:07Key Behaviors and Attributes of Agentic Systems

Key Behaviors and Attributes of Agentic Systems

04:21The Fundamental Shifts Driving the Rise of Agentic Systems

The Fundamental Shifts Driving the Rise of Agentic Systems

Part 2: Leadership, Mental Models

06:03Real-World Applications and the Shift in Leadership

Real-World Applications and the Shift in Leadership

07:41Harnessing Non-Determinism and Managing Agents at Scale

Harnessing Non-Determinism and Managing Agents at Scale

08:55Four Key Areas of Leadership Shift: Governance

Four Key Areas of Leadership Shift: Governance

10:48Four Key Areas of Leadership Shift: Risk Management

Four Key Areas of Leadership Shift: Risk Management

12:29Four Key Areas of Leadership Shift: Organization Structure

Four Key Areas of Leadership Shift: Organization Structure

14:54Four Key Areas of Leadership Shift: Culture

Four Key Areas of Leadership Shift: Culture

Part 3: Business Process Re-envisioning

16:30Re-envisioning Business Processes: Accounts Payable Example

Re-envisioning Business Processes: Accounts Payable Example

18:34Optimizing Cash Flow with Agentic Systems in Accounts Payable

Optimizing Cash Flow with Agentic Systems in Accounts Payable

20:21Orchestrators in Action: Amazon's DeepFleet Example

Orchestrators in Action: Amazon's DeepFleet Example

Part 4: Technical Capabilities: Intelligence

21:32Technical Capabilities for Agentic Systems: Intelligence, Context, and Trust

Technical Capabilities for Agentic Systems: Intelligence, Context, and Trust

23:15Agents Need Intelligence: Model Selection and Trade-offs

Agents Need Intelligence: Model Selection and Trade-offs

25:23Intelligence in Action: Amazon Pharmacy and Amazon Ads Examples

Intelligence in Action: Amazon Pharmacy and Amazon Ads Examples

26:11Intelligence in Action: Prime Video and Syngenta Examples

Intelligence in Action: Prime Video and Syngenta Examples

Part 5: Technical Capabilities: Context

28:04Agents Need Context: Beyond Data to Context Engineering

Agents Need Context: Beyond Data to Context Engineering

29:44Technical Capabilities for Context: Relationships and Semantics

Technical Capabilities for Context: Relationships and Semantics

31:41Technical Capabilities for Context: Memory

Technical Capabilities for Context: Memory

33:29Technical Capabilities for Context: Content and Machine-Readable Files

Technical Capabilities for Context: Content and Machine-Readable Files

34:53Technical Capabilities for Context: Access to Tools and Knowledge as a Capability

Technical Capabilities for Context: Access to Tools and Knowledge as a Capability

36:04Context in Action: Amazon's Rufus Shopping Assistant

Context in Action: Amazon's Rufus Shopping Assistant

Part 6: Technical Capabilities: Trust

37:03Agents Need Trust: Setting Guardrails

Agents Need Trust: Setting Guardrails

38:44Trust Through Automated Reasoning Checks

Trust Through Automated Reasoning Checks

40:24Trust in Action: Amazon Text and Compliance

Trust in Action: Amazon Text and Compliance

Part 7: Implementation, Future Outlook

41:11AgentCore: Building Blocks for Operating Agents in Production

AgentCore: Building Blocks for Operating Agents in Production

42:35AgentCore in Amazon Bedrock: Providing Primitives for Agentic Systems

AgentCore in Amazon Bedrock: Providing Primitives for Agentic Systems

44:11Getting Started with Agents: Focus on Dynamic Tool Selection and Adaptability

Getting Started with Agents: Focus on Dynamic Tool Selection and Adaptability

45:30Preparing Organizations with Skills and Training

Preparing Organizations with Skills and Training