Operating in the agentic era: AIOps sessions at re:Invent 2026

Telemetry volume and incident counts keep growing while operations headcount stays flat, and coding tools now produce changes faster than most release processes can review them. The AIOps topic on the re:Invent 2026 Cloud Operations (COP) track covers how AI agents handle operational work: investigating incidents, recommending fixes, reviews code changes to catch issues before the reach productions, handles on-demand SRE tasks, and drive proactive improvements, and leaving the audit trail production environments require. This post walk you through the key themes of AI Operations (AIOps) sessions and how customers are leveraging AWS DevOps Agent to help you build an agenda. Use the AWS Events app to favorite, build schedule, and reserve your seat.

Plan your AIOps sessions

The lineup spans breakouts, workshops, builders’ sessions, chalk talks, code talks, and lightning talks. Pair a strategy or architecture session with a workshop or builders’ session so you leave with both a model and working patterns. Chalk talks and code talks are discussion-heavy, so bring your own problems. Workshops and builders’ sessions have limited seats, so reserve early.

Jump to a theme:

• Agentic operations: strategy and architecture
• Investigate and resolve incidents with AWS DevOps Agent
• Building and extending agents
• Acting safely: remediation, auditability, releases, and modernization

Agentic operations: strategy and architecture

COP331 | AIOps strategy: foundations before firefighting | Chalk Talk

Location: Wednesday Dec 2 4:30 PM–5:30 PM PST | Wynn/Encore · Level 1 · Chopin 4
Repeat: Thursday Dec 3 1:30 PM–2:30 PM PST | Wynn/Encore · Level 1 · Chopin 4

Many teams turn on AI-driven operations before deciding what it should handle, and investment races ahead of maturity. At the whiteboard, this chalk talk uses real support-case patterns to show that most critical failures can be caught before production. It maps both halves teams miss: using AI to run operations, and the cost, security, and organizational foundations that let AI scale safely. Leave with a maturity checklist and a view of where agents earn autonomy.

Who should attend: Operations and platform leaders setting an AIOps strategy and deciding where to invest first.

COP301 | Stop fighting fires: operations that prevent incidents | Breakout Session

Location: Thursday Dec 3 10:00 AM–11:00 AM PST | Venetian · Level 2 · Ballroom F · Content Hub · Red Theater

The best incident is the one that never happens, yet most teams stay stuck firefighting. Learn to collect the right operational signals beyond observability and use AI to spot problems before anyone gets paged. The session covers auto-healing and remediation with both traditional techniques and AI agents that catch misconfiguration before production, and shows how agentic investigation turns past incidents into ranked recommendations that prevent future ones.

Who should attend: Operations teams and SREs who want to move from reactive firefighting to proactive incident prevention.

COP302 | Best practices for agentic-driven operations feat. Best Buy | Breakout Session

Location: Thursday Dec 3 10:30 AM–11:30 AM PST | Caesars Palace · Promenade South · Octavius 4 · Content Hub · Blue Theater

AI agents are moving from assistants to operators. This session covers the architecture of agents that detect anomalies, diagnose root cause across distributed systems, and run remediation such as rollbacks, scaling, and traffic shifts. Learn how agents reason over telemetry, sequence actions, and decide when to escalate, and how guardrails, approval gates, and feedback loops shape their behavior. Best Buy shares how it decides where agents act alone and where people stay in control.

Who should attend: Architects and SRE leads designing agent-driven incident response and remediation.

COP303 | Improving operations with agentic AI feat. Verisk | Breakout Session

Location: Tuesday Dec 1 1:30 PM–2:30 PM PST | Wynn/Encore · Upper Convention Promenade · Cristal 2 · Content Hub · Purple Theater

Operational telemetry grows faster than the teams that watch it. See how agentic AI works across your observability stack, correlating metrics, logs, and traces to detect issues, find root cause, and start remediation without waiting for a person to connect the dots. Explore patterns for agents that triage alerts, reduce noise, and run operational runbooks. Verisk shares how it integrates agents into monitoring and incident response to cut mean time to recovery and repetitive toil.

Who should attend: Observability and on-call teams drowning in telemetry and alert noise.

COP351 | How Infor Uses AWS Agentic AI to Transform Cloud Operations at Scale | Lightning Talk

Location: Monday Nov 30 2:30 PM–2:50 PM PST | Wynn/Encore · Upper Convention Promenade · Cristal 2 · Content Hub · Lightning Theater

Infor, a Koch Industries company, runs one of the largest enterprise footprints on AWS. As it grew, manual release gates, incident investigations, and periodic security assessments couldn’t keep pace. Hear how Infor is adopting AWS agentic AI: DevOps Agent for incident triage and release readiness, Continuum for continuous security validation, and the Well-Architected Agentic AI Lens for production guardrails. Leave with a blueprint for running agentic AI operations at scale.

Who should attend: Enterprises with large AWS footprints planning an agentic AI operations rollout.

Investigate and resolve incidents with AWS DevOps Agent

COP202 | Autonomous operations: reducing MTTR at scale, feat. BMG Bank | Lightning Talk

Location: Monday Nov 30 4:30 PM–4:50 PM PST | Venetian · Level 2 · Hall B · Expo · Global Impact Theater

Banco BMG, one of Brazil’s most established financial institutions, handled roughly 350 operational events a day, each taking hours to investigate manually. In this lightning talk, see how AWS DevOps Agent began investigating those events autonomously and cut root cause identification from hours to minutes. Learn how to integrate the agent with your observability tools, correlate telemetry, logs, and code, and turn each investigation into lasting MTTR improvements.

Who should attend: Teams facing high incident volume who want a concrete MTTR-reduction story.

COP352 | Autonomous SRE: AWS DevOps Agent in Financial Services | Lightning Talk

Location: Monday Nov 30 3:00 PM–3:20 PM PST | Wynn/Encore · Upper Convention Promenade · Cristal 2 · Content Hub · Lightning Theater

Learn how Capital One runs AWS DevOps Agent as an autonomous SRE teammate, cutting incident investigation time from hours to minutes and freeing engineers from reactive firefighting. Explore an architecture spanning Amazon CloudWatch, Splunk, GitHub, and Slack, and see how the agent correlates telemetry across sources to identify root cause and draft mitigation plans. Leave with patterns for deploying agentic operations in regulated, compliance-sensitive environments.

Who should attend: SRE and operations teams in regulated industries evaluating autonomous incident response.

COP362 | Troubleshoot Workloads End to End with AWS DevOps Agent | Lightning Talk

Location: Tuesday Dec 1 4:30 PM–4:50 PM PST | Venetian · Level 2 · Hall B · Expo · Builder Theater

Application issues can start anywhere in the stack, from compute and containers to networking, data, and AI services. In this lightning talk, see how AWS DevOps Agent investigates issues end to end across AWS, multicloud, and on-premises environments, analyzing telemetry and correlating signals across layers to pinpoint root cause and deliver targeted remediation. Learn how to resolve issues faster while spending less time moving between tools and data sources.

Who should attend: Engineers troubleshooting complex, multi-layer workloads who want a quick look at end-to-end investigation.

COP354 | Diagnose Complex Workloads with AWS DevOps Agent | Code Talk

Location: Wednesday Dec 2 11:30 AM–12:30 PM PST | MGM Grand · Level 1 · Boulevard 159

Follow an investigation of a complex workload with AWS DevOps Agent. Trace an application-level symptom through multiple layers of your environment across AWS, multicloud, and on-premises systems. See how the agent analyzes telemetry, correlates signals across layers, identifies likely root causes, and delivers targeted remediation. Learn practical techniques for investigating distributed workloads and reducing the manual effort of diagnosing problems.

Who should attend: SREs and developers who want to see a full investigation worked through in code, layer by layer.

COP363 | Investigate Incidents Across Teams with AWS DevOps Agent | Chalk Talk

Location: Thursday Dec 3 12:00 PM–1:00 PM PST | Caesars Forum · Level 1 · Alliance 305

Production incidents usually involve multiple teams, but tooling and investigations are often siloed. This chalk talk explores how AWS DevOps Agent meets teams where incident response already happens. See how engineers work with the agent directly from tools like Slack, keeping investigation activity, findings, and recommended actions alongside team communication. Bring your multi-team incident patterns to discuss faster, more connected response across AWS, multicloud, and on-premises environments.

Who should attend: Incident commanders and on-call leads coordinating response across several teams.

COP317 | Hands-on with AWS DevOps Agent | Workshop

Location: Monday Nov 30 3:30 PM–5:30 PM PST | Wynn/Encore · Convention Promenade · Margaux 2
Repeat: Wednesday Dec 2 12:00 PM–2:00 PM PST | MGM Grand · Level 3 · Premier 317

AWS DevOps Agent investigates, resolves, and helps prevent incidents. In this workshop you take the on-call seat: inject real faults into a live microservices store on Amazon EKS, work with the agent to find root cause in minutes, and judge its mitigation plans. Then extend it by encoding your team’s expertise as skills, connecting custom MCP tools, triggering investigations from alarms, and approving agent-executed fixes. Leave with patterns you can take to production.

Who should attend: Hands-on SREs and on-call engineers who want to try the agent on a live system.

COP355 | Resolve Incidents and Validate Releases with AWS DevOps Agent | Builders’ Session

Location: Wednesday Dec 2 11:30 AM–12:30 PM PST | Caesars Forum · Level 1 · Alliance 311

Follow an issue from detection to resolution. In this builders’ session, see how AWS DevOps Agent investigates incidents across AWS, multicloud, and on-premises environments, analyzing telemetry and correlating signals to pinpoint root cause and deliver targeted remediation. Then use the agent to review code changes for production readiness before they reach customers, with every review informed by the production intelligence the agent builds from your environment.

Who should attend: Builders who want hands-on time with both incident investigation and release readiness in one session.

Building and extending agents

COP320 | Custom SRE agents: From prompt to proactive ops with DevOps agent | Builders’ Session

Location: Monday Nov 30 3:00 PM–4:00 PM PST | Wynn/Encore · Convention Promenade · Latour 7
Repeat: Wednesday Dec 2 3:00 PM–4:00 PM PST | Wynn/Encore · Upper Convention Promenade · Cristal 2 · Content Hub · Builders’ Session 2

Learn to create custom SRE agents that turn a simple prompt into environment-specific operational workflows. Using pre-configured tools and integrations, you define agents for your priorities, such as anomaly detection, performance optimization, cost analysis, or resiliency recommendations. See how agents combine real-time observability data with structured reasoning to surface improvements without manual investigation. Leave having designed an agent that watches your operational patterns and recommends next steps.

Who should attend: Builders who want to tailor an SRE agent to their own environment and priorities.

COP402 | Scaling operations with MCP, ACP, and A2A | Chalk Talk

Location: Monday Nov 30 4:00 PM–5:00 PM PST | MGM Grand · Level 3 · Room 306
Repeat: Wednesday Dec 2 4:00 PM–5:00 PM PST | Caesars Palace · Promenade Level · Neopolitan I

Operational complexity doesn’t scale linearly, but your approach to managing it can. Using AWS DevOps Agent, this chalk talk explores three protocols: MCP for connecting to operational data and actions, ACP for delegating tasks to purpose-built agents, and A2A for coordinating autonomous agents. Discuss how protocol choice affects decisions, when to extend capabilities versus delegate, and how to design architectures that grow with your footprint. Leave knowing when to reach for each.

Who should attend: Architects deciding how to connect and coordinate agents across their operations stack.

COP404 | Go headless: managed harness to custom DevOps orchestration | Code Talk

Location: Friday Dec 4 10:30 AM–11:30 AM PST | Wynn/Encore · Convention Promenade · Margaux 1

Your AI agents need production-grade orchestration, but should you own the loop or let AWS run it? This code talk builds one on-call responder three ways: an Amazon Bedrock AgentCore harness with built-in memory and tool routing, AgentCore Runtime for full control, and AWS DevOps Agent as a ready-made DevOps harness. Then it builds the shape most teams land on: your own orchestrator with DevOps Agent composed headless over remote MCP and A2A.

Who should attend: Engineers choosing between a managed agent harness and building their own orchestration.

COP405 | When agents run code: sandboxes for reliable DevOps | Code Talk

Location: Monday Nov 30 3:00 PM–4:00 PM PST | Wynn/Encore · Convention Promenade · Margaux 1
Repeat: Thursday Dec 3 3:00 PM–4:00 PM PST | Wynn/Encore · Convention Promenade · Margaux 1

When an AI agent investigates a production incident, some answers need computation, not just reasoning. This code talk opens the sandbox in AWS DevOps Agent: an ephemeral, investigation-scoped Linux environment that runs code in a Lambda MicroVM with read-only AWS access and no lateral blast radius. See how network allowlists, pre-installed packages, and executable skills give deterministic, auditable results for log parsing, metric correlation, and multi-source analysis.

Who should attend: Engineers who need agent output they can trust and audit in production investigations.

Acting safely: remediation, auditability, releases, and modernization

COP353 | Take Action on Operational Issues with AWS DevOps Agent | Breakout Session

Location: Thursday Dec 3 8:30 AM–9:30 AM PST | Caesars Forum · Level 1 · Forum 120 · Content Hub · Red Theater

There’s a big difference between an agent that hands you a remediation plan and one that can carry it out. This session shows how AWS DevOps Agent goes beyond investigation to act on operational issues across AWS, multicloud, and on-premises environments. Learn how directed actions and defense-in-depth controls govern what the agent can do, with layered opt-in, per-action approval, and built-in guardrails that keep operations safe.

Who should attend: Operations leaders and SREs deciding how much remediation to let an agent perform, and under which controls.

COP323 | Trust but verify: Building the audit loop for AI-driven operations | Builders’ Session

Location: Tuesday Dec 1 1:00 PM–2:00 PM PST | Wynn/Encore · Upper Convention Promenade · Cristal 2 · Content Hub · Builders’ Session 1
Repeat: Wednesday Dec 2 3:00 PM–4:00 PM PST | Wynn/Encore · Convention Promenade · Latour 7

AI agents act fast, and auditors ask questions slowly. In this builders’ session, close that gap with an observability layer that serves both. Work with Amazon Bedrock AgentCore traces across multi-agent workflows, Amazon CloudWatch for decision-level visibility and drift detection, and escalation thresholds that keep people in the loop. Learn patterns for reconstructing agent reasoning after the fact, and leave with an audit loop you can run for agents in production.

Who should attend: Teams accountable for governance, compliance, and auditability of agent actions.

COP406 | Rethink release management for the agentic coding era | Code Talk

Location: Monday Nov 30 8:30 AM–9:30 AM PST | Wynn/Encore · Upper Convention Promenade · Cristal 2 · Content Hub · Code Talk
Repeat: Tuesday Dec 1 4:30 PM–5:30 PM PST | MGM Grand · Level 3 · Chairman’s 363 · Content Hub · Code Talk

GitOps was designed for human-speed change, and agentic coding tools now open pull requests faster than teams can review and release them. In this code talk, live-code a GitOps workflow where AWS DevOps Agent acts as release manager at every promotion gate. Instead of manual approvals, the agent uses runtime dependencies, service health, and deployment risk to make go/no-go calls, so releases keep pace with code generation without giving up safety.

Who should attend: Release engineers and platform teams whose review process can’t keep up with agentic code generation.

COP332 | Modernize confidently with production intelligence | Chalk Talk

Location: Monday Nov 30 4:00 PM–5:00 PM PST | Caesars Palace · Promenade South · Octavius 11

Static code analysis tells you what to modernize, but not what breaks when you do. Runtime dependencies, real traffic patterns, and operational risk stay hidden until cutover, when surprises cost the most. This chalk talk shows how production intelligence from AWS DevOps Agent feeds an AWS Transform modernization plan and an automated migration pipeline, so assessments and migrations reflect real runtime behavior. Bring your modernization challenges to discuss.

Who should attend: Teams planning migrations or modernization who want plans grounded in production reality.

Suggested paths

If you are new to agentic operations, start with COP331 and COP301, then COP302 or COP303, and COP351 for an enterprise rollout story. For customer results, see COP202 and COP352. To see AWS DevOps Agent investigate incidents, pair COP362 or COP354 with COP363, and reserve COP317 or COP355 for hands-on time. If you are building your own agents, COP320, COP402, COP404, and COP405 run from custom agents to protocols, orchestration, and safe execution. For governance, add COP353 and COP323.

Registration is open in the re:Invent portal. Stop by the AWS DevOps Agent and AI Observability kiosks in the AWS Village to explore the latest innovations in AIOps, see live demos, and connect with AWS experts to learn how you can transform your operations.

Suggested paths

If you are new to agentic operations, start with COP331 and COP301, then COP302 or COP303. For hands-on time, reserve COP317 and COP320 early. If you are building your own agents, COP402, COP404, and COP405 run from protocols to orchestration to safe execution. For governance and auditability, add COP323.

Registration is open in the re:Invent portal. Stop by the AWS DevOps Agent and AIObservability kiosks in the AWS Village to explore the latest innovations in AIOps, see live demos, and connect with AWS experts to learn how you can transform your operations.