The AI-Accelerated Incident

Incident Management in the Age of AI

As we lean into AI-assisted coding, the speed of delivery increases alongside the potential for new and complex errors. This 90-minute workshop explores the intersection of Incident Management (IM) and AI-Assisted Development, designed to bridge the knowledge gap and unify the team around maintaining a high bar for reliability.

What might happen when we tackle incidents as we introduce more AI into our software lifecycle?

What are the most important things we can focus on in terms of reliability to increase the chance of success and limiting potential live incidents?

We will move beyond standard process basics to scrutinise how AI can introduce non-deterministic errors, identify specific risks (like "AI-Induced Debt"), and practice high-stakes triage using a real-world-inspired scenario. Through this, the team will build a shared vocabulary and confidence when live incidents happen.

Who is this workshop for?

Junior & Mid-level Engineers: to learn the "why" behind incident protocols and how to scrutinise AI-generated output.
Engineering Managers: to understand the trade-offs between delivery velocity and system stability.
Senior & Staff Engineers: to share mental models of system architecture and calibrate their "smell tests" for AI-coded features.

What you'll learn

✔️

Level the knowledge gap: juniors will gain a framework for incident response, while Seniors will align on AI-specific guardrails.

✔️

Identify "AI-Induced Debt": recognize patterns where AI-generated code might ignore edge cases or bypass existing safety patterns.

✔️

Implement "The Human Circuit Breaker": define specific points in the PR and deployment process where AI-generated work requires manual validation.

✔️

Master incident triage: apply a shared vocabulary for high-pressure situations, ensuring communication remains clear even when the root cause is an obscure AI-generated logic error.

✔️

Practice high-stakes response: gain the confidence and shared vocabulary needed to effectively take on the on-call rota, knowing the whole team is aligned on AI-specific incident triage.

Course content

1. Introduction

2. The incident management lifecycle

3. AI: a double edged-sword

4. AI: a double edged-sword

5. Interactive exercise

6. Closing and next steps

Course format

Workshop can be delivered in person or online via Zoom
Workshop duration is 90-120 minutes (main difference is the time allowed for the interactive exercise, allowing participants to go deeper)
Can be delivered to engineering teams

Want this for your organisation?

Let's connect