The Importance of Context in AI x Infra: Your Infra is a Graph
A Talk by Roxane Fischer (CEO, Anyshift.io)
About this Talk
Large Language Models (LLMs) are powerful, but they often struggle when generating Infrastructure-as-Code (IaC) because they lack the full context of your environment. This session will explore:
Why LLMs fail for the last mile in IaC generation – AI tools like Copilot or ChatGPT can generate snippets, but they fall short when understanding dependencies, resource constraints, and configurations. We’ll break down intrinsic limitations and why tools like Cursor can only go so far without deeper awareness.
Your infrastructure is a graph – Modern cloud infrastructure is not just a collection of isolated resources; it forms a complex, interconnected graph. To reason about it effectively, AI needs to treat it as such.
Live demo: AI-powered infra insights with context – A simple Slackbot can provide precise answers about your infrastructure when it has access to the right contextual graph. I’ll demonstrate how having a structured, connected knowledge base changes the game.
How to build an infra graph for AI – Practical steps to construct and maintain a contextual knowledge graph of your infrastructure, enabling AI to provide relevant, accurate insights rather than generic guesses.
Understand why LLMs struggle with IaC, why infrastructure should be treated as a graph, and how to build an AI-powered infra knowledge graph for better insights.
Equal Experts is a global network of experienced technology consultants specialising in agile software delivery, digital transformation & data-driven solutions. Known for its collaborative, high-expertise approach, they help organisations solve complex challenges and build lasting capabilities.