Lennox Hill are a digital marketing agency in Perth who were looking to supercharge their web development capabilities. The team had already seen moderate benefits from using AI tools like ChatGPT and were looking to train their development team on how to use dedicated AI coding tools.
AI Coding platforms like Cursor and Claude Code have seen massive adoption across development teams and organisations who aren’t embracing AI coding are starting to fall behind.
With so many AI coding options now avaliable, it can be difficult to know where to start and seperate the hype from results. At AI Advancements we have been using AI coding tools for the past 2.5 years to deliver over 40 projects for clients across Perth, giving us plenty of experience to know what does and doesn’t work when it comes to AI coding.
Lennox Hill wanted to standardise how their developers interact with AI, ensuring they had the proper training and support to use these tools to their fullest potential.
AI Advancements ran a two-hour session with Lennox Hill’s development team to teach them the fundamentals of AI coding. The session was focused on using Cursor, an AI powered IDE which allows AI agents to directly read, modify and create files in your code base.
The session was focused on practical Cursor skills, ensuring developers left the workshop with the knowledge to immediately start getting value out of Cursor. We covered the different agent types within Cursor, equipping developers with a playbook of when to use each agent.
Figure 1: Different types of agents in Cursor.
We also covered how Lennox Hill can extend Cursor with their own company wide Cursor rules and custom MCP tooling. MCP is a standard protocol for connecting LLMs with custom tools, allowing an organisation to extend the capabilities of an AI system while being tool agnostic. As part of our MCP education, we identified existing 3rd party MCP’s which Lennox Hill could immediately start using, such as the Atlassian MCP for interacting with Jira tickets and the Figma MCP for quickly turning designs into code.
Figure 2: The purpose of the Model Context Protocol.
Developers left the workshop with all of the practical skills required to immediately start coding with Cursor and using MCP servers. AI Advancements provided recommendations for user licensing, establishing a company wide cursor rules, and the creation of a custom company MCP code library.
The Lennox Hill team gained the foundational knowledge to build their own AI systems and MCPs, extending their AI coding capabilitiy and allowing them to build their own internal solutions.
If you’re considering AI but aren’t sure where to begin, get in touch.
