Situation
When we started building the AI Learning Hub, Franco’s AI Taskforce had already spent more than a year rolling out AI tools and processes to our team. That work proved the value of AI across our communications and marketing work for clients and the agency, and it also showed us where people still needed support. The resources mostly relied on prompt templates, which often needed reworking from one project to the next and didn’t explain why a prompt worked or when to trust the results.
We surveyed our team about the tools and resources they needed to be more successful with AI, and the responses gave us a clear picture. People weren’t looking for shortcuts – they wanted the underlying skills (how to strategically use AI, how to use it ethically and accurately, how to build a strong prompt and how to apply it to their own work in a way that saved time and created space for the type of work we don’t want to use AI for). The survey also pointed out the high-priority tasks where the team most wanted help. The need wasn’t more access to tools. It was building capability and confidence.
Approach
We vibe coded the AI Learning Hub directly in Claude as a single internal home for learning and applying AI across our work. Building it in the tool itself, rather than through traditional development, let us move quickly, iterate in real time and model the same AI-assisted approach we were teaching.
We made it interactive on purpose. The guided, hands-on learning from the AI Taskforce was the part people valued most, but those sessions were periodic and easy to lose track of once the meeting ended. An interactive hub keeps that guidance available on demand, so anyone can work through a lesson or build a prompt on their own time and return to it whenever they need a refresher.
Everything starts from a shared standard for safe, ethical and effective use, captured in an adoption guide and a set of best practices. Before anyone optimizes a workflow, they understand responsible use and accuracy, so the work stays trustworthy from the start.
From there, the Academy covers the skills in short, focused modules that build from beginner to advanced.
- Prompting Fundamentals. The first module covers the difference between asking a question and engineering a prompt, the skill everything else builds on.
- Navigating Hallucinations. The next module explains why models make things up and the simple habit that keeps accuracy from becoming a problem.
- Context and Memory. The last module covers why AI can feel like it forgot everything you told it, and how to work within its memory limits to make it smarter over time.
Throughout, the content is grounded in our own knowledge and content as the basis for effective use. The best results come from bringing Franco’s expertise, context and standards to the tools, rather than expecting the tools to do the thinking for us.
That thinking carries into the prompt builders. Instead of handing over a static prompt, each one walks a team member through the components of a strong prompt and then assembles them into a clear, editable structure they can refine and reuse. We built them for those same tasks the survey surfaced, the everyday work where better prompting saves the most time (spanning media relations to operations) with room to grow as new needs come up.
Results
The Hub gave the team a single resource in place of a scattered set of templates, one that builds real skill and grows with us. It also gave everyone a common language and consistent expectations for responsible, accurate work.
Pairing the Academy modules with the prompt builders connects the why and the how, so people don’t just receive a prompt – they understand the thinking behind it and can adapt it to any situation. That is what makes AI a practical, everyday tool rather than an occasional experiment.
Since we built the Hub to evolve, we can add new lessons and builders as AI changes, which keeps it current. And building it for ourselves sharpened what we bring to clients through our AI Enablement services, from readiness workshops to prompt frameworks to ongoing consulting. It reflects how we approach AI in general. We start with what our team actually needs to do their best work, and we bring in the tools from there.
