This experiment started after reading a post by Chris Pearson, who shared a clever way to visualize Salesforce release features in a dashboard format. His idea immediately resonated with me — clear, structured, and visual instead of the usual PDF marathon.
I loved the concept so much that I decided to take it a step further and build my own version: a more interactive, filterable, and expandable Salesforce Release Dashboard.
You can explore the live proof of concept here 👉 sfrelease.enree.co.
The Challenge with Salesforce Releases
Every few months, Salesforce drops another huge release — hundreds of new features, updates, and enhancements scattered across multiple products, clouds, and PDFs. Each one is full of value, but navigating it all can be an endurance test.
The documentation is detailed (and impressive), but when you’re trying to answer practical questions like “What’s new for Service Cloud admins?” or “Which GA features have a high business impact?” — finding answers means jumping between dozens of pages.
So, I wanted to see what would happen if we made exploration simpler, more visual, and more interactive.
Building the Salesforce Release Dashboard
The idea was straightforward: turn static release notes into a searchable, filterable experience where you can quickly find what’s relevant to you.
The dashboard loads structured JSON data for each release (for example, Winter ’26), merges multiple files, and displays features in a clean grid. From there, you can filter, sort, favorite, and export features — all from a browser.
Some highlights:
- 🔎 Advanced filtering – Multi-select filters for profiles, products, and clouds, plus full-text search.
- ⭐ Favorites – Mark the features that matter and toggle a “Show favorites only” view.
- 📈 Live stats – Automatically count total, GA, Beta, and high-impact features.
- 🪟 Detail modal – Each card expands into a rich modal showing full descriptions, benefits, tags, and impact scores.
- 📤 CSV export – Instantly export your filtered view for documentation or sharing.
It’s built with Node.js and Express on the backend, and plain HTML/CSS/JS on the frontend — no frameworks, just clarity and speed.
The Prompting Approach
A key part of this project was automating the collection of release data. Instead of manually extracting from PDFs, I used ChatGPT Deep Research prompts to parse and summarize Salesforce release notes — one cloud at a time.
Each prompt was designed to extract clean, structured fields: titles, feature types, impact levels, benefits, and tags. The output went straight into JSON files, organized per release.
The dashboard then merges and normalizes everything on load. This workflow can evolve a lot further — better prompts, validation against the Salesforce Help portal, and automated pipelines for new releases. But even in its current form, it drastically reduces manual effort while keeping data consistent and explorable.
Why This Matters
This isn’t only about Salesforce — it’s about how we can use AI-assisted structuring to make sense of massive documentation sets.
Every large platform has the same problem: too much valuable content, not enough structured discovery. Combining good prompting with a simple visual layer transforms that chaos into clarity.
That’s what excites me about this experiment: it sits right at the intersection of AI, data structuring, and developer experience.
Big thanks again to Chris Pearson for the initial spark. Sometimes all it takes is one inspiring post to turn “this is cool” into “I need to build this.” and to Cursor for the cool session of vibe coding.