•September 22, 2025
Every quarter, our team unites in person for a few days of creativity and connection—our CCC Days, named after our core values: Character, Collaboration, and Craftsmanship. In a world where many of us work remotely, these days are more than just meetings; they’re a chance to spark cross-departmental partnerships and collaboration with colleagues we might not otherwise work with.
This quarter, our experiments were focused on AI enablement, exploring new tools, and improving workflows. Below are a few highlights from what we explored during CCC Days.
As part of our ongoing AI-focused experiments, we launched a bold experiment: Could a non-technical team build a functional internal dashboard tool using only natural language prompts and generative AI? The result is a standalone prototype of our Project Management (PM) Scorecard, created without traditional coding.
The PM Scorecard aggregates weekly project metrics from multiple data sources (task tracking, time tracking, and retrospectives) into a central dashboard. Historically managed through a complex and manually updated spreadsheet, the scorecard tracks key indicators like on-time delivery, roadmap confidence, team collaboration, and budget adherence.
Our MVP achieved automatic score calculations and a working interactive dashboard. While automated data collection remains unfinished, we proved the concept: lightweight AI tools can empower non-developers to build useful internal products quickly.
At RoleModel, we are always looking for ways to accelerate development by leveraging AI to streamline boilerplate code generation, enabling teams to focus on higher-order problems that will result in delivering customer value. Our goal was to cut down the time spent on initial codebase setup for a Minimum Viable Product (MVP) by using tools like GitHub Copilot to create entity relationship diagrams (ERDs), roadmaps, and user stories—transforming abstract plans into actionable steps.
Leveraging assets from our sales process, we generated a foundational roadmap with user stores, iterating on prompts to balance speed and accuracy. AI also helped us quickly scaffold views, forms, and navigation, while building tests, both model and system-level, that aligned with system goals.
Though this effort wasn’t without its challenges, we’ll continue to refine our prompts and examples as templates for future projects. This approach isn’t just about speed—it’s about creating a more complete and stable foundation for exploring complex problems earlier in the development cycle.
We explored building an AI agent in n8n, a self-hosted workflow automation platform, to streamline interactions with QuickBooks Online, aiming to allow our team to answer client-centric questions about invoicing, payments, and company performance.
Our key takeaway was the importance of structured data: preloading fixed datasets based on customer names for prompts allowed the tool to efficiently retrieve invoice and payment details for multiple clients in a single query. However, performance remained a hurdle, highlighting the need for optimization.
We learned that providing full context upfront, rather than expecting dynamic problem-solving, was critical for success. This experiment underscored the value of combining structured data, clear prompts, and thoughtful model integration to create a functional, albeit imperfect, AI-driven workflow for QuickBooks.
We’re pushing the boundaries of visual design with our new static Augmented Reality (AR) feature for Railing Designer, empowering users to overlay their 3D railings configurations within real-world photos. We’ve previously explored live AR; given the current state of hardware, we see value in a static version that allows for viewing the scene without being on-site.
This experiment demanded intricate math and iterative problem-solving, such as navigating complex research papers and refining trigonometric approaches to simplify calculations, but we made significant progress.
This innovation enables a seamless workflow: users capture a photo, load it into the app, and draw two sets of parallel lines to define horizontal and vertical surfaces. The 3D view then dynamically adjusts to match the image, enabling users to visualize railings as if they were part of the original photo.
This not only enhances the design experience but also strengthens Railing Designer’s ability to sell designs visually, helping customers better realize the value of their investments.
Managing design tokens between Figma components and CSS code has long been a source of friction, leading to errors, manual copy-pasting, and wasted time. To address this, our team embarked on a project to create a unified, centralized "source of truth" for creating and updating our Optics themes, aiming to streamline the process of defining and managing variables.
The goal was to build a data model for tokens based on the spec published by the Design Tokens Community Group that could seamlessly bridge design and development. Though the initial phase focused on defining the app’s structure and data model, challenging but critical, progress was made on UI for colors (including linked variables) and JSON exports to Figma.
Looking ahead, we’re prioritizing completing the UI for all token types, resolving Figma import issues, adding CSS export functionality, and deploying the tool publicly. This effort isn’t just about fixing a workflow—it’s about creating a foundation for greater consistency, collaboration, and agility as we continue to refine our design system and codebase.
Our team has been on a mission to elevate our design and collaboration workflows. After evaluating several alternatives to our current headless CMS, Butter, we think we’ve found a winner with Framer. Framer’s flexibility, robust collaboration tools, and seamless integration with design systems make it a game-changer for our goals.
To start, we overhauled our existing website to align with our new brand identity, refining navigation, card styles, button consistency, and spacing. Parallel to this, we launched a Framer CMS experiment, rebuilding navigation and footer systems from the ground up while testing its component structure and CMS capabilities.
The results? A powerful platform that blends Figma-like live editing with advanced collaboration features—like commenting, staging, and version history—making it a breeze for designers and non-designers alike. With Framer, we’ve unlocked faster iterations, deeper creative freedom, and a more intuitive workflow, especially as it evolves with features like custom page creation. The momentum from this CCC day experiment has us excited to push forward, ensuring our digital presence reflects the same excellence we bring to every project.
We are excited to continue to explore this opportunity and evolve the tools we use for us and for our partners.
This is our time to think bigger, experiment boldly, and push boundaries. From testing cutting-edge technologies to refining workflows that elevate our partnerships, CCC Days are where ideas take shape. It is also an opportunity for junior team members to lead, innovate, and grow. Whether we uncovered tangible improvements in our workflows or gained valuable insights to refine our approach, every effort contributed to our collective growth. We’re thrilled to see how these initiatives will continue to elevate our craft, empower our partners, and unlock new opportunities for their competitive advantage.
