IDC Season 10: Six AI Products Built from the Ground Up

IDC Season 10 set a new record for Accedia's internal innovation program. The season ran from November 2025 to March 2026. Over four months, 21 participants from across the company, spanning software and DevOps engineers, business analysts, and UI/UX designers, joined the IDC challenge to build six AI-powered products. The requirement this season was concrete: every team had to develop a real AI model, trained on prepared data. A dedicated mentorship team with hands-on product development experience supported them throughout. Accedia invested €9,000 in rewards, shared across the participating teams.

Here is what the six teams built:

• AImvest: a personal finance assistant that retrieves user data in real time and generates personalized budgeting recommendations through agentic workflows.
• ParAI: a parenting app offering evidence-based guidance from pregnancy through graduation, with an AI assistant grounded in AAP, WHO, and CDC medical guidelines. ParAI is already available in the App Store.
• SmartCharge: a SaaS platform for managing EV charging in parking and garage communities, using AI-driven prioritization to allocate capacity and detect anomalies in real time.
• StoryForge: a tool that converts meeting transcripts into sprint-ready Agile work items automatically, with a human review step built into the process to keep outputs accurate and accountable.
• VeriScript: a translation agency management platform that handles the full document lifecycle, from OCR and translation to quality checks, legalization, and delivery, with AI embedded at each stage.
• WildFire NowCast: a free, open-source wildfire intelligence platform that combines satellite detections, weather forecasts, and terrain data into one live map, updated every hour. No account required.

Across the six projects, teams worked with some of the most in-demand technologies in AI development today. Retrieval-Augmented Generation (RAG) featured in multiple products as the architecture of choice for grounding AI outputs in real, structured data. Agentic workflows, where AI takes sequential reasoning steps rather than returning a single response, appeared across several solutions. On the model side, teams used GPT-4o and GPT-4.1 Mini via Azure AI Foundry, as well as Google Vertex AI and the Gemini API, reflecting the multi-provider reality of production AI work. One team also implemented on-device machine learning with TensorFlow Lite for instant, offline intent recognition.

About Accedia Innovation Development Center (IDC)

IDC is a skill-sharing initiative that supports the realization of team members' innovative ideas into real software solutions. The program celebrates individual talent, entrepreneurial thinking, and cross-discipline collaboration. Interested in taking part in a future season? Browse our open positions.