AI-driven automation is reshaping how enterprises operate — unlocking opportunities at the intersection of innovation and efficiency.
Your mission: Use Beam’s no-code AI Agent platform to create a transformative solution for PedalWorks, a fictional bicycle manufacturing company.
- Target a workflow or department at PedalWorks where intelligent agents could drive massive value.
- Design and prototype AI agents that automate, streamline, or reinvent real business processes.
- Explore new modalities — integrate voice for natural interaction or use Beam’s APIs to build custom interfaces with tools like Loveable.
We'll provide platform access and starter examples — you bring the ideas.
Think you’ve got what it takes? Show us your vision at the hackathon. 🧠⚡
Insert Deep Dive Slides Here
PedalWorks is a mid-sized bike manufacturer operating across four locations with approximately 500 employees. The company spans six core departments:
- Procurement
- Manufacturing
- Quality Control
- Supply Chain
- HR & Onboarding
- Customer Support
Each team struggles with repetitive, low-value tasks that hamper productivity, increase errors, and reduce strategic focus.
Pick a real-world problem in any department, and demonstrate how your agent can:
- Automate with Intelligence: Deliver a working, end-to-end workflow powered by Beam.
- Deliver Tangible Impact: Quantify gains — hours saved, errors reduced, or satisfaction improved.
- Integrate Seamlessly: Show how your agent fits into existing tools and processes.
| Department | Department Role | Jobs to Be Done |
|---|---|---|
| Procurement | Handles the sourcing and onboarding of parts and materials (frames, tires, gearsets) from global vendors. Manages contract negotiations, PO generation, invoice matching, and payment processing for over 200 suppliers. | - Seamlessly onboard and verify new vendors without email back-and-forth. - Automatically match POs to invoices to minimize errors and manual checks. - Ensure timely supplier payments to maintain trust and avoid production delays. |
| Manufacturing | Runs high-precision bike assembly lines across four facilities. Teams manage production scheduling, coordinate component availability, and respond to equipment issues to meet shifting demand cycles. | - Detect and act on early maintenance warnings to prevent unplanned downtime. - Recalculate assembly schedules instantly when orders spike or change. - Dynamically allocate labor based on real-time line status and order backlog. |
| Quality Control | Inspects frames, welds, brakes, and electronic components across production lines to ensure every bike meets safety and design standards. Logs defects, flags issues to engineering, and approves final releases. | - Streamline inspection and defect logging to reduce product release delays. - Unify defect data to surface recurring issues and enable faster root-cause fixes. |
| Supply Chain | Manages the movement of bike components and finished products between warehouses, suppliers, and retailers. Uses ERP tools but struggles with visibility across multiple systems. | - Maintain real-time visibility into inventory levels at each location. - Use consolidated data to predict shortages and optimize restocking. |
| HR & Onboarding | Oversees hiring for seasonal production surges and ensures onboarding of factory staff, engineers, and support agents. Tracks training compliance and manages employee documentation. | - Digitize onboarding to get new hires on the floor within 24–48 hours. - Centralize training and certification records for easy compliance tracking. - Automatically screen and rank applicants to reduce recruiter workload. |
| Customer Support | Handles thousands of support tickets annually including warranty claims, service requests, and parts replacements. Agents work across email, spreadsheets, and CRM tools to resolve issues. | - Auto-extract and structure claim details from emails and forms. - Present agents with full customer history without switching tools. - Track warranty claims in one place to reduce response time and errors. |
You'll have access to curated datasets, including:
- Source-to-Pay data from Celonis
- Order Processing data from Beam
Want to solve a new problem? Go for it! Use our:
- Dataset Generation Helper – Upload a single example (PDF, JSON, etc.), and generate multiple variations for testing your agent effectively.
You’ll get full access to:
-
Beam AI Platform – A powerful low-code agent builder, featuring:
- Flow Builder: Map out workflows visually.
- Integration & Tool Builder: Hook into APIs and external systems.
- Tool Tuner: Improve accuracy and performance.
- Human-in-the-Loop: Add approval or feedback steps.
- Agent Analytics: Track behavior, performance, and usage.
-
Loveable Frontend Builder (credits provided) for custom UI design.
-
Beam APIs to embed agents into web/mobile apps.
- API docs and sample apps will be shared for guidance.
Your submission will be evaluated across the following five dimensions. Winning teams will excel not only in innovation, but also in execution, impact, and storytelling.
| Criteria | What We’re Looking For |
|---|---|
| 1. Use Case Relevance | How well does the solution align with real-world challenges faced by PedalWorks? Did the team pick a meaningful problem and clearly articulate the pain point? |
| 2. Workflow Design & Agent Logic | Does the agent demonstrate a thoughtful, end-to-end process? Are flows well-structured, modular, and effectively using Beam’s platform features? |
| 3. Technical Integration | How seamlessly does the solution integrate with other tools, datasets, or UIs (via Beam APIs or Loveable)? Did the team go beyond Beam basics to build robust systems? |
| 4. Business & Operational Impact | What measurable value could this solution deliver (e.g., time saved, errors reduced, faster onboarding, better forecasting)? Is the impact clearly demonstrated or well-reasoned? |
| 5. Presentation & Storytelling | Was the final demo compelling, clear, and easy to follow? Did the team communicate their agent’s purpose, design, and impact effectively — both visually and verbally? |
Earn extra credit for:
- Creative use of voice, natural language interfaces, or multimodal interactions
- Innovative use of the Dataset Generator or augmentation techniques
- Embedding the agent into a polished UI using Loveable or Beam APIs
Questions? Reach out to Burak, Mujtaba, or Aqib during the Deep Dive — or ping us anytime on Discord!
- JBL Charge 5 Bluetooth Speaker
- 1:1 Mentorship Session with Beam Leadership