Pulsato

Inspiration

Employee pulse surveys often fail to capture how people truly feel. We've experienced this firsthand.

Traditional surveys rely on numerical self-assessments for wellbeing, workload, or motivation—usually on a scale from 0 to 10. These numbers lack context, do not account for emotions, and ultimately leave employees feeling unheard.

As a result, most people-related issues in companies are detected when it’s already too late. Organizations need a more natural and empathetic way to understand employee wellbeing before problems escalate.

What It Does

Pulsato is an AI voice agent that enables employees to have short, meaningful check-in conversations about their wellbeing.

How it works

1. Conversational Check-Ins – Employees engage in a natural conversation with Pulsato, which asks relevant questions, follows up where needed, and adapts based on the discussion.
   
2. Insight Extraction & Anonymization – Pulsato processes the conversation, extracts key insights, and generates an anonymized wellbeing report. Employees can review and approve their report before submitting it.
   
3. Company-Wide Analytics – HR teams receive an aggregated wellbeing dashboard, allowing them to track trends over time while maintaining full anonymity and employee privacy.
   

How We Built It

• Frontend: Developed using Lovable, a platform for AI-driven applications.
  
• Backend: Powered by Supabase, providing authentication, database, and storage functionalities.
  
• Voice AI: Integrated with ElevenLabs API to enable a realistic and engaging AI voice agent.
  
• Knowledge Base: Fine-tuned using Mistral, allowing Pulsato to ask intelligent, context-aware questions.
  

Challenges We Faced

1. Row-Level Security (RLS) in Supabase – Initially, we encountered several access control issues, leading us to disable RLS for the demo. We plan to implement a secure alternative.
   
2. Design Limitations in Lovable – Making UI adjustments and customizing interactions proved to be more rigid than expected.
   
3. Defining Use Cases for Voice AI – Without prior experience in voice-based AI applications, we had to spend considerable time refining the right conversation flows and use cases.
   

Accomplishments

• Successfully built our first AI voice agent within less than 10 hours of development time.
  
• Created an intuitive, anonymous, and privacy-first wellbeing tracking tool.
  
• Developed a structured process for integrating voice AI into business workflows.
  

What We Learned

• How to develop and deploy AI voice agents using ElevenLabs.
  
• Integrating conversational AI into a structured workflow using Lovable.
  
• Designing meaningful AI interactions that feel natural and engaging.
  

Next Steps

1. Pilot Testing in Our Organization – We are deploying Pulsato in our own company (50 full-time employees) to gather real-world feedback.
   
2. Improving Security and Data Handling – Implementing a robust Row-Level Security strategy to protect individual responses while maintaining anonymity.
   
3. Enhancing AI Conversations – Refining Pulsato’s conversational flow to better detect subtle emotional cues and context.
   
4. Advanced HR Dashboard Features – Expanding analytics capabilities to allow HR teams to detect patterns, track long-term trends, and measure the impact of interventions.
   
5. Expanding to Other Organizations – Once refined, we will introduce Pulsato to select external companies for broader testing.
   
6. Exploring Additional Integrations – Investigating integrations with Slack, Microsoft Teams, and other internal communication tools for seamless adoption.
   
7. Multilingual Support – Expanding Pulsato’s capabilities to support multiple languages, making it more accessible for diverse teams.
   

Our goal is to build a tool that not only collects wellbeing insights but also helps companies take proactive steps in improving employee satisfaction and mental health.

Built With

• elevenlabs
• github
• lovable
• mistral
• resent
• supabase