Resolvia

• 

  Case journey

• 

  Case overview

• 

  Truth Graph

• 

  Resolution Readiness

Inspiration

Resolvia was inspired by a personal experience I had while trying to get a subscription refund. What should have been a simple process became difficult because I had to contact different parties, repeat the same information, and deal with inconsistent responses. I realized that many disputes become complicated not because the information does not exist, but because it is scattered across different systems and people. That experience inspired me to build Resolvia: a system that keeps evidence, context, and actions connected throughout the entire resolution process.

What it does

Resolvia is an AI-powered dispute resolution platform that helps investigate and resolve cases using evidence. When a dispute is created, Resolvia keeps it as a persistent case. The AI can collect evidence, review claims, identify inconsistencies, and help determine the next step. With WebMCP, Resolvia can also interact with supported web applications to retrieve information and perform structured actions. Important evidence and actions remain recorded so the final resolution is easier to understand and review.

How we built it

Resolvia is built around a case-based workflow. Each dispute has its own status, evidence, investigation history, and resolution. Instead of asking AI to immediately make a decision, the system guides the case through several stages such as evidence collection, investigation, and resolution. We use WebMCP to give the AI a structured way to interact with web applications, while keeping the process transparent and traceable.

Challenges we ran into

One of the main challenges was deciding how much control the AI should have. Disputes may involve important customer or financial decisions, so we did not want the AI to act without clear evidence. Another challenge was handling incomplete or conflicting information from different sources. We also wanted WebMCP to be a meaningful part of the product, not just something added for the competition. It needed to genuinely help Resolvia collect information and move cases forward.

Accomplishments that we're proud of

I am proud that Resolvia became more than a simple AI chatbot. Each dispute can keep its evidence, history, and progress over time, making the system more suitable for real-world resolution workflows. I am also proud that WebMCP fits naturally into Resolvia by allowing AI agents to interact with web applications while keeping important actions traceable.

What we learned

Building Resolvia taught me that many disputes are caused by fragmented information rather than a lack of information. I also learned that giving AI more abilities is not enough. A good agentic system also needs clear workflows, persistent context, evidence tracking, and human oversight. WebMCP showed me how websites can become easier and more reliable for AI agents to interact with through structured capabilities.

What's next for Resolvia

Next, I want to expand Resolvia beyond subscription refunds into other cases such as billing disputes, marketplace transactions, delivery problems, and service complaints. I also want to add more WebMCP integrations, improve evidence tracking, and introduce stronger human approval for sensitive actions. The long-term goal is for Resolvia to become a reliable resolution layer that helps users and businesses turn fragmented information into clear, evidence-backed resolutions.

Built With

• adk
• artifact
• build
• cloud
• fastify
• firestore
• gemini
• google
• identity
• manager
• next.js
• platform
• pub/sub
• registry
• run
• secret
• typescript
• webmcp