PurposeEra

The Problem

When people have an ambitious idea, the first question is often: “Who do I know?”

But your existing network isn't necessarily the network your mission needs.

A founder might know engineers, but be missing user research. A student might have a great product idea, but have no idea which roles, capabilities, or collaborators are actually needed to make it real.

We wanted to change the starting point.

Instead of asking “Who do I know?”, PURPOSERA asks:

“Who should exist around this mission?”

What We Built

PURPOSERA is a mission-first network that turns an unstructured idea into an actionable map.

You describe what you want to build. PURPOSERA then identifies:

The problem and outcome behind the mission
The capabilities required to move it forward
The roles that could provide those capabilities
People in the network whose skills match those needs
A visual map showing how everything connects
A concrete first move worth making this week

The goal is to move from a vague idea to a clear understanding of what is missing, who could help, and what to do next.

How We Built It

PURPOSERA is built as a full-stack web application using Next.js, TypeScript, React, and React Flow.

The analysis pipeline runs server-side through the Gemini API. A mission is sent through a structured analysis pipeline, where the model identifies the capabilities and roles required. The resulting structured data is normalized and transformed into the interactive network graph.

We also built a fallback keyword-based analyser so the core experience can continue working even when the AI provider is unavailable.

The network matching layer compares the capabilities required by a mission against the skills represented in the network's profiles.

What We Learned

One of the biggest things we learned was that AI is more useful when it produces structured, actionable outputs rather than simply generating a block of text.

By turning the model's output into typed data, we could use it to drive an interactive experience rather than presenting users with another chatbot response.

We also learned how important it is to design the AI experience around the user's next action. Identifying what someone needs is useful; identifying the first thing they should actually do is much more valuable.

Challenges

One of our main challenges was making the AI output reliable enough to power a visual interface.

A free-form model response can be unpredictable, so we designed the system around structured output, normalization, validation, timeouts, retries, and a fallback analyser.

Another challenge was making the network feel purposeful rather than like another directory of contacts. That's why the matching starts with the mission's required capabilities and works backwards toward the people who could help.

What's Next

The current network uses seeded sample profiles to demonstrate the interaction model.

The next step would be connecting PURPOSERA to a real network, allowing people to create profiles, verify their skills, discover missions that need their capabilities, and turn matches into real collaborations.

Ultimately, we want PURPOSERA to make building ambitious things less dependent on who you already happen to know.

Don't start with who you know. Start with what your mission needs.

** NO DIVERGENCE LOG WAS SUBMITTED BECAUSE NO FEATURES WERE CHANGED**