Evenner

Framing the Problem

What is the problem you are trying to solve? Who does it affect?

Many event attendees struggle to find the most relevant sessions, build meaningful connections, and retain what they learned after the event ends. Most events focus on delivering content, but they do not provide a personalized journey before, during, and after the event.

This problem affects both attendees and organizers. Attendees feel overwhelmed by parallel tracks and packed schedules, while organizers lack real-time insight into participant engagement and long-term impact.

Idea Explanation

What is your idea? How does it fix the problem?

Evenner is an AI-powered event companion — an agentic event planner that accompanies the attendee before, during, and after an event.

Before the event, an AI onboarding interview learns who the user is: their profession, interests, goals, and preferred session formats. When an attendee joins an event, the AI generates a personalized agenda from the event's real schedule — an ordered selection of keynotes, sessions, and workshops, each with a one-line explanation of why it was picked. When the agenda leaves a long gap in the day, the platform fills it with a "micro-meetup": a short networking slot built around an interest the attendee shares with others.

During the event, attendees check into sessions from their live agenda. The app can capture the talk via on-device speech-to-text, and after each session the attendee rates it and leaves feedback — stored together with the transcript. Organizers follow a realtime dashboard that streams session check-ins and feedback as they happen.

After the event, attendees get a recap of everything they attended and rated, plus AI-generated follow-up recommendations — articles, repositories, and videos found via live web search, each tied to a session they actually attended and explained with a reason. Their ratings also feed back into the AI as taste signals for their next event's agenda.

This transforms events from one-time experiences into ongoing learning and networking journeys.

Implementation

How do all the pieces fit together? Does your frontend make requests to your backend? Where does your database fit in?

The frontend is a Next.js 14 application using KendoReact components and Tailwind CSS, with two role-based experiences: attendees get Events, a personal hub, and Profile; organizers get a console for creating events, building multi-day agendas, managing sponsors, and a dashboard with live stats (sessions running now, next up, registrations and capacity).

The backend uses Next.js Server Actions on top of Supabase: Postgres for data, Supabase Auth for sign-in (LinkedIn OpenID Connect with consent, or passwordless magic links), Row Level Security scoping every record to its owner, and Supabase Realtime for the organizer's live dashboard.

The database stores profiles, onboarding answers, events, agendas (keynotes, sessions, workshops, panels), sponsors, personal agendas, session check-ins, feedback with transcripts, and recommendations.

The AI layer is provider-agnostic: it works with Anthropic, OpenAI, or OpenRouter (selected via configuration), and degrades gracefully to a deterministic local heuristic when no API key is available — so the core experience never breaks. AI services are used to:

• Generate the personalized agenda with an explainable reason per session.

• Produce post-event recommendations using real-time web search with structured, validated output.

• Incorporate past session ratings as taste signals for future recommendations.

• Suggest interest-based micro-meetups to fill agenda gaps.

A core guardrail: every recommendation must cite the signal that drove it, and no user's interview answers are ever visible to another user (enforced at the database level).

Challenges

What did you struggle with? How did you overcome it?

One of the biggest challenges was balancing innovation with simplicity. We generated many feature ideas, but kept the experience focused by organizing the platform into three clear stages: pre-event personalization, in-event engagement, and post-event continuation.

A second challenge was making the AI trustworthy and resilient. We solved this with explainability (every pick comes with its reason) and a fallback chain — multiple LLM providers with a local heuristic backstop — so the product demos and functions even without API access.

Finally, capturing in-event signals without friction was hard: nobody fills in long forms at a conference. Lightweight check-ins, ambient speech-to-text capture, and one-tap ratings keep the cost of engagement near zero while still feeding rich data to the AI.

Accomplishments

What did you learn? What did you accomplish?

We built a working platform that:

• Personalizes event participation with an explainable, AI-generated agenda.

• Encourages active engagement instead of passive attendance through live check-ins, transcripts, and instant feedback.

• Creates networking opportunities via interest-based micro-meetups.

• Extends learning beyond the event with AI-curated follow-up resources.

• Gives organizers a realtime view of engagement as it happens.

The project demonstrates how events can become long-term learning journeys rather than one-day experiences.

Next Steps

What are the next steps for your project? How can you improve it?

The core product is functional end-to-end. Next, we plan to deepen each stage of the lifecycle:

• Interactive quizzes and challenges during sessions.

• Digital achievement badges and gamification rewards.

• Advanced AI matchmaking for one-to-one networking (beyond interest-based micro-meetups).

• Trend analysis and topic-level insights for organizers, built on the feedback and transcript data we already collect.

• "Future Note" reminders sent after the event.

• Community groups and post-event discussions.

Our long-term vision is a platform that keeps attendees connected, engaged, and learning long after the event has ended.