Inspiration
Fundraising is one of the most time-consuming challenges for early-stage founders. Finding relevant investors, understanding their investment preferences, writing personalized introductions, and tracking conversations across multiple investors can quickly become overwhelming. We wanted to build an AI associate that simplifies this process while keeping founders in control.
What it does
Raise acts as a fundraising and investor-relations assistant that founders can interact with through iMessage. It helps founders:
- Onboard their startup: Build a profile covering the startup's name, sector, stage, fundraising target, location, and traction.
- Find relevant investors: Research 20–30 real pre-seed angels and funds, with fit explanations, recent deals, and source links.
- Draft personalized introductions: Generate concise investor outreach emails with evidence-based reasons for investor fit.
- Track fundraising progress: Manage investor statuses from research and drafting to approval, contact, replies, or passes.
- Prepare investor updates: Draft monthly updates covering achievements, metrics, and requests.
Raise also supports multiplayer conversations, allowing co-founders and advisors to collaborate in the same thread.
How we built it
We built Raise on the Plow Hermes base (OpenClaw 2.0), extending its capabilities with a fundraising-specific persona and five skills: onboarding, investor research, introduction drafting, pipeline management, and investor updates.
The project uses Python for pipeline and profile state management, JSON for persistence, Docker for containerization, and Docker Compose for local deployment. The agent's persona and skill instructions define its responsibilities, workflows, and safety boundaries.
Challenges we ran into
One major challenge was designing an assistant that could provide useful investor research without inventing investors or unsupported claims. We addressed this by including source links and evidence-based fit explanations.
Another challenge was managing fundraising workflow states and enabling collaboration without compromising approval authority. We designed explicit pipeline stages and restricted draft approval to the startup owner or a co-founder.
Accomplishments that we're proud of
- Built an end-to-end fundraising workflow accessible through messaging.
- Combined investor discovery, personalized outreach drafting, pipeline tracking, and investor updates in one assistant.
- Designed a source-backed research process that prioritizes real investors and verifiable information.
- Implemented human-in-the-loop approval instead of autonomous email sending.
- Enabled founders, co-founders, and advisors to collaborate through multiplayer threads.
What we learned
We learned how to extend an existing agent framework with specialized personas and skills, structure persistent workflow state, containerize an AI agent, and design safety constraints for real-world business automation. We also learned that trustworthy AI requires clear approval boundaries and evidence-backed outputs, not just automation.
What's next for Raise?
We plan to improve investor-fit ranking, strengthen research verification, add richer fundraising analytics, and enhance pipeline insights. Future iterations could also support follow-up recommendations, investor relationship history, and progress summaries while preserving human approval for external communications.
Log in or sign up for Devpost to join the conversation.