AI agents that evaluate investment opportunities

About the project

RunwayFBU evaluates enormous amounts of data daily. From startups sending their pitch decks to receiving newsletters with new startups being seen in various channels. In order to focus their time on the best fitting companies, they needed a way to sort through the noise and find the startups that closely match their needs.

Client Challenges

With Artificial Intelligence growing fast, they wanted to jump on the train and take leverage of the capable models and implement a solution that solves their pain points. The main challenges in this project included:

1. Create a data researcher AI agent crawls the internet about the founders, gets financial data from various APIs and collects other critical information
2. Create an AI evaluator agent that scores startups based on custom metrics and what RunwayFBU are looking for in startups with a high confidence rate
3. Launch the prototype so it is easily adopted by their evaluations team

The role of Magebit
^{Where digital magic comes to life}

Magebit has been innovating with AI since 2016. In this project we were tasked to build a functioning AI agent system that drastically reduces the need of manual human labor so evaluation team can focus on startups that are closely matching what RunwayFBU is looking for. Our responsibilities included:

• Agentic workflow planning
  ‍Full strategy and AI-communication mapping of how AI agents will interact with each other as well as how they will interact with humans.
• AI design and development
  ‍Flows, data retrieval, evaluation and complete implementation of the project.
• Model training and calibration
  ‍To ensure the results are actionable and highly accurate, we trained and calibrated the AI agents to get meaningful results.
• Prompt engineering
  ‍As with any AI project, prompts, system messages and context are critical.

_HOMEPAGE