🤯 What we are currently working on is insanely powerful...
🏗 We have been building, testing and fine tuning an AI agent on the first party public data of VC’s to understand their thesis. ⌗ A data-driven approach, based on a dynamic taxonomy, will instantly and accurately lift Founder <> Investor fit, scaling towards Investor <> Investor fit. With a middle layer of specific datasets from 1st party public data directing LLMs, founders can instantly qualify and connect with investors. Using VC public domain statements, we get nuanced insights into their strategies and behaviors and reduce the information asymmetry to level the playing field.
🥱 💻 A taxonomic approach
🥱 💻 We use a taxonomic approach for hierarchical categorization, interpolating possible interests from subcategories. Each node and its child nodes carry dynamic confidence weights adjusted through learnings and feedback. This setup captures the current landscape and absorbs innovation effectively.
✅ The great thing
✅ The great thing about the confidence level weightings is they are a direct function of public information and portfolio co. data. The less transparent, ambigus, fluffy the info is, the lower the confidence… transparency boosts confidence and attracts founder interest.
😀 Fun and insightful
😀 Some of the fun and insightful highlights to see are the gaps between what is said and how it reflects with actual actions. By comparing a thesis to global startup capital trends across geos and verticals, it’s easy to spot the trend followers, adventure capitalists, and the biases.
📈 Some analysis
📈 In an analysis on E-mobility after the spike on the West Coast, we saw startups globally funded by VCs with no previous exposure or interest. With near-zero interpolated interest from their thesis, these VCs jumped on the trend. How could a founder know to approach them?
⭐️This insight is gold, when major trends occur we know who might be influenced.
🤯 What is mindblowing is every newsletter, social media post, podcast, website/ blog update will be ingested, contextualised, analysed, scored, and applied, continually fine-tuning confidence levels through learning rom irrefutable first-party data.
✅ Data driven is the future
✅ Data driven will soon be the core of every introduction made IMO. Figuring out the 5 W’s and how just to get a meeting is hugely time consuming. Learning to rank models with LLMs allow us to 10x the qualification process and make relevant connections with VC’s who can directly do their Je ne sais quoi.
for the ❤️ of startups