Building the AI Super Connector with Andrew D’Souza, Founder of Boardy and Clearco

The best startup introductions happen when someone who truly knows you connects you with exactly the right person at exactly the right moment. But that magic relies on human memory, limited networks, and sheer luck. Andrew D'Souza thinks AI can do better.

D'Souza, founder of fintech unicorn Clearco and now CEO of Boardy, is building what he calls an "AI super connector" that optimizes startup fundraising introductions. His central thesis cuts against Silicon Valley's obsession with scale: the best AI applications don't maximize matches or minimize friction. They maximize goodwill.

"Boardy is when you think about any unsupervised learning system, you're trying to optimize an objective function or a cost function. For Boardy, that is goodwill," D'Souza explains. "How much goodwill can I actually accumulate across my entire network? And can I make intelligent bets in the introductions that I make that will lead to good conversations and economic value and good outcomes of deals such that I continue to expand the goodwill."

This framing inverts how most people think about AI-powered matching. Dating apps optimize for engagement. Job platforms optimize for applications submitted. Professional networks optimize for connections made. D'Souza argues these metrics miss the point entirely. Goodwill is the only currency that matters in venture capital introductions, because a bad introduction destroys trust faster than a good one builds it.

The technical challenge becomes fascinating once you accept this premise. Traditional databases reduce people to searchable fields: location, industry, funding stage, check size. But chemistry between founders and investors operates in what D'Souza calls "multidimensional space." Great human connectors intuitively understand this complexity.

"Software technology and databases have been used to reduce the dimensionality of what a human being is," he notes. "What our goal is to actually expand that dimensionality and capture in. Like, really good connectors do that. They hold you in this sort of multidimensional space. Very charismatic people, they ask you really interesting questions. They can kind of understand you as a complete 360 degree human."

This explains why scraping LinkedIn profiles and Crunchbase entries fails spectacularly for venture introductions. The data points that predict successful founder-investor relationships live in conversation, not in structured databases. When D'Souza talks to founders, he captures "everything from a founder's current fundraising status, sector, and location to their personal values, operating style, and even the kind of people they wanna meet." For investors, he tracks "what investors are actively looking for, their recent deals, and their appetite for new opportunity."

The key insight is that AI excels at pattern recognition across high-dimensional data once you feed it the right inputs. A twenty-minute conversation with a founder reveals their unique positioning, what drives them personally, and why they believe they're the only person who could build their specific company right now. A similar conversation with an investor uncovers their worldview, investment thesis, and the subtle preferences that determine whether they'll take a meeting.

D'Souza learned this lesson scaling Clearco to nine figures in revenue. The company provided growth capital to e-commerce businesses, which meant constantly matching entrepreneurs with capital sources. The difference between a successful introduction and a wasted one often came down to factors you couldn't find in any database: Does the founder's communication style match the investor's decision-making process? Do their timelines align? Are their risk tolerances compatible?

The measurement challenge becomes clear once you optimize for goodwill instead of volume. D'Souza tracks "match acceptance rates when Boardy proposes an introduction" as his core metric. This number has to stay high or people stop responding to introduction requests entirely. Every bad match destroys the platform's credibility with both sides of the network.

This creates a fascinating constraint: AI systems typically improve through iteration and feedback loops. But in venture introductions, you can't afford many misses. The cost of false positives is relationship damage that compounds over time. D'Souza essentially has to build an AI system that works correctly from the start, then gets better through careful calibration rather than trial and error.

The broader implications extend beyond fundraising. Most professional networking platforms treat connections as binary events that either happen or don't. But D'Souza's approach suggests that AI's real opportunity lies in understanding and optimizing the quality of human relationships over time. The goal isn't to facilitate more introductions. It's to facilitate better ones.

For founders currently fundraising, this suggests a different approach to networking. Instead of optimizing for meeting volume or trying to game algorithmic matching systems, focus on having substantive conversations with the connectors in your network. The people who can help you most are those who understand you in multiple dimensions, not just your pitch deck and traction metrics.

The real test will be whether Boardy can actually capture and operationalize the intuition that makes great human connectors so valuable. If D'Souza is right, the future of professional networking won't be about connecting more people more efficiently. It will be about connecting the right people in ways that build lasting goodwill across entire ecosystems.