AI, Data and the Death of Investing From the Gut: Inside TRAC.vc’s Model

By Alastair Goldfisher
Veteran journalist and creator of The Venture Lens newsletter and The Venture Variety Show podcast. Alastair covers the intersection of AI, startups, and storytelling with over 30 years of experience reporting on venture capital and emerging technologies.

Khaled Kteily, founder and CEO of Legacy, is no stranger to the world of venture capital. Over the years, his male fertility startup has raised $50 million from investors, including Bain Capital Ventures, FirstMark Capital and celebrity backers like Justin Bieber and The Weeknd.

But when TRAC.vc reached out to him in a cold call last year, it wasn’t just another investor pitching to get a slice of the action. Venture capital has been undergoing a data revolution, and a new breed of firms like TRAC represent this “Moneyball” moment where data and algorithms are increasingly driving investment decisions.

“What impressed me most was the transparency and efficiency,” Kteily said. “TRAC made it clear that their algorithm dictated their investment decisions, and their process was seamless.”

Data is the Investment Committee

TRAC.vc was founded in 2020 by Joe Aaron, 75, who previously ran hedge funds, and Fred Campbell, 69, who was formerly an entrepreneur and most recently co-founded CrowdSmart.

TRAC describes itself as an AI-driven venture firm, and what sets TRAC apart is that it is built entirely on data. The firm, based in Sonoma, California, uses predictive AI technology to pore over market and third-party data sources. Once TRAC flags a company, they reach out directly to founders with the intent of investing.

Kteily said that he was surprised when TRAC cold-contacted him and identified Legacy as having a 20% probability of becoming a unicorn within five years, and this was before he even started actively fundraising the next round.

“We cast a much wider net than a traditional VC working off their network,” Campbell said. “If a startup has traction, media mentions or funding activity, we’ll find it.”

That net is cast with machine precision. Campbell and Aaron said that TRAC’s AI-powered system enables a more efficient funding process, allowing the firm to quickly and objectively find deals, invest and cut a check. Campbell, who boasts of his sales ability to successfully close the deal when he gets an entrepreneur on the phone, doesn’t have final say over where TRAC invests. He recalled a moment when he wanted to back a particular startup, but the algorithm said no chance.

“It was frustrating, but the data showed gaps that I hadn’t considered,” he said. “We’re not in the business of discovering hidden gems. We’re here to make sure we never miss a unicorn.”

Campbell’s experience highlights a truth about AI-driven investing: No matter how much AI powers the process, human judgment still often plays a role, whether through data interpretation, deal structuring or the final decision.

Many firms using AI have their own “wizard behind the curtain,” whether it’s a partner overriding a decision, a data scientist tweaking inputs or an unseen behavioral bias shaping the outcomes.

How TRAC gets into deals 

TRAC has raised two main funds, with its AUM exceeding $80 million. It has invested in 105 unique companies, overwhelmingly early-stage deals in such sectors as deeptech, fintech, medtech, gaming, AI and software, among others.

TRAC claims strong performance across its funds, with its second fund ranking among the top performers in its category, according to third-party data cited in a Fast Company article. If accurate, it suggests that TRAC’s AI-driven methodology is achieving results that rival traditional VC models.

TRAC’s portfolio includes Human Interest, which operates a digital retirement benefits platform and which reached unicorn status in 2021. Several other TRAC portfolio companies are reportedly not far behind Human Interest in terms of nearing billion-dollar valuations.

Of course, identifying high-potential startups is one thing. Getting into those deals is another challenge entirely. A spokesperson from another firm underscored this when they noted that 40% of their potential deals come from AI sourcing, but fewer than that are converted to deals. If TRAC has invested in 105 unique companies in almost 5 full years, it is noteworthy they’re getting into so many deals, especially given the competitive nature of rounds.

For seed-stage startups that it has identified as promising, TRAC has little difficulty securing meetings. Once TRAC connects with founders, the firm’s close rate with seed companies is an impressive 90%. But when a company is at Series A and B stages, where rounds are more competitive, TRAC faces a different challenge.

Only about 20% of the “hottest” founders agree to a call or a Zoom meeting with TRAC. However, once they do, TRAC almost always secures an allocation. From mid-2023 through the next 14 months, Campbell said he personally closed every single deal after getting a founder on a call.

So, what makes TRAC so compelling once they get in the room? Campbell says the firm’s edge is the TRAC Intelligence Dashboard, a customized tool they make available exclusively to their portfolio founders. The dashboard provides market insights, benchmarking data and competitive analysis, research tools that typically could cost companies upwards of $200,000 per year if sourced independently.

“Founders can easily see that with TRAC, they get valuable data and intelligence they cannot get elsewhere, or certainly not without incurring great costs and time,” Campbell said.

Kteily agreed that TRAC’s compelling value extends beyond the capital. “They provide tools that help me track competitors, understand market valuations and target future investors,” Kteily said.

TRAC’s allocation strategy also works in its favor. The firm typically requests 10% of a round, and never more than 20%. While they don’t always receive their full requested allocation, their minimum check size is $250,000, making it relatively straightforward for founders and lead investors to fit TRAC in on the cap table.

And after five years and more than 100 portfolio companies, the firm is now leveraging its built network of co-founders and executives from its portfolio to help secure the next wave of warm intros with potential targets, something that wasn’t possible when TRAC first launched.

What founders say about TRAC

Sunil Thomas, Chief Business Officer of Zendar, had a similar experience to Kteily when TRAC reached out to him cold in 2022. The firm used its AI-powered database to identify Zendar, which develops radar-based sensors for autonomous vehicles.

Initially skeptical, Thomas said he quickly realized TRAC’s data-driven insights were invaluable for their fundraising process. Zendar, which came out of Y Combinator, has now raised $55 million to date.

“They [TRAC] provided valuation benchmarks that helped us negotiate terms confidently, even in a challenging market,” Thomas said.

AI in venture is evolving

Data-powered investing isn’t anything new to VC. Firms like Correlation Ventures, SignalFire and EQT Ventures (with its Motherbrain platform) use AI and data analytics to help source potential deals. This 2019 Forbes article highlighted several adopters of data-based investing. I’d argue that today, every early-stage venture firm incorporates some level of data analysis to score investments, track market trends, support portfolio companies or identify exit opportunities.

“This is the natural evolution of venture capital,” Campbell said of TRAC.vc. “Historically, firms made decisions based on networks and intuition. Now, we have access to so much data that it would be negligent not to use it.”

However, if AI is a universal disruptor across venture, especially in early-stage investing, there are downsides. AI struggles when the data is sparse. Consider, for example, how many pre-seed and SAFE rounds are not reported. Pre-seed companies may not yet be on the radar of data-crunching analysts even if they have proprietary AI-driven computer systems to turn over every digital stone.

In addition, the AI-driven approach raises questions of accessibility and equity. While algorithms might aim to eliminate human biases in investment decisions, they can also reinforce systemic disparities. Underrepresented founders who have less digital visibility, fewer connections to established networks, or operate in communities that receive less media coverage may be overlooked by data-scraping algorithms.

Startups led by female founders or people of color historically receive less press coverage and appear in fewer databases, potentially creating blind spots for AI systems like TRAC’s. I addressed this question to a proponent of AI, and they acknowledged that their system requires founders to have some existing digital footprint, but they argued that the AI approach widens the aperture beyond traditional referral networks in Silicon Valley. This, they said, potentially helps to uncover talent outside the usual Sand Hill Road radar.

Also, AI in venture isn’t a one-size-fits-all approach. Early-stage firms using AI leverage more alternative data sources, while later-stage firms have more structured and verified information to analyze. But that raises another question. At what stage does AI-driven sourcing become a competitive advantage instead of just another tool in an investor’s quiver?

Every firm approaches AI differently 

More VCs are experimenting with AI to enhance deal sourcing, diligence and founder support. Another firm that has developed an AI-driven tool for investing is Flybridge Capital, which this year launched an AI-powered investment memo generator to analyze investment memos and help founders refine their pitches.

“We built the tool to help our team evaluate deals more efficiently, but it also has potential benefits for founders,” said Daniel Porras Reyes, an associate at Flybridge. “Imagine being able to submit your data and receive an AI-generated feedback loop, showing how investors might evaluate your company, potential red flags and areas to improve.”

Flybridge’s tool is open sourced, and Reyes said that other investors are accessing the source code, experimenting with how they can adopt it for their own uses.

Kanu Gulati, a partner at Khosla Ventures who led the firm’s investment in Zendar, sees AI as an invaluable tool in venture. But she said that human judgment remains critical.

“We use AI to help us spot trends and analyze markets, but investing is still a deeply human process,” Gulati said. “At the end of the day, relationships, conviction and experience still matter.”

To some, that represents a potential downside of going all in on AI: the risk of losing the human element. Even firms that say they’re fully in support of AI efforts often have a human factor, that wizard behind the curtain who influences the final investment decision. Even at TRAC, it takes a human like Campbell to call a founder and close the deal.

While AI offers numerous benefits, one concern some expressed to me is the reliance on quantitative data, which may overlook such qualitative factors as a founder’s potential, a resilient entrepreneurial mindset or a passionate vision. I wonder if an algorithm can truly replace the “gut feeling” that historically guided a VC’s investment. Is there a risk of potentially missing out on truly groundbreaking companies that don’t neatly fit into pre-defined data categories?

SuperTRACers: Predicting Investor Success 

While the debate continues about whether algorithms can fully replace human intuition in identifying promising startups, TRAC has developed an interesting solution that merges the two worlds. Rather than relying solely on startup metrics, they’ve also turned their AI lens to the investors themselves, creating a data-driven approach to identify what could be called “the smart money.”

TRAC’s AI-driven model also analyzes the investors backing the unicorns. This is where the firm’s predictive analytics take a fascinating turn. Instead of just identifying high-potential founders, TRAC also tracks high-performing investors, whom they call SuperTRACers.

One of TRAC’s key predictive models revolves around identifying these SuperTRACers, a select group of investors with a track record of success. The firm has found that 60% of US-based unicorns have a SuperTRACer in their cap table.

TRAC’s co-founder Aaron demoed their database to me and explained that their algorithm prioritizes startups with multiple SuperTRACers. A SuperTRACer’s involvement increases the probability of a startup’s success. Aaron said that Human Interest, for instance, has 58 total investors, 21 of which are considered top tier. Aaron also pointed to Mark Andreessen as an example of a SuperTRACers, noting that TRAC’s data shows he has a 46x return multiple and has backed at least 18 unicorns.

The Economist recently highlighted TRAC’s methodology and ranked some of Silicon Valley’s sharpest talent spotters. The article reported that TRAC’s data identified a select group of more than 250 early-stage investors—the SuperTRACers—who have generated outsized returns. Among them is Sara Thomas Deshpande, a general partner at Maven Ventures and one of only about a dozen female GPs on TRAC’s list.

Deshpande, who invests in consumer and health applications of AI, posted on LinkedIn that TRAC’s data validates what her firm already strives for: identifying founders positioned for hypergrowth and steering them toward their next funding round.

The Future of AI and Investing

Another question I had when speaking with TRAC is whether more firms will follow their lead of going all in on AI, or is relationship-driven investing going to continue to play a role in venture? As AI continues to reshape venture capital, the future of investing will likely mature into a hybrid, somewhere between relying on some instinct feeling and the fully automated algorithmic decision-making we see increasingly today.

“Ultimately, venture capital is still about relationships,” Campbell said. “But if we can use AI to make smarter decisions and support founders more effectively, everyone wins.”

For Kteily and Thomas, they said that the results of the TRAC’s AI-driven approach speak for themselves. “TRAC isn’t just an investor,” Kteily said. “They’re a partner in every sense of the word.”

The venture world is evolving, and firms like TRAC are leading a shift toward data-driven decision-making. Whether AI becomes the dominant force in venture capital or simply an assistive tool, one thing is certain: the industry will never look the same again.

Final takeaway for founders

As more VCs embrace AI, founders need to adapt. Here are some steps to follow:

• Ask VCs questions: Just assume that VCs everywhere and at every stage are using AI. Ask firms how they use the data for their sourcing, due diligence and portfolio company support.

• Manage your data: AI-driven investors rely on available data and metrics. Inaccurate or misleading fundraising data can thus harm startups and possibly exclude them from future deals. So make sure that databases, such as PitchBook and Crunchbase, have your accurate info. And manage that data so it remains clean and error-free.

• Storytelling still matters: AI might put you on a VC’s radar, but landing an investment still depends on a compelling narrative. This is where marketing and media trainers and content specialists, like yours truly, come into play to help you hone your story.

• Leverage the AI tools for yourself: Flybridge’s memo generator and TRAC’s valuation models provide founders with methods to assess their fundraising and market potential before meeting with investors.

🎙️ Be sure to follow Alastair - he has an upcoming podcast on The Venture Variety Show with TRAC’s Fred Campbell, as we talk about AI-driven investing.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.