Over the past two decades more and more investing decisions have been made, in part, by machines. With the newest generation of AI-tools upending every area of business, the industry faces a new inflection point – one that opens a world of opportunities, and the world itself. Gartner has estimated that AI and data analytics will inform more than three-quarters of venture capital investments by 2025.
With AI tools that can work seamlessly alongside human analysts, VCs, Seed Funds and other early stage investors have the chance to significantly reduce risk at every stage of the investment journey, whether that’s to identify new startups, assess the performance of later-sage businesses eyeing an exit or IPO or hedge against currency fluctuations overseas when raising a fund.
We examine what the new playing field looks like, and what will set the winners apart.
Finding the signal in the noise
“We're at a stage right now where you can upload a deck into an AI system and it generates a summary and determines if it aligns with your investment thesis. The AI can even identify your investment thesis, potentially before you've fully realized it yourself,” explains Ashley Groves, CEO and Founder of Deaglo.
“This technology not only clarifies your investment strategy but also guides you in finding suitable candidates within the ecosystem for lending or investment opportunities for your firm at whever stage you’re active."
Every day, investment analysts face a dizzying volume of sell side research, earnings transcripts and pitch decks. One of the most immediately obvious and applicable uses of the newest generation of AI tools is to turn this morass of financial boasting into usable insights.
Not only does this improve the accuracy and efficiency of financial decision making, but it also expands the range of data without impacting the performance of analysts, enabling measures such as web traffic and new users. Academic research has already found that ML models can outperform the average VC by between 25% and 29%.
As reported by the FT, these models are already in use at firms, including KPMG, PitchBook, Headline and Moonfire Ventures. According to research by DDVC, penetration is currently highest as early stage firms, for sourcing and screening, followed by portfolio value creation and due diligence
"We've built various models to enhance our investment processes, including a startup grading model for prioritizing companies aligned with our thesis, company industry classification for benchmarking and targeted vertical segments, and matchmaking to connect companies with our network." says Jan Arp, Founding Managing Partner at The Holt Xchange.
This can be even more valuable in later stages, where decisions become more complex and time consuming.
“The longer a business has been running, the larger the volume of data to review. AI systems can drastically streamline the due diligence process, turning the pile of information from a headache into a chance to make real benchmarks among peers and competitors.” says Ashley.
With AI tools now built to be used by anyone, the ability to teach and tailor these models to an individual firm's investment thesis, risk model and customer profile will be a new core competency for analysts – so what does this mean for the humans working with the machines?
De-risking the personal
Human-risk is one of the most difficult to quantify elements for investors – and potentially the most volatile.
While AI has a clear advantage and application when it comes to data and analysis, the personal element of investing in making judgements on character and ethics is still, for many, an essential element. So what happens when AI’s data-driven ruthlessness comes up against an investor’s gut?
“If anything, AI is complimentary as it helps us to better prioritize our efforts”, says Jan. “By building models from the start, it forces us to be very systematic about our venture processes, and further unpack our 'gut' instinct, so that we may better gauge our decisions going forward for us all to learn from thereafter.”
This can include using objective elements to benchmark founding teams’ experience and network relative to their current endeavor and using data to understand the people themselves.
“We’ve seen that lending based on psychometrics tests produces a 40% gini score– sufficient for lenders to lend against. At the very least regarding psychometric testing for venture investing, you learn more about the personality and character of the founder in order to improve your communication.” says Jan.
For larger businesses, this can also be applied at the workforce level, looking at collaboration between teams, analyzing communications data and drilling into historic compliance and financial management.
Expanding investment horizons
As we have covered elsewhere, the range of data points, local knowledge and sector-specific considerations in foreign investing can quickly spiral out of control, blurring the risk profile for GPs and LPs.
In its work managing FX hedging and risk for funds and investors, Deaglo has worked as a human-technological interface between domestic capital and foreign opportunities. AI is already playing a key role in widening the scope of these opportunities while reducing volatility risk, especially for long-hold funds facing turbulent currency markets.
"Imagine a Brazilian VC seeking to raise capital in the U.S. and invest in Brazil and Mexico. By using LLMs, we gain deeper insights into their specific needs, whether it's raising funds or deploying capital,” says Ashley.
“We analyze their investment stage, fundraising goals, and compare onshore versus offshore returns. This technology enables us to tailor the investment journey, providing VCs with crucial insights and strategies, ensuring they're fully informed and prepared for every aspect of their global investment venture."
In the world of currency hedging and FX management, a complex task that GPs have long skirted, AI not only enhances the analysis of historic trends and future strategies, but plays a key role in interpreting the results of sophisticated risk models for a non-specialist audience.
Where does the race go next?
While investment funds have traditionally been at the vanguard of adoption – armed with both the means and the motivation to grab every advantage – this latest development has been far more democratic. With everyone from college students to grandparents jumping on the AI bandwagon, the genie is truly out of the bottle.
Already, one can see a correlation between the size of the engineering team and the AUM of VC firms, with larger funds having a higher average number of engineers, signaling the potential for a widening gap.
Currently, only 1% of VC firms have internal data-driven initiatives, but 84% of VC firms express a desire to increase their data-driven efforts. "AI's integration in the investment journey is inevitable; it's a case of use it or get left behind." says Arp.
We’re likely to see AI both expand the opportunities available and concentrate attention on those that can go the distance, while reducing risk for those firms that can combine their models with robust theses and management However, in a world where machines choose the players, it may be those who go outside the expected that become the true outliers.
Even as AI supercharges the game, the winners, as always, will be those no one saw coming. But with the right approach, a clear eyed, AI-assisted view of risk can transform balance sheets.