How AI helps evaluate investment decisions at Hustle Fund
At Hustle Fund, we review over 1,000 deals per month. Across our three funds, we currently support over 500 portfolio companies. And, we invest in 250 companies per fund.
The most common question people ask our investment team is some version of: How? How can you process so many deals? How is it possible to support such a large portfolio? How do you make time to sleep?
The answer is leaning into what we’ve done our entire careers, which is build software. All our investment team members are former founders and senior operators. We believe that modern venture capital shouldn’t run on pen and paper, but leverage the same software stack and AI solutions which we expect our own portfolio companies to use.
Below are five examples of how we’ve been able to scale our investment team via software and processes.
1. A standard data ingestion process for every deal
Probably the most important thing we’ve gotten right at Hustle Fund is insisting that all deals (warm or cold) go through the exact same ingestion process. Every founder must: go to our website; click the “What are you hustling” button; and fill out our 5-10 minute form to provide detailed information about their company.
This form allows our team to convert all deals coming in as structured data. We leverage this data to automatically filter/reject companies that are definitely not a fit for our team.
Example rejection flags include: the valuation being too high; the industry/sector being outside our fund’s focus; the startup operating in a geography we don’t support. About half of all the companies we see will not pass this initial filter, and each rejected company will receive an email explaining why we will not invest.
Now that we have collected tens of thousands of deal records, this database is also the basis for training our own LLMs. We are currently training an AI bot to assist our investment team in recommending action on every single deal.
We never make an investment decision based off an AI recommendation alone, but it's a helpful aide.
Our deal flow engine represents proprietary data that serves as a huge moat for Hustle Fund. In theory, the more deals we see, the smarter our team (and AI) should become over time.
2. Standardizing deals, KYC, wires via software flow
Executing a deal and sending a wire sounds easy at a high level. But it’s easy to miss details like: ensuring we have all the key company data (articles of incorporation, operating agreement, etc); OFAC KYC security checks; double-verification of wire details; etc.
We use a software called Process Street, which we have heavily customized, to write data into our company record database; queue wires to our fund admins; run KYC checks; integrate with DocuSign to send SAFEs and side letters.
This has greatly helped with our audits to ensure that for each audit season, we have all necessary files to send to our third-party auditors for review. This is why even with over 500 companies, we delivered our audits early last season!
3. Leveraging shortcuts for follow-ups and rejection
Of the 1,000 companies we review each month, about 500 will be automatically rejected as part of the form ingestion process mentioned in #1 above. But that leaves 500 deal opportunities we need to manually review. That’s a lot. How do we manage that load?
We created the Pipeline Review Tool, which allows our team to process a lot of deals manually. For each hour that our team spends together to look at deals, we can usually process about 200 deals. We have also created pre-built due diligence templates for questions we’d like to follow-up with the founder, as well as templates for rejection. It’s as simple as pulling a selector option on a pull-down.
From the 500 companies we manually review from this tool, we’ll on average ask 50 companies to meet, and then select 7 companies for investment for a given month.
It’s super efficient!
4. Running continuous audits
VCs often treat audits like how a lot of people approach going to the dentist—only go when you’re in acute pain! It’s easy to push off important but not urgent items related to your deals until audit season begins—but ultimately, that leads to a ton of thrash and wasted team time.
We work with our awesome auditors in designing a quarterly audit form that our companies are required to complete. It’s automatically sent to our portfolio each quarter, and takes 10 minutes to complete. The form captures items like: new money raised, latest valuation of raise, current burn, etc.
Our portfolio company data is constantly being analyzed by our investment data analyst, whose core job is to verify, clean, and update all the company data across our portfolio. By running these audits continuously, it allows for our team to have an up-to-date perspective on our fund’s performance, as well as makes our annual audits far easier.
We are constantly experimenting with more software and integrations to make our audit processes even easier, and quicker.
5. Ingesting updates from portfolio companies
This one is super neat.
Beyond our quarterly audit process described above, many of our portfolio companies also send us monthly or quarterly updates voluntarily. We’ve created a process where we can email these updates to a specific email address, and an AI bot will then analyze/summarize the information in a structured way.
The data is then added to the portfolio company records in our core database, and (very soon) will automatically update the markup/performance value of that company.
The five examples above are just a sample of some of the great work our team is pursuing to scale Hustle Fund. Huge credit to our incredible past and present teammates, who have driven the creation of our unique internal software over the years!
Happy efficiency-finding!
Eric
Eric Bahn is one of the Co-Founders of Hustle Fund. He might be better known as that guy on Twitter who keeps talking about minivans. No one on the Hustle Fund team can understand that obsession.