I ask nosy questions.
As an investor who used to run his own machine learning business, I’ve put money into five promising AI startups in the past years.
These investments are more than an exchange of money. After signing a check, I want to mentor entrepreneurs, make introductions, and do anything I can to help my portfolio companies succeed. I have more to offer AI entrepreneurs than money — I have technical expertise and hard-won familiarity with the industry.
But before reaching any kind of agreement with a company, I meet and speak with the team doing the work. I want to see what they’re working on, ask them questions, and get a vibe for how they operate. Do they seem capable and committed to the task at hand? Does their approach to the problem make sense? As much as I evaluate a company’s technology for investment, I also evaluate its people.
No two founder conversations are the same when I kick tires on potential investments, but I find myself asking the same questions each time. I’m not trying to learn trade secrets. These questions are my tools for teasing out enough details so I can determine if a company is working on something valuable or not.
Here’s what I ask when I’m vetting companies to be part of my portfolio.
Who are your founding team members and what are their technical backgrounds?
I like to see multiple co-founders attached to a startup instead of a standalone solo entrepreneur. These companies prove to be more robust and enjoy more success. Because I specifically invest in artificial intelligence, at least one of these co-founders ought to be a technology rainmaker with street cred or some other proven track record.
I like to see academic credentials in data or computer science — a bachelor’s degree or better is great. If a company’s tech guru doesn’t have that education, I’ll defer to his or her field experience. They ought to have a history of completing real-world projects related to the work they’re doing now.
In simplest terms, I don’t want to work with tourists. I want to invest in people who already decided to specialize in AI (just like I did) and help them meet their goals.
Do you have domain expertise?
Suppose a startup built an artificially intelligent algorithm that monitored your securities trading history, learned your approach to the market, then started executing trades on your behalf. For that system to reach its maximum potential, it needs a markets expert working alongside an AI wonk.
If an AI system solves a retail problem, the team ought to have someone with corporate retail experience aboard. If it solves real estate problem, they need a realtor or broker. Whatever niche the system touches, I want to see an appropriate industry expert helping carry the team.
The domain expert and the technical expert need to feed positively off of each other. They keep each other’s expectations in check.
Does your solution actually use AI?
It’s happened too many times: a startup has some data (and even uses that data), but gratuitously labels this practice “artificial intelligence.” I’m an investor technical enough to ascertain if a so-called “AI system” is actually just checking a rule-based threshold. That’s conventional computer programming, and that’s not what I’m looking for.
I’m looking for companies who take artificial intelligence seriously. These will be the companies that don’t lean on buzzy terms for marketing purposes. When companies carelessly invoke artificial intelligence and machine learning, it’s sensationalist hype at best and it’s misleading at worst.
How close are you to your customers?
There’s nothing more pleasing to an investor than seeing a startup have deep customer empathy and understanding. Customers are the people providing the all-important revenue that carries a business through its quarter — they’re presumably the whole reason the company exists.
I don’t want to see arm’s-length situations here. Failure to consider the customer is always a misstep. The better you understand them, the better you can serve them. So show me that you understand them.
Where does your data come from?
Data enables just about everything that AI technology does, so it needs to be as pure as possible. But the real world has very little in common with a laboratory — things can get messy and imperfect. That’s why I value startups that automate their data pipelines effectively.
If I learn that their people spend lots of time cleaning and organizing data, it’s a problem. The idea is to scale a process to be bigger and better — if people play a big role in gathering the data that powers the system, then the company will never reach its full potential.
Are you just looking for money, or do you want mentorship as well?
I don’t just want to invest in artificial intelligence companies — I want to pour gas on their figurative fires.
My resources here go far beyond the financial. My runs in business and computer science have been successful, and I want to bring my knowledge and network to their playing field. I want to make introductions, I want to troubleshoot technical problems, and I generally want to do everything I can to help a company succeed.
I’m not in this for transactions. I’m in this to help bring the next generation of world-changing technology to life.
Who are your competitors?
Don’t tell me there’s no competition. That just signals to me that you haven’t done appropriate due diligence. Some founders will even cling to technicalities that make the “we have no competitors” sound byte true. If another startup solves the same problem by running a different algorithm on a different dataset, they will point to this as evidence of their own unique, unbeatable solution.
But this is self-deception and an unwillingness to observe the field for what it is. Every time I’ve heard the “no competition” line, I’ve been able to find the competition within a few minutes of searching. There is always competition, even if you say there isn’t.
What’s your go-to-market strategy?
A go-to-market strategy is like the final mile of a startup’s operation. It’s their specific plan for delivering value to the market and winning customers, so this question deserves a thoughtful, comprehensive answer.
I smile when founders answer this question in a way that reflects a lot of care and consideration for the customer buying journey (you’ll recall my earlier point about startups being close to their customers). It only gets better when the team values segmentation — it means they’re trying to pick the best targets instead of trying to sell to everyone.
Asking the right questions doesn't mean I'm being nosy. It just means I'm finding the companies who most closely align with my expectations. A "one size fits all" approach just won't do.