Whether you want to start an artificial intelligence company as an entrepreneur, evaluate AI-driven vendors as an enterprise client, or simply learn more about new technology breakthroughs, you’ll need to understand how successful investors think.
Through funding, advisory, connections, and operational support, investors enable innovative companies to grow and drive overall industry trends. That’s why we spoke with dozens of venture capitalists who’ve made recent investments in artificial intelligence companies.
To help you understand how different investors approach AI, we compiled 15 unique and varied opinions that exemplify the diversity of opportunities for artificial intelligence to change our world.
1. Max Gazor, General Partner at CRV
Max Gazor of CRV believes AI will be most disruptive to industries where there’s “too much data for humans to process, a severe shortage of expert talent, and a high willingness to pay for even small productivity boosts.”
Cybersecurity fits these criteria perfectly. Business endure relentless amounts of attacks which fatigue an already limited number of expert security analysts and cause high churn rates in the industry. “Any improvement in productivity is very well received,” says Gazor.
Automated driving presents similar opportunities for computers to outperform humans. “Auto insurance companies have recently seen an anomaly in accident rates. Texting has gotten to the point where it’s causing a lot more carnage on the road,” Gazor explains.
CRV’s portfolio of AI-related investments represent the cream of the top as far as technical pedigree is concerned. Their portfolio company Rethink Robotics creates friendly manufacturing robots that work side-by-side with humans. The company was started by Rod Brooks who previously headed the artificial intelligence lab at MIT and founded successful robotics company iRobot. Jibo, another CRV portfolio company, is founded by Cynthia Brasile of the MIT Media Lab who is a world renowned expert on social robotics and human-robot interactions. Jibo is a family-friendly robot who interacts with you in a thoughtful, emotional way and slowly builds memories and relationships.
When asked about investment opportunities, Gazor points out that “the feedback loop between training and building inference systems is very long. This creates an opportunity for parallel and distributed computing to tighten those feedback loops.”
He’s particularly intrigued by the promise of Cerebras, a stealth startup that promises to architect processor chips specialized for deep learning which should outperform GPUs. “Not a lot of people have the requisite expertise to innovate in this space.”
2. Varun Jain, Early-Stage VC at Qualcomm Ventures
Although we still have a long way to go before we deem any AI startups to be unicorns, Varun Jain of Qualcomm Ventures already has two major hits with Cruise Automation (acquired by GM for $1 billion) and Clarifai, a computer vision startup that has raised over $40 million. Of the dozens of VCs we’ve spoken to, many regretted not having participated in Clarifai’s early rounds.
Jain looks for investments which, like Clarifai and Cruise, employ artificial intelligence at the core of their value proposition and not just as add-on features. He also prefers that companies build their own scalable, proprietary data source, whether through crowdsourcing, exclusive business partnerships, or other defensible methods.
For consumer-facing AI companies, optimization is particularly challenging and time-consuming. Many of these products produce the wrong input the first few times of use, causing major user churn. Jain advises that entrepreneurs building consumer experiences ask themselves “What is my trojan horse? What can I do to get a consumer to stay and generate valuable data?”
“I wish Alexa came out of a startup and not out of Amazon,” muses Jain. “It would have been a great investment.” Amazon took the Amazon Echo mainstream despite the overhead of requiring consumers to purchase a separate device. Additionally, they are not stopping at proprietary software, but turning Alexa into a platform to extend the reach of their API.
3. Sumant Mandal, Partner at March Capital
Sumant Mandal of March Capital employs an application-focused thesis for AI investments. He looks for companies which can provide a 5-10x efficiency multiplier on existing enterprise functions.
Mandal also co-founded The Hive, a Palo Alto-based early-stage incubator focused on data-driven, artificial intelligence startups. Portfolio companies include E8 Security, an enterprise security firm that combines hundreds of disparate sources into intelligent reports for analysts, Perspica, an analytics product that performs complex root cause analysis for data center issues, and Snips, a cross-platform virtual assistant builder that lets you spin up a smart assistant on any device with minimal setup.
For new investment opportunities, Mandal is primarily focused on healthcare, but also sees opportunities in hardware acceleration for AI as well as in insurance. He anticipates that insurance companies will experience a talent crunch and rush to buy tons of startups, but “people building insurance-related companies are still quite new.”
4. Colin Beirne, Managing Director, Two Sigma Ventures
“We don’t really believe in narrow technical approaches to A.I. or machine learning,” says Colin Beirne of Two Sigma Ventures. “Actually we believe in the opposite – using an ensemble of different techniques is required to solve most hard problems today. But targeting A.I. on a narrow domain space reduces learning complexity and improves the chances of success.”
Portfolio company Kasisto has built a virtual specialist AI to help you understand and manage your personal finances. “They work with bank partners to make conversation with a virtual agent another way you can interact with that bank’s offerings (the next wave in the progression from brick and mortar teller, to ATM, to a website, to a mobile app, and now to a bot),” explains Beirne.
Other notable investments include Amper Music, which takes in user specifications and composes a unique customized piece of music in seconds using AI, and Anki, which makes best-selling robotic toys powered by AI.
Like most of the other VCs we spoke with, Beirne is particularly excited about the applications of AI in healthcare. “We’re going to break the cost escalation spiral that U.S. healthcare has been in over the past few decades. At scale, the interaction of doctors and much more intelligent machines, all armed with much easier to access data, will be able to provide far better preventative care and spot problems before they cause massive cost.”
5. Dharmesh Thakker, General Partner at Battery Ventures
Dharmesh Thakker of Battery Ventures has a strong BS detector to filter out whether companies are truly working on next generation artificial intelligence. One key criteria he uses is the complexity and speed of the data they process. Unstructured images and video are more challenging to analyze versus text, as is fast-moving data versus static data.
One of his investments, Saffron Technology (acquired by Intel), analyzes copious amounts of email to detect fraud patterns within a financial institutions. Within the retail industry, portfolio company Reflektion is used by major brands like Disney, TOMS, and Godiva to enable real-time personalized merchandise offerings to consumers.
Thakker’s excited about the potential created by satellite companies like Orbital Insights and Planet as well as drone startups like Kespry to solve world problems. Equally exciting are opportunities in healthcare such as Imagia, which uses AI-driven image analysis to detect cancerous tumors.
“The whole healthcare system is so inefficient in terms of how doctors are utilized. You’re not really measured by the effectiveness of the care you provide.” Although healthcare startups will need to overcome regulatory approval and won’t see commercialization as quickly, Thakker is still positive about the space. “If possible, we like to make a return while doing good.”
6. Brad Gillespie, General Partner at IA Ventures
“If I had to pick one, domain expertise trumps machine learning expertise,” states Brad Gillespie of IA Ventures.
Gillespie is interested in two broad areas of investment: 1) Tools to advance AI and help people build better models from data, such as DataRobot, and 2) Vertically integrated solutions where AI is highly useful, but doesn’t make or break the company. In other words, “can you solve some meaningful problem in a more scalable way?”
Gillespie has seen firsthand how teams who put artificial intelligence before problem solving lose out. When a competitor to Vectra Networks, an IA Ventures enterprise cybersecurity company, pitched one of Vectra’s customers on their allegedly superior technology, the buyer’s response was: “These guys are smart but they don’t understand my business. Their product has lots of bells & whistles, but I can’t understand what it does.”
IA Ventures is also interested in markets with “winner takes most” dynamics and network effects, which is why they invested in x.ai, an automated scheduling assistant. While their competitors use human-in-the-loop systems which cost significantly more, x.ai aims to win more customers through affordable AI-only solutions. When both parties schedule a meeting via x.ai, the meeting is negotiated and scheduled with zero back-and-forth. Additionally, once you understand the nuances of managing calendars and schedules, numerous adjacent business opportunities arise.
Within healthcare, a popular space for AI investors, Gillespie is particularly interested in companies that go after secondary uses for drugs. In those situations, the drug has already overcome regulatory approval which significantly reduces company risk.
He’s also fascinated by efficient green chemical companies like Lygos, which uses yeast to create the same chemicals typically made with petroleum. By applying machine learning to genetics to design customized yeast which can produce target chemicals, Lygos has the potential to dramatically reduce our petroleum dependency.
7. Krishna K. Gupta, Founder at Romulus Capital
“Machine learning has been around for decades, as has the quest for artificial intelligence. But true AI won’t be achieved until we understand and replicate emotional intelligence,” says Krishna K. Gupta of Romulus Capital.
Gupta’s emphasis on emotional AI is reflected in Romulus’ portfolio companies, many of which spun out of the MIT Media Lab. Leading companies include Cogito, which uses AI to analyze voices and emotions and is used by companies to give their call center employees real-time feedback as they speak on the phone with customers. Another company, Humanyze, helps companies analyze how their employees are interacting, producing office environments that are more efficient, productive, and fun. A third, Ginger.io, tracks how mental health patients use their phones and alerts caregivers to any patterns or changes that may signal a problem.
One company Gupta wishes he had invested in is Geometric Intelligence (acquired by Uber to create Uber AI Labs). He explains why he found the company so compelling: “I find the approach Gary Marcus was taking to be bold and in line with some of what Marvin Minsky was promoting. I really believe we need to do more work to understand and emulate how humans learn.”
8. Robert Mittendorff, Partner at Norwest Venture Partners
Dr. Robert Mittendorff, with over a decade of operational and medical experience, is the ideal person to lead Norwest’s investments in healthcare AI companies. “I specifically look in areas of healthcare where we can help to scale or automate human labor, like radiology and coaching,” he clarifies.
Norwest counts AnalyticsMD, NextHealth and CognitiveScale as the leading AI companies in their portfolio. According to Dr. Mittendorff, “AnalyticsMD employs AI and ML to predict, prescribe, and persuade actions in real time to improve health system efficiency. NextHealth uses AI to inform and nudge patients of health plans to use their benefits more effectively. CognitiveScale uses AI to deliver insights as a service for clients in multiple verticals.”
9. Michael Dolbec, Managing Director at GE Ventures
Michael Dolbec of GE Ventures has a very useful rule for filtering out bad investments: “The more the CEO talks about AI, and not about their customer’s problems, the less interested I get. We fund valuable outcomes, not science projects.”
Dolbec has a point. Enterprises who buy AI solutions from vendors are looking for “outcomes and solutions, not flash and technique.” Instead of tricking people into clicking on ads, Dolbec prefers companies that solve important industrial problems, such as eliminating unplanned downtime of industrial assets, optimizing their long-term performance, improving the workflows and operations of industrial assets, logistics, and supply chains.
10. Suresh Madhavan, Investment Manager at Verizon Ventures
Like many other investors, Suresh Madhavan of Verizon Ventures, thinks AI has the most potential to disrupt legacy industries with massive inefficiencies. However, he emphasizes that startups stick to a very specific vertical focus in order to generate a high quality data set that can be expanded to other use cases in a vertical.
Madhavan provides an example: “AI techniques that determine the best areas for oil wells to be drilled, based upon topology and yield history, become incrementally valuable as additional data is collected from other oil drillers. Being very specific on focus allows for that data network effect to be seen in operation.”
SparkCognition is the leading pure play AI company within Verizon’s portfolio. The company helps industrial asset owners and operators optimize asset performance (like machinery) and predict and prevent potential issues by deriving complex relationships between data elements to highlight where potential failures or malfunction may exist.
“An example would be helping wind farm owners or oil and gas pipeline owners proactively determine any need for preventative maintenance based on a disturbance in behavioral data,” explains Madhavan.
11. David Cheng, Investment Manager at DCM Ventures
Like IA Ventures, DCM also invested in x.ai due to the company’s extraordinary potential to scale. “Most knowledge workers would hire a human assistant to run their calendar if they could; however, at a cost of $50,000 per year in salary, they can’t afford to,” explains David Cheng, Investment Manager at DCM. “At $39/month, X.ai’s AI personal assistant is within the reach of nearly all 90 million US knowledge workers.”
Additionally the fund invested in AVA, a personalized nutrition coach that combines the expertise of expert dietitians and cutting edge artificial intelligence. Cheng highlights why the company uniquely addresses the overcrowded dieting space: “AVA solves one of the most frustrating aspects of healthy eating — tracking and managing what you eat. By providing the first auto calorie estimation tool combined with image and human recognition, AVA is putting intelligent systems in the hands of skilled nutritionists to provide users with highly personalized eating programs”
Outside of their portfolio companies, Cheng is excited about Textio, a pioneer in natural language understanding (NLU). Textio’s first use case is to assess job postings for attractiveness to potential candidates, but the technology easily be extended and scaled to other business use cases.
12. Minal Hasan, General Partner at K-2 Global
Minal Hasan, General Manager of late-stage fund K-2 Global counts Spotify and Magic Leap as investments. Since Hasan’s fund invests more heavily on the consumer side, she looks for AI applications that fit seamlessly into our daily lives and automate tasks we currently do ourselves.
“As a working mother, Alexa has improved the quality of my life significantly,” says Hasan, “I use it to play lullabies on a loop to my toddler at nap time, tell knock-knock jokes to my seven year old, keep track of my grocery shopping list, tell me what restaurants are nearby depending on what I’m in the mood to eat, and order household items on Amazon with a simple request.”
Hasan brought up one issue no other investor mentioned to us: privacy. She wants to see more artificial intelligence companies actively protect consumer privacy, which she cites as “a big barrier to widespread adoption of technologies like Alexa.”
13. Bruce Cleveland, Founding Partner at Wildcat Ventures
Since founding Wildcat Ventures two years ago, Founding Partner Bruce Cleveland has made over 17 investments, many of which entail an AI or machine learning component.
One of them is Amplero, a “digital intelligence platform” that uses a self-learning, unsupervised approach called “multi-armed bandit experimentation” that enables brand marketers to automatically optimize customer interactions so they can maximize customer lifetime value. The companies customers span a diverse range of industries – including telecom, finance, gaming, and Saas – and seen, on average, more than 3% incremental growth in customer revenue, 5x retention benefit, and lower retention costs through less frequent customer interaction.
One company Cleveland wishes he invested in, but had closed their first round before Wildcat was founded, is 6Sense. 6Sense provides sales predictions to companies like Dell, IBM, and Dropbox, to better predict “winnable” opportunities and booking events, improving upon the static “system of record” functionality offered by most CRM systems.
Cleveland suggests that entrepreneurs and AI researchers thinking of starting a company look into innovating in field service, a highly costly area for any enterprise with physical assets which must be replaced or replaced. “According to some industry reports, the total addressable market for the field service market is expected to grow from $2B in 2015 to $5B by 2020, an estimated CAGR of 21.0%,” Cleveland highlights. “Applying AI/ML (and AR/VR) to this industry in order to reduce inaccurate repair diagnoses and optimize routing, etc. to reduce operational costs is a large opportunity.”
14. John Frankel, Partner at ff Venture Capital
John Frankel’s ff Venture Capital runs a four-month incubator called AI NexusLab to help early-stage AI companies prepare for seed and Series A rounds of funding. While Frankel acknowledges that AI is currently being hyped up, he also believes “it will be imperative for companies to develop an AI strategy in order to have competitive longevity, much in the way that today the necessity of a mobile strategy is implicit.”
Wade and Wendy, an ff Venture portfolio company, applies AI to capture new types of datasets (via a chat interface) and facilitate employment opportunities, candidate competencies, and general communication throughout the recruiting process. Clarity Money, another investment, looks at financial transaction histories to identify money-saving opportunities, such as better credit products or runaway subscriptions.
15. Neil Callahan, Managing Director at Pilot Growth Equity
Pilot Growth Equity, a growth-stage VC, built their own AI platform NavPod to improve the speed, accuracy, quality, and cost of their deal sourcing. Neil Callahan, Founder & Managing Director, explains: “NavPod learns about our interests and companies on their way to success, enabling us to spend time more efficiently. As a result, we can spend more time to assess code, learn from customers, and see prospective investments’ features in action.”
Companies that NavPod has helped Callahan discover include CB Insights, which uses AI to create a “Mosaic” score – like a FICO score – for private company health. The company has the most comprehensive database of private companies which is used by corporate customers to evaluate M&A targets. Perthera, another investment, applies AI and bioinformatics to genetic and proteomic analysis from a patient tumors and medical histories to develop personalized cancer treatment plans.
Marc Andreessen’s statement that “Software is eating the world” should be updated to be “Artificial intelligence is eating the world. AI techniques have the potential to disrupt virtually any industry through cost reduction, automation, and scaling.
Virtually every investor we spoke to is incredibly excited about the outsized benefit of AI applied to healthcare, but we also heard about numerous opportunities in hardware acceleration, insurance, field service, and more. If you’re an entrepreneur or AI researcher evaluating which space to start a company in, you’ll also want to learn how investors separate hype from reality.