In the eye of the artificial intelligence market, how do investors choose a good company?

In the past year, artificial intelligence has successfully captured the public's attention, dominated the leading media reports, and drove a large number of investment and acquisition activities. On the cusp of this hype cycle, how to identify a truly valuable startup is a big challenge for VCs.

We interviewed experienced venture capitalists from top companies like CRV, IA Ventures, Two Sigma, etc. to see how these successful investors are evaluating artificial intelligence startups. If you are the founder of a company that wants to create artificial intelligence as its core, please focus on the issues mentioned below.

Is artificial intelligence the core of the company's value proposition?

Varun Jain of Qualcomm Ventures warned: “Many companies that are unable to raise funds are trying to position themselves as artificial intelligence companies. Jain said that he has seen companies that only use Wi-powered artificial intelligence. The -Fi router claims to be an artificial intelligence company.

In these cases, artificial intelligence is just an added feature to these companies, not the core of the company's value proposition. Jain explained: "Traditional Wi-Fi routers can use artificial intelligence to detect anomalies in network data and flag those errors, but this feature does not change or add value to the company.

In contrast, the High Ventilation Investment Company invested in two artificial intelligence startups, Clarifai and Cruise AutomaTIon (which have been acquired by GM). Cruise is a company that provides intelligence and power to autonomous vehicles, while Clarifai uses advanced depth learning and computer vision technology to accurately identify objects in images and video.

How reliable is your technical team?

According to CRV's Max Gazor, "Companies with cutting-edge and innovative artificial intelligence technologies will have recognition from AI or a well-known industry organization such as Google Brain or Facebook's artificial intelligence team."

CRV's commitment to the company's technology systems is reflected in the extraordinary experience of the founders of their investments. Rob Brooks of Robhink RoboTIcs is the founding director of the MIT Artificial Intelligence Laboratory and the founder of the successful robotics company iRobot (NASDAQ: IRBT). Jibo's Cynthia Breazeal founded the Personal Robotics Group at the MIT Media Lab, a world-renowned social robotic expert. Pullstring's Oren Jacob used to be the chief technology officer of Pixar, working with Steve Jobs since the company's inception. (Robhink RoboTIcs, Jibo, Pullstring are all CRV companies)

David Cheng of DCM Capital added that at this stage of the artificial intelligence industry's strong market, only a limited number of artificial intelligence experts have the necessary experience in large companies or top universities to truly build cutting-edge and innovative solution. If a team claims to use artificial intelligence technology in their products and the company does not have a matching technical team, this will cause us to doubt its authenticity.

Can you solve the customer's actual problems?

“One rule I found was that when CEOs kept on emphasizing the company’s artificial intelligence technology, not the customer’s needs, my interest in them was lower.” Michael Dolbec of GE Ventures “We invest in valuable The results, not the research projects.

All the investors we have contacted in the industry agree with this view.

Brad Gillespie of IA Ventures added: "If I don't want to do this single-choice question, in my opinion, domain expertise is better than machine learning expertise. IA Ventures invests in Vectra Networks, an experienced field expert. Leading network security company that focuses on solving important customer issues and maximizing the use of security analysts.

Vectra's competitors repeatedly emphasized their complex machine intelligence, but the buyer's feedback is that these guys are smart, but they don't understand my business. Their products have a lot of slogans and terminology, but I don't understand what it is.

Effectively solving business problems requires not only a professional team, but also a specific business area. Colin Beirne of Two Sigma Ventures points out that existing solutions with different technologies can solve most problems, but artificial intelligence for narrow areas requires more precise solutions.

Do you have related, proprietary and scalable data sources?

Jain of Qualcomm Ventures always asks this question to potential investors: How do you get the data? Do you rely on large companies to provide data, or do you have an independent collection method? Both methods are feasible, but having an independent collection is often more popular.

Self-driving cars have traditionally been tested in suburban areas, parking lots, and closed environments that do not reflect driving reality. Qualcomm's portfolio company, Cruise AutomaTIon, obtains many missing data by operating test vehicles in an urban environment and being monitored in real time by professionals. Similarly, its other artificial intelligence investment company, Clarifai, began working with a popular consumer app to get unique data and then expand it to get the specific data needed to handle a specific business.

In addition to the uniqueness and discernibility of data sources, they must also be relevant to real-time issues that need to be addressed. According to Dharmesh Thakker of Battery Ventures, the next generation of artificial intelligence technology will depend on the complexity of the data you are mining. Unstructured image, video, and audio data are much harder to collect than text data. Thakker also considered whether companies need to use fast-moving data or stop at static data. Fast moving data, such as real-time images processed by autonomous vehicles, is often much more complicated than static data.

Finally, the team must demonstrate their ability to continuously improve their performance based on unique data. Highly ventilated Jain regularly checks whether the team can demonstrate the ability to quickly process data and effectively optimize technology, making the system more reliable.

Have you built your own unique technology or rely on open source code?

Verizon Ventures' Suresh Madhavan said: "Using open source code frameworks or developing proprietary technologies is often an important criterion we will look at. Using open source code will allow you to analyze some surface data, but it can't be solved. The best solution for every difficult business problem."

Cheng from DCM Ventures also agrees. DCM's investment team relies on a strong network of industry consultants and technical experts who can help review technology heaps, data architectures, and determine if teams are properly collecting, storing, parsing, or annotating data.

Do you have a sticky product?

Sumant Mandal is a partner at March Capital and co-founder of The Hive, an “incubator” focused on driving the development of artificial intelligence startups. Mandal emphasizes: "If your robot can't improve the efficiency of your customers by at least 5-10 times, then as a new company, it's hard for customers to see your value," Mandal stressed. "As a Startups, you have to see more than just your own products, but the benefits your products bring to your customers. For example, if you want to apply robots to recruit for your customers, then ask yourself If my robot can increase efficiency by five times, can the recruited person increase the profit of the client company by 100 times?"

In addition, he warned that any improvement must be made under the premise of responsibility.

Even if you have an ideal product, letting the customer commit to a single pilot program is not a viable business. Woodtik Capital's Kartik Gada seeks a diverse income and diverse clientele, often asking these two questions: “Is your income growing steadily? Has your customer become a repeat customer?

Do you have a diverse team?

The last but equally important point is that investors are looking for a diverse team that can solve the initial challenges of the company and have the confidence to overcome all the challenges of expanding the artificial intelligence business process. Kiersten Stead of Monsanto Growth Ventures explains, “Successful companies hire the entire team, including the domain experts, business leaders and sales people, not just the engineering team.

In contrast, Stead observed that a single startup team, especially when composed of artificial intelligence researchers with no industry-specific experience, often fails more frequently.

“The technical team can't communicate well with the sales staff, and vice versa,” she stressed. “We are looking for a very experienced, artificial intelligence founder with a more mature career or a A team with a long history.” Sales and marketing management are often overlooked in the process of technology-based artificial intelligence entrepreneurship, but for companies that want to succeed, excellent market management is often indispensable.

Woodside Capital's Gada warned: "The biggest mistake made by artificial intelligence startups is to stifle marketing. Most customers don't know they need these high-tech products."

PMMA Cement

Pmma Cement ,Pmma Bone Cement,Contain Powder,Polymethyl Acrylate

Dragon Crown Medical Co., Ltd. , https://www.dragoncrownmed.com