We Used a Company Analysis Chatbot to Analyze 50 Unicorns, Here’s What We Got

Which company has a bigger investment in patents – Lyft or Uber? Which tech sector recently suffered from layoffs or legal issues? What kind of newly established companies will face a blue ocean, and who will face a red ocean? Which is investors’ favored industry today, frequently infusing hundreds of millions of dollars, and which one is forgotten, thirsty for money and attention?

In order to answer these questions, one can hire a group of analysts and have them spend a day or two processing data and reviewing the market. However, this time we used the advice of a company analysis chatbot named Emmet to get the information faster.

Meet Emmet: The AI Chatbot for Company Analysis

Emmet is the product of Zirra, a company that built AI-based tools to analyze other companies, currently serving VCs, accredited investors, as well as startups, accelerators and stock exchanges.

Using Emmet (f.k.a Nigel), we get instant smart insights on companies using simple English phrases such as “show me competitors for Uber,” or “opportunities for Github”, or “what are the risks for Airbnb?”.

At this moment, Emmet can produce insights on subjects such as:

-Level of direct competition: Emmet counts up to 20 competitors and evaluates the competition intensity, or the degree of overlap in their offerings. 50% and up is considered as a direct competitor. In our Unicorn analysis, we counted the number of competitors with more than 40% and 50% overlapping offerings.

-Time since last funding: It also calculates how much time has passed since the last funding round for a particular company in units of: 3 months, 6 months, 12 months, two years, three years and longer. A company, or cluster of companies, that has not attracted investors for a substantial amount of time can indicate it is either, 1) suffering from a lack of growth or 2) producing enough cash that allows it to reject investment offers.

-Layoffs: Emmet can detect the occurrences of layoffs that were published in media outlets along with other sources where it is listed on the open web.

-Legal Problems: The chatbot also identifies and determines the total number of legal problems that were published in the media or on litigation websites.

-Patent investment: Normalizing number of patents, Emmet shows them on a scale of 1-5.

To achieve this level of sophistication, Emmet uses a huge array of structured and unstructured data, from multiple sources, and then applies NLP and a variety of A.I. models to analyze companies and markets.

Using direct conversation with Emmet, we analyzed 55 private tech companies (that weren’t being acquired or merged), many of them unicorns, in 11 categories such as ride-hailing apps, self-driving cars or enterprise software. With the help of the tech behind the chatbot we could characterize companies, and clusters of companies, by risk and success criteria, establishing a better notion of their behavior and uniqueness.

Clustering companies into industry segments (i.e. putting Magic Leap, MindMaze, Improbable, Blippar and Niantic into AR/VR category) helped us better understand them, recognizing the uniqueness found in each sector.

Here is the analysis based on the AI chatbot’s findings within some of the reviewed industry segments:

Ride Hailing Apps

Ride Hailing Apps are characterized by a very dynamic funding activity, according to Emmet. In order to keep expanding into new territories, subsidize rides and attract more customers, ride-hailing apps are in constant need of cash. Out of the five companies analyzed here, Uber and Lyft raised hundreds of dollars each in the last three months, while Careem and Ola have done so over the previous six months.

Also, according to the AI chatbot, more ride-hailing apps are in need to create partnerships. Careem, the famous ride-hailing app in Arab countries, announced partnerships with taxi agencies, in-car WiFi providers, and payments services to adapt their service to local communities with local services.

In contrast to what you might have thought about ride-hailing apps, the total number of competitors is low, based on Emmet’s appraisal. After all, how many Ubers and Lyfts are competing for each territory at the same time? However, the intensity of the competition remains rather fierce.

Self-Driving Cars

According to Emmet, self-driving cars technology startups invest in patents, but at an average level that is not so different from most of the startups. Contrary to their innovative, scientific image, their total patent investment is lower even than cybersecurity or adtech firms, in regards to the number of patents.

Food Delivery

Food delivery startups are low on patents but packed with legal issues, according to Zirra’s chatbot.

For example, DoorDash had to pay $5 million to settle a class-action lawsuit that had stated that the company misclassified its employees as independent contractors. Instacart settled a class action lawsuit for $4.6 million after being accused of improper tip pooling and failure to reimburse workers for business expenses.


We also found quite a few legal issues in the fintech sector. Disruptive, innovative solutions that collide with the incumbent, traditional banking world is a high-risk potential. In addition, finding new models of dealing with people’s money can create lots of legal issues if not done right.

But not all of these lawsuits go around money. SoFi’s CEO Mike Cagney resigned last September following a sexual harassment lawsuit. Credit Karma was sued by the FTC after disabling security certificates, leaving social users’ data vulnerable to attacks.

In addition, funding activity in the Fintech segment is slow, as most of the companies sampled in this analysis didn’t raise funds for more than a year. Investors have cooled down their enthusiasm about fintech in comparison to 2016 or 2015.

Enterprise Software

However, the most dynamic industry segment we’ve seen so far is in enterprise software. Unicorns in this area are frequently funded, rated high on patent investment and partnership activity. However, they also suffer from a high degree of competition with other companies who have a similar offering, higher than any other industry segment examined in this project. This may also explain the high rate of layoffs in the enterprise software sector.

Are VR, AR and Drones still Trendy?

 Other notable tech segments that didn’t raise money recently include drones (19 months on average since prior investment), AR/VR (13 months), and adtech (more than two years). The data supports the fact that AR, VR, and drones are buzzwords that attract less money these days than what it used to be one or two years ago. Flying cars, self-driving cars, and ride-hailing companies were probably much more relevant to investors in 2017.


The adtech industry has suffered a blow after it had been accused of creating ad networks and servers that generated profit from an arbitrage in media trading. Consequently, our research shows the industry is highly competitive and experiences layoffs. Investors also have not put their money into this industry for more than two years on average.

Emmet, our chatbot, is just making its first steps into the world of AI-based company analysis.

It’s not perfect but its ability to find critical data and present it within seconds, like no human analyst can do, provides great benefits. Over time, we’ll be adding probabilistic NLP techniques to find new sorts of information, producing more elaborate and accurate insights.

The days when a chatbot will be able to supply analysts and investors with insights needed to make an investment – including private companies – are not far off.

About Zirra

Zirra’s platform offers users company analysis tools powered by data science and machine learning. Zirra’s customers include global and enterprise brands such as Microsoft Ventures, Deutsche Telekom Capital Partners, Silverlake, and Verizon Ventures.

Recognized by the Office of the Chief Scientist in Israel, Zirra is backed by top-tier investors such as Moshe Lichtman, former Corporate VP at Microsoft and now a general partner at IGP, AOL’s Nautilus (now Verizon ventures), Professor Dan Galai, co-inventor of the VIX Index, and Professor Zvi Weiner, Dean of the Business School at the Hebrew University.

Professors Galai and Weiner are both globally recognized researchers in the field of company rating and market risk analysis.


So How Much Is WeWork Really Worth?

Commercial real estate startup WeWork has attracted some attention recently after being Softbank’s latest multi-billion dollar deal, raising $4.4 billion at an estimated $20 billion valuation. WeWork is by far the largest shared workspace startup, but is a $20 billion price tag justified? Such valuation makes WeWork the 3rd highest valued American startup after Uber and Airbnb, but also raises some questions in light of the criticism over Softbank’s policy of investment resulting in extremely high valuations for its newby portfolio companies.

In order to answer this question, we used Zirra’s company analysis automation technology that helps get the fuller picture about every company in question. First, we asked Emmet, a company analysis bot on Zirra’s homepage, to analyze WeWork’s risk and success criteria:

Source: Zirra.com

**Try Emmet the company analysis bot here**

Emmet is a first of its kind AI-powered analysis bot that understands simple English and answers high-level company analysis insights within seconds. Emmet uses Zirra’s company analysis technology, aggregating data from a myriad of public and private information sources. It then utilizes Natural Language Processing (NLP) techniques to process large volumes of unstructured text, to detect companies and their semantic context, model their interrelationships, and derive dynamic indicators.

Emmet’s output gave us a few interesting points to think about when analyzing WeWork. Emmet confirmed some of the indicators we already sensed about WeWork: the company is well-funded, it has raised much more than its average competitor, and it has created many partnerships, including its most recent partnership with Airbnb, in which business travelers will be offered spots at WeWork locations.

However, Emmet also shows that it’s not all rosy in WeWork’s kingdom, despite the massive PR the company attracts. The company had to cut 7% of its staff last year, it suffers from a certain level of competition, it lacks IP (while holding many trademarks), and it suffers from some legal issues.

Let’s tackle the legal concerns. After learning about their existence from Emmet, let’s go to Zirra’s automatic company report about WeWork, accessible to anyone from Zirra’s homepage. In the section ‘latest events’ we will look for recent legal problems and then click on one of the articles offered.

Source: Zirra.com

One of the articles describes WeWork’s recent legal battle against UrWork, a Chinese competitor, over a trademark infringement. According to WeWork, UrWork copied a part of its name and its logo. The lawsuit is also in context with other problems Emmet has raised such as the growing competition and the lack of IP: on the one hand, WeWork cannot register a patent over shared working space, on the other, it competes over global domination, just as Uber in the automotive space, especially in Asia-Pacific, where UrWork is a strong player.

While UrWork expanded into Singapore and invested in an Indonesia-based rival, WeWork bought Singapore-based SpaceMob. UrWork, which is funded by Sequoia and Jack Ma, also announced it plans to expand into the US and UK. In fact, a part of the financial round led by Softbank was dedicated to WeWork’s expansion in Asia.

According to analytics service SimilarWeb, WeWork’s traffic mainly originates in the US, followed by the UK, Brazil, Canada, and Japan. The recent Softbank round was a part of WeWork’s effort to penetrate into Asia-Pacific.

Source: SimilarWeb

**Try Emmet the company analysis bot here**

Emmet also raised an HR issue. Checking WeWork’s LinkedIn employee and job opening count reveals an interesting angle. September 2017 was the first month in which WeWork significantly reduced its growth to almost none. Also, according to the company’s job openings chart, WeWork is now reducing its operations efforts in favor of marketing and engineering.

Source: LinkedIn

So how much is WeWork really worth? Despite raising a few risk criteria, Zirra’s company analysis algorithms forecasts a bright future for the shared workspace company known for their free beer kegs.

Zirra values WeWork even higher than the valuation given by Softbank last August. WeWork’s worth is estimated at about $25-$26 billion. In the case that the company chooses an exit strategy, most likely to go to an IPO, it can ask for almost $39 billion in valuation. According to Zirra, WeWork can go public within 2-3 years, at a probability of 60%-70%.  

Source: Zirra.com