In this article, I want to showcase some disruptive AI startups by application. My first article outlined the overall fundraising trends across the global AI landscape. My second article took a closer look at how conglomerates invest/acquire AI startups. Now, I hope to shine light on actual use cases of AI and what startups you should be paying attention to.
I mentioned before that I reviewed 1,600+ startups globally. After looking through each company, I was able to split most startups into two main categories (Horizontal and Vertical), of which was split into 22 subcategories to better define their application.
In this article, I only want to focus on the Horizontal AI startup category. This is made up of 9 subcategories, which are Analytics, Assistant, Data, Infrastructure, NLP, Robotics, Security, Tool and Vision (The definition of each subcategory is outlined in my previous article). I’ll do my best of giving a general overview of the subcategory and then I will mention a few startups that are excelling in this subcategory.
Analytics: (AI built solely to analyze general data and provide insight)
The Analytics AI subcategory is challenging because it’s hard to distinguish startups that are building a glorified business intelligence tool and an analytics application that has a form of machine intelligence. Most startups claim they can provide “transformative” information based on existing big data, but the trends that I’m seeing are Analytic AI startups are building tools that can automatically identify patterns in big data sets. This is different from most BI tools that require users to understand what they’re looking for or ask questions to probe for patterns. These tools simply figure out all possible combinations and suggests patterns/factors you may not have noticed.
Most of these Analytic AI startups cater to enterprises trying to optimise operations, understand customer behaviour, manage vendors/finances, etc. Outside of enterprises, this subcategory uses AI for a number of analytical purposes. Examples include natural disaster damage estimates, predicting maintenance of machinery, discover patterns/information across social networks, optimise mobile networks, identify opportunities in financial markets, identify economic or political changes, predicting price changes in consumer products or airfare, analyse scientific or legal literature, helping realtors figure out when someone is moving, predicting success of books or movie scripts before release and much more.
Startups in this subcategory you should watch include: Ayasdi*, ThoughtSpot, Tamr, INCORTA, Banjo, FLYR* and Interana*.
In total are about 20 out of 100+ AI startups in this subcategory that I’m watching closely.
Assistant: (Tools that are built to be forms of chat bots)
The Assistant subcategory is comprised of mostly startups that create chat bots. Assistant AI startups can take many form factors but their ultimate goal is to help you complete a task through a simply question/answer system.
Applications of chat bots are vast but each chat bot is only capable of handling a single use case without human intervention. Applications include setting up meetings, querying information from science/financial/legal databases, managing customer service questions, travel/flight/hotel assistant, personal on-demand assistant, virtual sales teams, banking requests, buying gifts and more.
It’s notable that Tech Giants were very active in this category between 2013-2015, but interest dropped off afterwards.
Startups in this subcategory you should watch include: Eloquent Labs, Octane AI, Hyper Anna, X.ai, Mezi (Recently Acquired), Magic*, Recast.ai (Recently Acquired), Action.ai and Sensay*.
In total are about 10 out of 60+ AI startups in this subcategory that I’m watching closely.
Data: (Startups that acquire, analyze and sell private data sets)
The Data subcategory is very similar to that of the Analytic subcategory. The only difference is that these startups hold proprietary data. Many of these startups offer services to help you cross reference their data with yours, others provide data for educational purposes, while most simply sell their data/analysis to give companies an edge over their competitors.
Startups in this subcategory you should watch include: Near*, Qloo*, Node*, Premise*, and Descartes Labs.
There are 5 AI startups out of 15 that I’m watching in this subcategory.
Infrastructure: (Startups building the physical or software infrastructure to empower AI tools)
The Infrastructure subcategory is comprised of mostly hardware AI startups. These companies are focused on building servers, chips or any hardware system that can process information faster, handle more complex calculations and/or can operate more efficiently. There is a blend of quantum computing in this subcategory. Almost all of these startups are focused on solving the infrastructure challenges required for AI tools to perform optimally.
Startups in this subcategory you should watch include: Graphcore, Kneron, Cambircon Technologies, Rigetti Computing*, Thinkforce, LeapMind, Bigstream* and Wave Computing*.
There are 10 AI startups out of 30 that I’m watching in this subcategory.
NLP: (AI startups focused on understanding speech, voice or text)
The NLP subcategory is made up of startups building tools that pick up on human voice/text, understand the meaning behind what is said/written, and sometimes take action upon that information. What differentiates most NLP’s is how they compare your voice/text to a particular data set. Applications of NLP include translation, interacting with electronic devices, querying a knowledge base of an organisation, analysing or optimising meetings/sales/customer service and more.
Startups in this subcategory you should watch include: MindMeld (Acquired), Unbabel, SoundAI, AISense, Insight Engines*, Entefy*, Snips*, Agolo* and Semantic Machines*.
There are 11 AI startups out of 60+ that I’m watching in this subcategory.
Robotics: (AI built for robots to perform any types of tasks, movements, etc)
The Robotics subcategory is broken in to two parts. First, there are startups focused on the ‘brains’ of a robot. Their goal is to train robots to perform various tasks on their own. This is done training the robot to teach itself or users guiding the robot to perform a specific task. Second, there are startups building robots solely for a specific task.
Robots are built to be companions, some for healthcare monitoring, managing warehouses or retail stores, security guards, inspection, delivery, restaurant catering, education and more.
Startups in this subcategory you should watch include: Vicarious*, Intuition Robotics*, CloudMinds*, Embodied*, ROBART GmbH and Embodied Intelligence.
There are 14 AI startups out of 45+ that I’m watching in this subcategory.
Security: (Tools built to protect companies from security threats)
Security is a challenging subject when reviewing startups because it is very arbitrary how startups measure success. Most Security AI startups claim they are better a spotting abnormalities. How security startups take action upon their findings are different. Some simply notify companies of patterns that aren’t normal, others provide action for specific cases and others try to solve the problem.
Examples of applications in this industry include identifying malware in emails/computers/networks, monitoring phones/IoT/Vehicles/enterprises, various forms of authentication, identifying fraud in various financial services, monitoring your online presence or your online community and more.
Startups in this subcategory you should watch include: Emailage, Versive, UnifyID*, Callsign*, Simility, NS8 Inc, Blue Hexagon, ZingBox and Socure
There are 25 AI startups out of 60+ that I’m watching in this subcategory.
Tool: (Startups building AI algorithms)
The Tool AI subcategory focused on building tools to optimise, deploy and manage machine learning models. Tool AI is mostly built for developers but there are many startups trying to help non-programmers deploy machine learning models. Many of these startups offer different analytical services based on the common trends they find among their clients. These services are very similar to that of Analytics AI subcategory. Many of these startups offer a consulting services to deploy their tools.
This subcategory is popular among investors from all industries. In addition, there a handful of startups that have multiple investors from the same industry.
Startups in this subcategory you should watch include: Kensho*, Bonsai*, H2O.ai*, CognitiveScale*, DataRobot*, Element AI* and Neura*.
There are 30+ AI startups out of 95+ that I’m watching in this subcategory.
Vision: (Startups building image recognition tools)
Vision AI startups are applying forms of image recognition to various applications. Examples include indexing content of videos, visual search engines, analytics of earth imagery, security and monitoring of offices or cities, identifying people or understanding their emotions, capturing key moments of an event, building eyes for autonomous vehicles/drones/objects, home designing, analysing safety in construction sites, building logos and much more.
As you may recall from my previous article, this was the most heavily invested category by large corporations. Tech Giants, Top VCs and Commerce were the most active in this subcategory.
Startups in this subcategory you should watch include: Mighty AI*, PointGrab Ltd.*, Vidrovr, Clarifai*, PrecisionHawk*, Skycatch*, Shield AI*, Vion Technology and SenseTime
There are 40+ AI startups out of 130+ that I’m watching in this subcategory.
I hope this review gives you more clarity about what is actually happening in the AI ecosystem. My goal was to share different examples of AI in order to make the subject more familiar. Ultimately, I hope this serves to inspire new founders, help corporations discover new applications of AI and help investors find new opportunities. I will publish an article soon on the Vertical AI category.
* Is for investors to know which AI startups from this list will be fundraising again in the near future, are currently fundraising for their next round or they could announce the closure of a new round very soon.