AI: More Than Just a Buzzword
The F2 Way
When we think of artificial intelligence (“AI”), a few things come to mind: Arnold Schwarzenegger in the masterpiece Terminator 2, C-3PO in Star Wars, and the world of Westworld. AI as many see it is evident in each of these movies and series — a computer or robot uses intelligence that originates in humans to learn how to independently “think”, developing its own identity.
While these movies and TV shows are fictional (and fantastic, by the way), the truth behind the robots is the same on screen as it is in real life:
through AI, we can create machines that can perform endless tasks at a rate faster than humans and limited only by physical computing power. Pretty cool, right?
AI, machine learning (“ML”), and deep learning are all related and evolved over time like this:
Especially these days, AI is changing the shape of our world and automating manual processes across industries. In order to build the most effective and efficient models, however, there needs to be a solid foundation in place. Take machine learning for an example. When thinking about all ML can do, we often look at the end result, the product. That machine, however, only functions as a result of access to data, clean the data, and learn from the data. Which brings me to a vital topic: data.
Data.
At the beginning of this year, the digital universe comprised of 44 zettabytes of data. As of 2019, more than 4.5 billion people use the internet. The stats speak for themselves: there is a ton of data out there and the volume is only getting larger. Most of this data, however, remains untapped. On average, 73% of a company’s data goes completely unused, which means companies are losing out on major opportunities to analyze all aspects of their work streams.
As a result, every organization is becoming a data organization. The only way to make sense out of this big data is to leverage AI.
Because of this, applications and frameworks of AI and ML that allow enterprises to retrieve, digest and analyze the data, and turn that data into tangible solutions, are more important than ever.
The AI value chain.
In the past decade, we’ve seen the proliferation of software development tools that make building software easier and more effective. AI is beginning to follow suit. AI products rely on their own value chain. ML is a great example: Its models are only as good as the data they are based on; the more robust the data is, the more accurate the models. Entrepreneurs are racing ahead to fill the gaps. Just as with software development tools, AI tools that target a specific aspect of the ML value chain are popping up all over.
There are also incredibly interesting companies striving to provide full-blown ML platforms that cover all, or almost all, of the value chain. These companies provide solutions in verticals such as healthcare, retail, insurance, and you name it!
Here is the value chain for ML as we see it:
Notice how small the category dedicated to actual ML code is. Whether a company is targeting a specific problem on the value chain or creating holistic solutions to vertical-specific problems, AI companies are red hot.
Our thesis.
F2 is a specialized, seed-stage venture capital fund backing Israeli deep technology companies at the junction of big data, AI and connectivity.
The value chain and frameworks that help to build and implement an ML model and solve a specific problem are not mature enough. We see this not as an issue, but as an opportunity to help develop and fund such infrastructure. We also believe there is opportunity in companies in different verticals that are creating holistic solutions by leveraging AI as part of their core technology.
How our companies fit.
Following this thesis, we have led many pre-seed and seed stage investments in AI that can be divided into two main categories:
framework companies that fit into a specific part of the ML value chain and vertical holistic solution companies whose core IP is partly based on AI.
Below are some examples of companies from our portfolio that are targeting specific aspects of the ML value chain:
- Dataloop: develops a one-stop-shop platform for labelling datasets for ML. It includes a data management environment that includes automatic annotation capabilities and data quality control management.
- Datomize: generates high quality synthetic data through advanced AI, that can be used both on-premise and over the cloud. It empowers R&D and Data Science teams by enabling effective AI modelling, 3rd party collaboration, testing and development, while complying with the most stringent privacy regulations.
- Explorium: unleashes data silos and enriches them with external sources online, enabling the ML model building process to begin with an optimal data set that could not have been created otherwise.
- Superwise: helps monitor and analyze the critical data points affecting the performance of a ML model.
Below are examples of companies from our portfolio that are using AI to provide solutions to vertical-specific problems:
- Realfriend: develops a chatbot that leverages various state of the art ML techniques to deliver personalized services to real estate agents so that they can support multiple customers at once.
- Zero Networks: ensures that each user and machine is only utilizing the network resources they should. This enforces an airtight, least privilege networking stance at scale to eliminate internal attack vectors and allow companies to trust their networks again.
And here is how these portfolio companies fit into the ML value chain:
To sum up.
This is truly an exciting time to take part in all things AI. There is no telling what trend will lead the industry and what changes may occur over the next few years. I am confident, though, that because of its ability to apply AI to almost any vertical, there is a blue ocean for game changing founders.
If you are an early-stage start-up that speaks and breaths AI, and thinks big and out of the box, we want to help!
Contact me @ maor@f2vc.com
Hasta la vista baby!
Maor Fridman
* The author in an Associate in the Investment Team at F2 Venture Capital
