Several AI startups in the NLP, legal, MLOps, and data training markets feature in the list of upcoming AI vendors to know in 2022. As the development of AI continues to accelerate into 2022, AI startups from all around the globe increases in their hundreds, and perhaps even thousands. As the AI world progresses rapidly, startups are incorporating sophisticated technologies into their services and products.
New and existing technology vendors, law firms, retailers, financial institutions, as well as manufacturers – virtually all giant businesses in most industries are utilizing or selling one form of AI or the other, be it advanced analytics, deep learning, or natural language processing (NLP).
AI companies compose just a small fraction of the group of vendors that have been identified to look out for in 2022. Nonetheless, it is likely that these vendors will transform the AI world, either with simple and easy-to-use AI-based products or innovative technology. Although there are several other startups that are worthy of tracking, we have chosen these vendors because of their exceptional potentials.
Law and innovation
A plethora of law offices and technology vendors use and sell AI-driven legal software designed to ease the work of paralegals, lawyers, etc. Perhaps the most useful function of these tools is their capability to create quick contracts or identify errors in legal papers. An example of such a startup is FairShake, which has been developed on a special business model.
FairShake is a three-year-old vendor based in Oakland, California. They provide consumers with a simplified method of submitting complaints and initiating legal actions against 50+ pre-listed organizations.
According to the founder and CEO of the company, Teel Lidow, a large number of the preselected businesses are telecommunications or financial corporations, as clients tend to often file claims in these sectors. Lidow expatiated that FairShake typically receives a lot of claims in which an organization uses misleading sales tactics or has failed to honor a promotion.
The FairShake website allows users to file their complaints in plain language, highlighting details about their issues, how they happened, and even when they started. Behind the curtain, machine learning (ML) and natural language processing (NLP) models assist in converting the natural language of users into legal documents that FairShake can automatically file with the American Arbitration Association.
Users can register for the service free of charge. However, FairShake receives a commission if they are awarded any cash settlement or a refund.
Generally, generating and filing documents are tedious processes, Lidow remarked. He said, “Your average consumer would be unable to grapple with it themselves,”
FairShake simplifies the whole process. The organization has processed as many as 10,000 legal claims since its inception.
The Hypergiant industry is another notable startup in this niche. The company deliberately tries to project a futuristic outlook, employing science fiction-style art on its website, with the catchy tagline “Tomorrow today” as well as galactic systems divisions and space-age solutions.
At first, the futuristic look may seem rather disorienting, however, it becomes logical after digging to know what Hypergiant does.
The company boasts big clients including Shell, NASA, the U.S. Air Force, Department of Homeland Security, GE Power, and Booz Allen Hamilton. Hypergiant is involved in the development of AI-driven products that, at a certain time, would have appeared extraordinary.
According to the CEO, Ben Lamm, “We formed Hypergiant to effectively tackle AI and emerging tech. We try to be innovative around solutions we think the world needs.”
The vendor’s R&D Laboratories is developing an astronaut helmet that displays information, including the vital signs of the wearer, on the helmet. Hypergiant is also designing another helmet that will be used by people on the ground. It aims to utilize augmented reality (AR) to present users with a video game-like map on their screen, allowing them to view objective points and the real-time locations of members of their team.
Hypergiant is also working on a robot that can autonomously tidy up a room and featuring germ-killing UV light. With laser and optical imaging, the robot can detect when humans are within close range and turn off the light, to avoid potential harm to a person.
“We try to be innovative around solutions we think the world needs,” the CEO said, further commenting that, for the most part, they are making investments in the R&D products that have talked about publicly “for the long haul.”
Hypergiant’s products and services, while space age-sounding, are based on real technology, including deep learning, natural language processing, robotic process automation (RPA), and computer vision.
Created back in 2018, the startup also has a remarkable track history with numerous standard projects, having developed various applications for its clients, one of which includes a predictive fueling system for Shell customers, an application for automating shipment documents, and a rail fleet management system.
Pure Machine Learning
Established by a team that developed Uber’s ML platform – Uber Michelangelo, Tecton.ai came out of hiding in April 2020 with an intention to assist data scientists to easily develop production-level ML models.
Tecton.ai sells a feature store, a storage location to store and manage ML features as well as run data pipelines to cover raw data into feature values. Tecton says that with the feature store, users can use their features more easily into production at scale, keeping track of the lineage and versions of their features, and tracking the condition of their feature pipelines.
In April, the feature store was released by Tecton as a limited beta and generally made available in early December. As the vendor puts it, the feature stores control data pipelines that transform raw data into feature values, and at the same time managing and store feature data. Using the feature store of Tecton, customers are able to easily ‘productionize’ the latest features; monitor feature versions, metadata, and lineage; and keep track the of the condition of their feature pipelines.
An estimate of $25 million has been raised in the seed and Series A financing of the startup and $35 million in its Series B financing co-led by Sequoia Capital and Andreessen Horowitz.
“We think that the Tecton team is so strong that we really focused on them from an early time on,” Martin Casado, a Tecton board member, and general partner at, had said.
Training and Document Data
DefinedCrowd was founded in 2015 and had the backing of Amazon. The startup sells training data for developing image and video recognition models, NLP engines, and intelligent speech models.
Mentioning that they hear a lot about the challenges faced by enterprises in finding high-quality training data, Nick McQuire, the CCS Insight analyst commented that startups such as DefinedCrowd “help with the ability for customers to crowdsource high-quality training data for a range of applications.”
When it comes to training speech models, the vendor is responsible for sourcing, transcribing, correcting, and validating speech data? In terms of training text-based Natural Language Processing models, the startup can also carry out the collection of text data, its annotation, validation, and tagging it. In like manner, it can identify, collect, and validate videos and images for those models.
Earlier this year, the company completed a round raising Series B funding in which it raised $50.5 million. Before then, the startup had raised $1.1 million and over $11 million in seed funding and a Series A funding round, respectively.
Around this time, yet another new Artificial Intelligence vendor can’t out of concealment at the beginning of 2020. The company, which was Alkymi, had generated $5 million in seed funding and released a product known as Alkymi Data Inbox. It enables the automatic extraction of data from business documents and emails.
Alkymi Data Inbox “tackles the problem of complex data and content within email inboxes,” a major issue as email boxes are normally locked silos, said founder and principal analyst at Deep Analysis, Alan Pelz-Sharpe.
The product gathers users’ documents (such as PDFs, images, and text-based files) and emails, all in a single platform. Leveraging image recognition and NLP, as well as input from users themselves every once in a while, the vendor extracts vital data and positions it under the right tables. For instance, the platform is capable of pulling payment data, including how much is to be paid and the date due, from a document or an email discussing billing.
Although Alkymi’s platform is targeted at the financial sector, the startup, however, has plans to venture into other industries in the future.
As hundreds of AI companies are launched every year, it’s tempting to believe that the number of new AI companies is getting too much. Contrary to that, AI is one of those fields that only experience growth when more participants are establishing platforms and more applications are learning from the data and environment. These artificial intelligence (AI) companies will determine the future of their respective verticals and specialized markets.
The AI companies discussed above are proof that ML (machine learning) has disrupted several industries and will continue to do so in the future provided that sufficient funding is available to these top AI companies over time.