Atomian: The powerful cognitive software that thinks, works like the human brain
Spain · Jun 04, 2019· By Amalia González
Photo by ra2studio/Shutterstock
Combining natural language processing with big data, Atomian enables easy, quick and real-time access to information in databases and documents
Imagine a digital brain that can process any and every type of data, and sift out useful data from dark data. A brain that can be customized and used across a variety of sectors – law, business, healthcare, banking and insurance – for better decision-making and more efficient processes.
“This is what the Atomian founder, Miguel Montero de Quadras, has been working on for the past 18 years,” said Beatriz Cabrera, Atomian's business brain and newly appointed CEO.
As early as 2004, Montero, who was Atomian’s CEO until October 2018 and is now its CTO, had said that it was possible to construct a set of symbols, data structures and algorithms that could reproduce the same behaviors and abilities as the cognitive processes of the human brain, Cabrera told CompassList at the recent 4YFN conference in Barcelona.
Based on that idea, Atomian developed a digital brain for large-scale organizations, one that can exploit data using natural language processing (NLP) as well as to learn and automate business processes. For this, the Atomian founding team spent 15 years researching and constructing a software and a complete cognitive architecture that enables people to establish natural communication with information.
Known as cognitive computing software, a term popularized by IBM with its Watson computer, the Atomian software can be applied to online platforms and interact and think like humans, finding solutions to problems that humans are typically tasked with or to complex situations. The software can also synthesize data from various information sources, because of machine learning.
The cognitive computing market is segmented by product into NLP, machine learning and other technologies. According to a report in March last year, the size of the global cognitive computing market is expected to reach US$49.36 billion by 2025. The findings in the report also suggest that the NLP technology segment has dominated the industry, accounting for over 40% of the cognitive computing market by volume.
According to Cabrera, although there are other companies that work with data and can adapt their solutions more or less well to client’s needs, “today there is no other software [other than Atomian] capable of extracting data from either unstructured or structured documents, assemble the information, and answer queries about it just using NLP."
The founder, Montero, who has degrees in industrial engineering and software engineering from the Polytechnic University of Catalonia and the Open University of Catalonia, respectively, began to develop his product while he was a student. His aim was to design and develop a humanoid cerebrum. He got his breakthrough in 2011 when he developed his first working prototype. Montero then formed a company together with Luis J. Garcia, Pol Vives i Fàbregas and Guillem Ysàs Miralles. That's how Atomian was born in January 2014.
The success of the company’s proprietary technology and the opportunities to scale up the business meant the Atomian required a new CEO with experience in large companies. Montero chose Cabrera, a telecom engineer with an MBA from the Manhattan College of New York, and more than 30 years’ experience within the IT and enterprise software companies.
Her background included management positions, developing new sales channels at an international level and managing direct and virtual multicultural teams in matrix organizations, including being Director of Sales in Europe (Channel South Director) at CA Technologies and Vice President, Partners and Alliances, at Sage.
"Atomian has incredible potential,” Cabrera told CompassList. “The first thing I did when I started talking with Miquel Montero was to ask for participation in the company. I started seven months ago, especially to take care of the expansion of the company. What Montero saw in me was a person who could really launch the business plan, and help the company get new customers and monetize the product."
Cabrera, who was accustomed to working for multinationals, found in Atomian a startup providing a very powerful software to several companies and sectors, but with a lack of focus. Her first job was to determine which products to focus on. This led to the development of Atomian Discover and Atomian Highlight.
"We know that our software can be applied to many tasks and has a great future, but we also realize the advantages of continuing the development, while monetizing only some of Atomian’s possibilities now,” Cabrera said.
R&D + commercialization
Atomian has released two products for large companies. The first, Atomian Highlight, automates company processes that require reading documents. The product can read and classify documents and can be incorporated into management and service request systems to automate decision-making processes.
"This is a product that allows you to extract data from any unstructured source, such as pdfs and email and structure the data", Cabrera said. “It is able to learn and solve tickets with structured information and text fields or to read and understand new documents, extracting key information.”
The product has three versions: Atomian Class Highlight, Atomian Business Highlight and Atomian Legal Highlight.
Its other product, Atomian Discover, is able to answer questions formulated with natural language on the basis of the information provided. In the case of unstructured information, it can first be structured using Atomian Highlight. "It has a bar in which you can ask Atomian using natural language, currently Spanish or English, and Atomian responds," said Cabrera.
Atomian Discover also allows the creation of dynamic dashboards (for example, chats, tables, alerts, specific answers, progress bars, KPIs, insights, or forms) by writing your questions in the mentioned bar and sharing them.
In many cases, the two products are combined. For example, in Healthcare, where Atomian Highlight reads the pdfs and extracts the patient data, and doctors can then ask in natural language and get answers from Atomian Discover about patients’ data. It is also possible to access Atomian’s products in the cloud (Atomian Cloud).
As the company works in predictive analytics and simulations, it is aware that there is still room for further development. “Atomian’s business model is SaaS but we make a cognitive ETL ("extract, transform, load") to adapt the product to the client's data,” said Cabrera.
“Users are billed monthly. We license the product and charge a fee per user per month, in the case of Atomian Discovery, or a fee for each processed document for Atomian Highlight," she added.
Evolution and future
Since the company was formed in early 2014, Atomian has tapped external funding over four rounds to maintain the business and increase hiring. It currently has a staff of 12, of which more than half are engineers. The company is preparing for another funding round with a target to raise €5 million.
Focused on growth in a sector that grows annually by almost a third of the volume of the previous year, Atomian wants to reach as many customers as possible. The next step will be internationalization.
"Having the product already in English and Spanish, the idea is to go to the UK and the US, so that they can see the potential of the software and the products," said Cabrera.
Although Atomian is present in various sectors, it is currently specialized in healthcare, banking, law and insurance. Healthcare is where Atomian started: "For a doctor, it is a great advantage to have immediate access to any processed data," Cabrera said.
Atomian has also been targeting medium-sized and large companies in other sectors, including multinationals, such as those in the automotive sector. "Our goal for 2019 is to reach €1 million in sales, and for 2020, €3.5 million," she added.
Edited by Bernice Tang and S. Mani
At CompassList, we pride ourselves on the accuracy of our information and reporting. Please help us by letting us know of any incomplete or inaccurate information on our website. Rest assured that we will handle every piece of feedback with care.