See the original story in Japanese.
Tokyo-based Cognitee announced last week that it had raised 150 million yen (about $1.3 million) in its pre-series B round. The lead investor was Global Brain this time and Mitsui Fudosan (TSE:8801) joined anew as an investor. This follows their series A round conducted in 2016 having investors including Global Brain, Alps Electric, beBit UCD Ventures, Glocalink (affiliated by Leave a Nest), Active and Company, plus SMBC-VC.
Cognitee has fundraised 280 million yen (about $2.5 million) in total since its angel round. With this investment, the firm will strengthen its organization system by hiring additional core members and prepare for overseas development which will take place next year and beyond. Additionally, the firm announced it invited Kazuhiro Kondo as COO; he has careers as Corporate Officer of DeNA and VP of DeNA West (subsidiary of DeNA covering the West) after serving Sony, Sony Ericsson and DeNA where he engaged in development of social game platform.
Visualization of communication
Cognitee was founded in 2013 by CEO Rie Kawano who formerly worked at business strategy department in Sony or DeNA and also experienced startup in her university days. The firm develops and provides UpSighter, enabling employee training / monitoring excluding cognitive bias targeting enterprise users. The service offers easier communication methods for conveying meaning by analyzing and digitizing speech, sales talk or presentation. Kawano demonstrated the service:
For example, we picked up five Steve Jobs’ speeches and five Japanese politicians’ speeches and then calculated the mean number of topics incorporated into these speeches. In the Jobs’ speeches, topics referring the background of the launch or the reason for the size of the product accounted for 40% of the whole story. On the other hand, as you can see, each politician made mentions of some different topic in one speech and it includes too many content.
Cognitee had initially tackled ‘visualization of thinking’ through a provision of services for meeting efficacy improvement / brainstorming support / planning correction. However, the firm pivoted its approach later; with the basic concept as it was, it changed the application field to ‘visualization of communication’ and then succeeded in gaining strong supports from major enterprises that conventionally had to depend on external consultants for human resource development such as training of salespersons.
Taking pharmaceutical companies as an example, MRs (medical representatives) have to learn about clinical trial data and give presentation to doctors to sell new drugs. You may think they should show a larger amount of data, but too much data will not result in good sales. Our analysis revealed that the MRs who spare much time to present specific data about efficacy, benefit and especially risk of the drug are more likely to be in the high-ranking group.
The same can be said of settlement negotiation with non-life insurance companies. UpSighter figures out modeling of communication style that leads to good negotiation. By applying it to each practical negotiation case and distributing assessment sheets, staffers in charge of these cases can compare their own methods with the model case. For enterprises, operation improvement can be achieved without requiring instructions cost.
Implementation of UpSighter will benefit enterprises in reduction of instructions cost, as well as the fact that each staff can grasp unnecessary / insufficient factors in his / her own speeches quantitatively. The Cognitee team consists of several members including Kawano and Kondo, and they do not carry out outbound marketing so actively but often receive inquiries from top executives of major enterprises who intend bottom-up of internal human resources.
Pursuing scalability without AI
The firm’s name ‘Cognitee’ and its service contents — analysis / digitization of speech is somehow associated with artificial intelligence (AI), but no general AI is used in UpSighter services at this time.All processes required for the service provision of UpSighter are contributed by 150 remote workers in and out of Japan. They carry on transcription of speech, subdividing of the contents, and analysis / evaluation of subordinate relationship of the contents. It is like a form of Factory Automation (FA) for intelligent labor driven by Internet; by dividing its process thoroughly and assigning each small process unit to individual workers, the factory-like quality control and production management is realized.
In the process from original speech content (the input) to analyzed / digitized assessment sheet (the output), more than 10 remote workers relay intermediate products in succession. A worker understands what to do with his / her own task but does not need to have knowledge about the other tasks. With this system, UpSighter needs no consultants with specialized knowledge but can leave all works to housewife workers who passed the test after one-month training.
We did not even know what tendency exists until we stored data about 1,500 matters. However, the details of analysis gradually dawned on us and we have been applying patents relating to our rule or framework in Japan, the U.S and EU countries.
Typical AI learning require Big Data. Moreover, I think it is impossible to trust what human does not understand to AI.
As AI is applied to everything these days, the firm’s approach appears quite fresh. Even AI takes certain cost and time for implementing or tuning, and above all, training data is necessary. UpSighter can extract model case if only it has data about three cases (for example, speeches by three high-achieving salespersons) no matter what kind of business type / condition they are, and assess other speeches based on it.
In general, it is difficult to secure scalability in tasks depending on the person, but there is no limit on the business development of UpSighter which implemented the concepts of FA or remote workers in its work process. As a larger amount of data is accumulated, the Cognitee can replace from man to AI in some process units partially as needed.
Currently, UpSighter is mainly used in sales department that tends to invest a large amount of budget into human resource development in enterprise, but the firm will positively explore the possibility of application or implementation into other departments in the future.
Translated by Taijiro Takeda
Edited by “Tex” Pomeroy