CEOCFO-Members Login

December 5, 2016 Issue



Tactical Cognitive Computing Platform helping Finance, Healthcare and Knowledge Enterprises Automate their Workflows



Praful Krishna

Chief Executive Officer




Interview conducted by:

Lynn Fosse, Senior Editor, CEOCFO Magazine, Published – December 5, 2016


CEOCFO: Mr. Krishna, would you tell us about Coseer?

Mr. Krishna: Coseer uses tactical cognitive computing to automate language driven workflows at enterprises. Think about any kind of workflow, where the input is language or unstructured text and some decision or some task or some process has to be completed based on that language. Coseer ingests such language, either from consumers, from the internet or from data bases and completes one of these tasks or workflows automatically.


Think about what an intern does in a typical company. He or she would read the language, make their own assumptions, extract some idea, build some hypotheses and then, without having benefit of judgment or creativity, the intern tries to complete a tedious or mechanical process. You would never trust an intern for your corporate strategy or M&A decisions. However, you can rely on the intern to research things for you, to summarize things for you, to build a scorecard, to go read some stuff and extract important values for you. Think about what you could do if you had a thousand such interns on your team. Coseer can manage all of these processes.


CEOCFO: How are you able to do it effectively? It seems that speech recognition has been around a long time, but never quite where people feel confident.

Mr. Krishna: When I say language I mean written language or unstructured data. The speech recognition is just an aspect of converting spoken language in to written language. It should already be written when it gets to us.


Our technology has essentially been designed to emulate the human thought process as it relates to language. There is a set of five different things we do for that. For example, humans never deal in the keyword domain or just the spoken word or written word domain. Humans deal in ideas. That is the first thing we do. Whenever language comes to us we take that from the articulation, to certain ideas and manage it in the idea domain. The fact that our technology can process ideas instead of keywords gives us a huge leg up compared to other companies.


CEOCFO: Would you give us a simple example?

Mr. Krishna: Let us say that you have two words “red ball”. Then there is angry bird. There is the baseball ball. There are the Christmas decorations. There is the planet Mars. The ability to connect the phrase “red ball” to each of these objects that I described with some level of association, is what we call dealing in the idea context. None of them is called a red ball, but each of them is spherical, each has red tinge or the red color to it. That is just an example of dealing in ideas.


Another example would be the country of Great Britain or the UK. Every dictionary or ontology maps them together, but how do you recognize them as ideas, saying that these two are the same things. When we read through a corpus and find that these two behave very similarly, we conclude they must be the same idea. That is what we call thinking in the ideas domain.


CEOCFO:. Would you tell us a little bit about the different products, how they work, when they are used and who might use them?

Mr. Krishna: If you think about our product suite it is really a continuum of our technical capabilities. Each product is based on top of one another. However, because our clients need some basic functionality, we productize it and bring it out.


CEOCFO: What are the names of the different products?

Mr. Krishna: I think our most evolved product is called DocSeer or document seer. DoSseer is really what reads documents. One document or millions of them, it is the same thing; it does not matter because it is a computer after all. It will read a set of documents and then form a cognitive model, which is an understanding of what is written in the document in the computer language. If it is a legal document, then it means arguments and causalities and implications. If it is a healthcare document then the treatment plans or the drugs the physician has prescribed, etc.


This cognitive model is used to power multiple interfaces. Once you have this cognitive model you can power a chatbot. You can power a mobile interface which is completely customized and interactive. You can power a web interface or a search which is very relevant, to the point and very personalized. I will give you a specific application for this. If you have to take a legal document which is five hundred pages and you are looking for something very specific, instead of going through each page and trying to figure out where it is you can ask DocSeer a question. DocSeer will understand your question. It already has the cognitive model so it can find that answer in a very natural language way.


DocSeer will just make your document come alive!


CEOCFO: What other applications do you have available today?

Mr. Krishna: You have multiple things that build up to it. There is a LongSeer, which is a summarizer, which can take all these documents, ask the questions itself and tell you what is going on, in ten bullet points or a two page summary, whatever you need.


Similarly, another product called InfoSeer basically puts out all of the information contained in there in a structured format so that classical structured big data technologies can take over. You can take a document, extract all the information, put them in a structured format, then you can do statistical analysis, use big data tools and everything can come to life. We enable all of that.


Similarly, if you can figure out which is the sensitive information and which is non-sensitive, you can go back and redact the sensitive information. That is one of the big projects that we are doing where their information is so sensitive they do not even want their colleagues to know about it. We are implementing a product called RedSeer, where we will be redacting all the sensitive information from documents. We will be ingesting it, understanding it and redacting it and then eventually making the documents available when all the sensitive information is cleaned out, but a sense of the structure they used or the strategy they used is still there.


CEOCFO: How do you help clients get comfortable with this approach?

Mr. Krishna: I think that is where we have to earn our trust before we earn our revenue. That is why our engagement with a typical client is multi-phased. Before we go live we spend three months of what we call the co-build phase. There we are developing the solution, training our platform for that particular problem along with the client and three months later if the client does not like it, or the client does not trust the solution, then there is no further commitment.


Even after the deployment we periodically audit the results. We pick up some samples in a very standardized process and run through manually. If we identify problems where some tweaks are needed and we go and do it, because we are looking for 95% to 98% accuracy. No one trusts a black box, so we try to stay as far away from the black box as possible and be very transparent with the client. That is how clients trust us. As I said, we have to earn the clients trust before we can earn any revenue from them.


CEOCFO: What types of companies have turned to you and where do you focus your outreach efforts? What industries?

Mr. Krishna: We have been very blessed in that all our leads are inbound. There is a Contact Us link on our website and that is responsible for the entire pipeline and entire business of the company. It is, of course, a great thing. What it does not let us do is be very selective of our clients or plan very focused outreaches. In general we work in finance, healthcare and what we call knowledge sector, which is a consulting, publishing, tech or legal where there is a lot of language driven things. Industrials is also an emerging sector for us. Those industries are becoming our sweet spot. We usually find that companies with more than $100 million in IT spend have the relevant budget to be able to engage with us, because with a smaller company there is that trust factor and that of justifying the spend.


CEOCFO: Are there many companies that are trying to work in this space? What is the competitive landscape?

Mr. Krishna: Artificial intelligence is what everyone is talking about in the main street, but Silicon Valley is already starting to talk about cognitive computing. There are three different segments. One is general purpose startups like us, which are very few. There also are many companies that focus on vertical applications, like healthcare diagnostics or financial fraud management. They prefer specific applications because accuracy goes higher if you are focused. That is one set.


The second set is a general purpose platforms by IBM or Microsoft; IBM Watson is the one we compete against the most. They are putting serious money behind it.


The third set is the independent software vendors powered by the API’s of these big companies, such as Watson, Microsoft and Google as well. Therefore, many software companies are basically repackaging those APIs and training them and trying to go and sell them. We do not worry about these companies because we know that training a super computer like Watson needs the resources that a super computer deserves. It needs money, time and people. Therefore, I may be wrong, but I think that at least for the next two or three years we will not see any serious competition from the third segment.


Today, in most cases we compete against nothing. Whenever we are in a competitive situation it has been against Watson.


CEOCFO: What did you learn in Harvard Business School that has been helpful in creating the business?

Mr. Krishna: I think business school taught me leadership without hierarchy. In a startup such as Coseer, there is definitely the need for outstanding team members, plus the most valuable people for Coseer are our scientists.  They are distinguished people in their own fields. Someone has thirty years experience in artificial intelligence. Someone is a professor of computer science and an expert in algorithms, and another in big data. This is where the technology has come from. I think what I learned at business school was how to get these guys motivated and actually be focused on the task without being the hierarchical leader.


CEOCFO: Put it together for our readers. Why is Coseer noteworthy?

Mr. Krishna: It is because Coseer makes possible the utopian scenario of cognitive computing where computers are taking care of everything mundane and you only want to focus on creativity and judgment. It brings it from a boardroom decision, where the CEO has to write a $10 million check just to try it out, to the offices of VPs and directors of enterprises. We practice a brand of cognitive computing called tactical cognitive computing, which means you can touch it, you can try it and it is within your budget. It is applicable to small problems. We are not out there curing cancer. We are out there automating your workflows. The whole idea of taking the cognitive computing a bit down market, more affordable, more tactical, is something that very few companies are doing and Coseer is lucky to be one of them.


“Coseer makes possible the utopian scenario of cognitive computing where computers are taking care of everything mundane and you only want to focus on creativity and judgment. We practice a brand of cognitive computing called tactical cognitive computing, which means you can touch it, you can try it and it is within your budget.”- Praful Krishna





Praful Krishna

617 412 5405






Any reproduction or further distribution of this article without the express written consent of is prohibited.



Cognitive Computing and Big Data Analytics, Coseer, Workflow Automation Software, CEO Interviews 2016, Praful Krishna, Tactical Cognitive Computing Platform helping Finance, Healthcare and Knowledge Enterprises Automate their Workflows, Technology Company, Processing, Automate repetitive workflows, language driven workflow, identify a repetitive, automate workflows on premise or in the cloud using tactical cognitive computing, Coseer Press Releases, News, Tech Stock, Companies looking for venture capital, Angel Investors, private companies looking for investors, cognitive computing companies seeking investors, workflow automation companies needing investment capital does not purchase or make
recommendation on stocks based on the interviews published.