Mumbai Analytics Startup vPhrase Analytics Brings Natural Language Generation To Make Reports More Insightful

Mumbai-based startup vPhrase Analytics, founded in 2015 by Neerav Parekh has given a new spin to everyday dashboards and visualizations drummed up by analysts. The core AI-based platform Phrazor, analyses data, derives insights and communicates these insights in multiple Indian languages. At a time when Big 3 of BI world – Tableau, Microsoft PowerBI & Qlik dominate the business landscape, vPhrase Analytics is demystifying the visualization landscape by drumming smart insights alongside text that makes the job of analyzing easier. Phrazor is now deployed across domains –  banks, brokerage firms, healthcare, CPG companies, media and entertainment and makes the job of reporting easier by giving a personalized narrative and highlighting areas that need due attention.

Let’s Talk About the Eureka Moment

Prod Parekh about the idea behind starting vPhrase and he reveals how in his previous role as the founder of a digital marketing agency, he found his project managers spending an inordinate amount of time in the clients’ office, explaining the report. “We were a Performance Marketing agency, responsible for getting day-to-day sales for clients like Dr Batra and Hiranandani. Our project managers had to send reports, to assess how performance pans out. They would take half a day to create this report and eventually when they send this report to the customers, customers would invariably say why don’t you come and explain to me,” he explained. That’s when Parekh came up with the idea of a product that could build up on the visualization end by drumming up text alongside charts. “The only way to do it was automating the task and that is when I thought of building Phrazor. We built a prototype, it worked out and when the prototype was successful I started thinking of potential use cases for such a technology,” he said. Today, the use cases have gone beyond the scope of marketing and expanded to healthcare, BFSI, retail, CPG companies, media and many more.

Understanding Phrazor

Essentially it is a combination of two things – first is the data analysis bit and the other is the language writing skills. “Language writing bit is where computational linguistic comes in and that is where Natural Language Generation comes which is the subset of AI,” said Parekh. The entire product is built on Python and the team took the NLTK language toolkit for building the platform. “But it was not production ready and was not sufficient for the kind of use cases we were doing, so we built a library from the ground up ourselves,” said Parekh. Essentially, the platform takes in data, structured and unstructured, analyses it, comes up with insights and then it communicates those insights using bullet points. The platform is capable of using any language, but currently, libraries have been developed for four different languages -- Gujarat, Hindi, Tamil and English. “For us tomorrow if a client comes from France and has a requirement in French, we can do that too,” he said. According to Parekh, this is the first of its kind product in India. But globally, companies like Arria NLG, Narrative Science and Automated Insights work in the same space. Parekh emphasizes his startup scores over the global competition in a lot of ways:

• All of us have a DIY product and all of us also provide customized solutions. In terms of customized solutions, all of us are almost at par. But in terms of DIY products, our product is superior because of the kind of use cases and complexity our product offers which is currently not possible with those products
• In short, the complexity of use cases those products can handle is far less than the kind of complexity we can handle. You can have real-life cases directly created by yourself using our product without any involvement from our side.

Use Cases

In terms of use cases, it can be bucketed in two broad areas: Internal Communications: There are several clients who use this platform for communicating with their customers. Case in point --  Motilal Oswal Securities use it to send out reports to their own investors. HDFC Bank that has more than 4500 branches in the country use the platform to send out personalized reports to each of the branch managers which gives details about performance vs target. Measuring Performance: FMCG leaders like Unilever and Marico use the platform for Business Intelligence BI to analyze market research data from Nielsen to take key decisions like advertising, price points for a particular region. The task, that was previously performed manually and would take days on end is now automated in seconds by Phrazor. Besides, there are also research and compliance use cases where banks and brokerages have to send out reports to regulators. “These reports are in a fixed format and they come from a fixed set of data. Instead of spending days analyzing data and writing reports, our platform automates the task in seconds,” shared Parekh.

Phrazor to come as a plugin for PowerBI soon

Parekh makes this abundantly clear that in the BI landscape, vPhrase Analytics is not competing with the Big 3 of BI. “These platforms give you the data analysis capabilities and capability of making dashboards but they don’t give you the capability of storytelling. They do not create insights for you,” he said. Currently, the team is in the process of creating plugins for these platforms and the plug-in for PowerBI will be out by March-end. This will be extended to other players as well like Tableau and Microsoft Office as well. On the client-side, the Mumbai-based analytics startups count some of the biggest names from banking, FMCG, retail and media world like HDFC Bank, Motilal Oswal Securities, Unilever, Accenture, consulting companies like Accenture, Wipro, TCS, Viacom18, Sony, Moneycontrol.com. Plans for expansion are also afoot and soon the startup will have their second office in Singapore, revealed Parekh.