My digital watch beeped. 12:00 a.m. It flashes, reminding me it is no longer Friday night, and I am starting to spend my Saturday in the office as well. I glanced back at my computer screen, wondering when I am ever going to finish manually typing and copy-and-pasting data into an excel. I looked around, ironically, I am not alone in the office. Many analysts across industries have days similar to mine - spending hours trying to find that number in the prospectus, verifying a number quoted by a data provider, copy and pasting information from a scanned PDF. There is so much information out there - online, offline, internal, external, etc. Yet, it's difficult to find, consolidate, and analyze such data to drive meaningful insights.
The problem can be generalized and broken down into three-folds:
- Data format: For the past 4,000 years, humans have recorded data in books, documents, and most recently, webpages. Most of these data are unstructured and cannot be readily used for analysis.
- Data consumption: Over 90% of the world's information lies within the deep web and remains unutilized. While most people use Google – it only surfs 4% of the Internet and is built for consumers. Hence, we see many analysts or professionals still spend the bulk of their time reconciling and collecting data manually.
- Data-driven actions: many businesses have not figured out how to use data to drive insights and incorporate those into our workflows. The possibilities can be limitless if we use data well - verifications, comparisons, predict, risk manage, and more.
That's when I decided to quit my banking job of 4 years to tackle the problem of how to make the data analysis process faster and easier, with a focus on finding new and interesting alternative insights instead. I spent a year picking up coding again, starting from the basics I studied in school and learning about the latest trends and technologies.
My idea began in a relative niche space - data for credit analysis where operating data is difficult to find. However, after speaking to countless mentors, peers, and professionals, I realized the problem is much bigger and applicable to many financial institutional processes, research, and analysis. Further, many corporates in Hong Kong also face the same problem.
Wizpresso's goal is to solve these 3 problems and help businesses extract valuable market insights from data, improve workflow, and make informed strategic decisions.
At the end of 2018, I pitched my idea to Cyberport, where my concept was recognized and I joined their incubation program. In early 2019, I recruited my founding team Calvin, KT, and Zac who have the technical experience as well as business experience to build and run technology platforms. Together, we built our first product – a business-oriented search engine for business insights and data. Using NLP (natural language processing) technology, we can analyze PDFs, websites, images, forms, etc. efficiently and automatically, and present data in graphs, visuals, or outputs that allow businesses to analyze and make decisions. We closed our first client within 3 months, an international investment banking firm, helping them drive insights from e-commerce data. This proves the significant need for such a product. We eventually applied and built upon the same technology to subsequent clients including international law firms, and local corporates.
Within a year, we went from no product to having over 200 corporate users, and petabytes of data on our platform with sources globally, generating insights, improving workflow, and saving time. We are also recognized by etnet as an Outstanding Alternative Data Platform in 2019. However, this is only the beginning of our journey. We hope to develop an ecosystem of individual and businesses that actively uses data to generate new insights and improve workflow. We hope to build a community where users will give us feedback as to how to continue to improve the data analytical experience.