Skim AI explores and solves the problems a researcher faces in the times of information overload. The product enables researchers in accessing machine learning tool to make better, more informed decisions, faster.
The client came up with an idea of creating a tool to help readers, researchers in saving their time while they search for news and information over the internet. Basically, we were supposed to create an A.I.-powered research assistant that extracts information from a webpage and processes its content through machine learning and summarizes it in a short paragraph.
The idea was to provide accurate information with facts, present in the web page in a shorter format to save readers’ time. For this, we first needed to identify the sources of information of the target audience for trial versions. We needed to set up a rough estimate of several websites to be used in machine learning process. The source of information, i.e., the websites must range through different categories and usage such as news, finance, asset management, academic research, consumer insight research, marketing insights, etc.
While access to the information is one thing, saving it in a user-friendly format is another, we decided to extend the features so that users can save the info over Google drives, Dropbox accounts. One offline feature was also to be added where user can upload a PDF and the tool shall summarizes its content.
Innovify started with a user-friendly and flawless design which was approved in the first sitting itself. Generally, chrome-extensions follow the same boring color-codes and design, however we reversed that trend in this case. Some design inspiration was taken from screen-capture tools, and it was decided to use a unique one in this product. A third-party login option was given along with tri-option of getting, facts, quotes, or summary from the article. One just needs to install the extension and visit any news website and voila! All the necessary information is just 1-click away..
Here is the most exciting part of the project. The reason why Innovify took up the challenge, and the reason this product is quite unique. We used machine learning in summarizing content in three categories, i.e., facts, quotes, and summary. Our process helps in identifying valuable pieces of content with color-coded quotes and facts throughout a given article. All this happens in a fraction of a second. This model was put to intense testing as we never wanted to leave a discrepancy even for an isolated case. The result is something to boast of..
As we have a flawless machine learning process to summarize content, it was inevitable that we add few features for the users. Essentially, we added an option to clip a summary, store a said article in user accounts, organize saved articles, and retrieve the information when in need. These features were introduced by keeping research scholars in mind, they tend to return to critical information which helps in their work.
Using machine learning to process content of around 50,000 websites every day to present a summary with accurate information is itself a quite demanding task. Innovify took this as a challenge and got started with dividing the whole pie into smaller pieces of work. Once we accomplish one task, it was time to test, test, and test until the outcome is achieved precisely. No amount of discrepancy was left unchecked as emphasize on our automation testing service to get the best for our client. The client had a CTO at their end who could understand the technicalities and complexities of the tasks, which was a blessing for our team. Innovify released the extension within the specified time and continues to extend support, adding features according to user feedback since its release.
Co-founder & CTO, MD – India