A company on a mission to save people drowning in data

In 2014, the year we founded Skim Technologies, the world was largely the same as it is today. Granted a few curve balls, namely Brexit and Trump have changed the political landscape, but technologically the way we search for, consume and communicate information is still the same. We share links in emails or chat; our favourite news app tells us what’s happening in our echo chambers, that are then compounded by the social platforms we communicate on with our friends. We have access to so much information that we’ve created an ever-growing thirst for knowledge that drives the creation and re-creation of information, resulting in a world that is now drowning in data.

When Lloyd Jennings and myself founded Skim Technologies, we could see this problem needed to be tackled in an entirely different way. We wanted to fundamentally change the way people discovered, consumed and shared information, in a way that would enable, not disable a person. We wanted to make consumption and communication of information more efficient. We founded Skim Technologies in the belief that we could use Machine Learning and Natural Language Processing (NLP) to do this, and to ensure we always had a guiding North star to help keep us focused, we inked the company’s overall mission:

To save people time searching for and consuming relevant content.

Quite early in our research, we started working with text summarisation using NLP, we thought that summarisation would be a good way of making information easier to consume. Initially we wanted to take a large ~5000 word article and summarise it down to within 5% of the original text. Although the results were good, it only accounted for about 30% of all web content, and therefore wasn’t a total solution. We started to consider what summarisation would mean for other content types, take a homepage or a listicle, and think about how we can faithfully represent that page in a digested format. That’s when we developed the idea for our new information format.

A skim: A snippet of a webpage, in card form, that’s generated automatically, by extracting only the most important parts of a page, and presented in a digestible format.

The skim allowed us to take a variety of content types and present the information in a more consumable, bite-sized format — with the aim of saving a person time, part of the company’s mission.

As we developed the Skim API, we wanted to get first hand experience building the API into our own product. That’s when we designed and built Skim.it, which started out as a bookmarking app which then pivoted into its current form of a speed-reading plugin that’s used by thousands of students and teachers across the world. We’re really thrilled to be making a difference to a sector that embraces technology so fervently, and whose needs can be met by this more efficient form of information consumption

The Skim API is also available to developers to build into their own products, as we encourage product owners to think about adopting this new format of information to help their users save time, and work more efficiently. Whether that’s someone searching a database, two colleagues sharing links in a chat window, or a person reading an article in a mobile app on the train home.

Although we feel the skim is a very good first step towards revolutionising the way information is consumed and shared, it still doesn’t satisfy the overall mission of the company.

We recognise that we’ve only begun to tackle the issue of information consumption which doesn’t solve the problem of discovery, or relevance.

With that in mind we’ve embarked on further research into Machine Learning and have begun development of topic classification and alerting API’s. We’re currently building this into a free chatbot, Triggers.ai, that alerts a user to a ‘trigger event’ based around topics and entities. Due to the nature of our technology that means a developer could build an alerting system within their own platform, drawing on data from random unstructured data sets.

To give you an example of how we see our Triggers API being used in other products. Take Supply Chain Management Systems — a Logistics Manager can set up and receive alerts about a trade route disaster that needs their attention. Or, take a CRM — a sales manager receives news about their competitor signing a partnership with a customer whilst on their way to a meeting. Or in a consumer travel app — it might be a news alert to a traveller in a foreign country.

We’re still experimenting at this stage, however we feel there’s a lot of value in using our existing unstructured data extraction and summarisation tools to develop interactive, and perhaps even proactive bots or virtual assistants.

We believe the Conversational Interface will become much more prevalent in our daily lives. It’s easy to be overwhelmed by the idea, but it’s the most natural form of communication. We’re very excited about designing our technology to work in these new interfaces. Making the discovery and communication of information even more natural, ultimately saving people even more time, searching for and consuming relevant content.

Visit Triggers.ai to join the waiting list and test out our new bot. Or follow this blog to get the latest company updates.