Introducing HyperScience: Better, Faster, Cheaper

Back

While the promise of artificial intelligence has experienced a resurrection over the last few years, we have found ourselves in a phase where “the future is already here, it is just not evenly distributed.”

Currently, almost all of the benefits of artificial intelligence accrue to only a small handful of companies, like Google, Facebook or Baidu. Those companies have big ambitions, deep pockets and access to gigantic amounts of data. Perhaps most importantly, they are able to attract many of the best machine learning engineers, by offering both high compensation and an opportunity to work on a range of fashionable problems, from large-scale image recognition to self-driving vehicles.

This leaves everyone else – both Global 2000 companies and government institutions – in a tricky situation. These large organizations will not be able to attract top machine learning talent, at least for the foreseeable future, both because they won’t offer them the same levels of compensation and also because many of the problems they need to solve through AI appear, at least on the surface, to be hypnotically uninteresting.

What Global 2000 companies and government institutions do, however, often have is very large amounts of data. They also have clear use cases where AI, through a combination of automation and augmentation, can make a tremendous difference to their bottom line.

At HyperScience, we believe that large enterprises present an untapped ocean of opportunity for AI and machine learning applications. That’s the opportunity we are pursuing: to help large enterprises perform business-critical functions better, faster, and cheaper, through artificial intelligence.

Our initial focus is on the world of enterprise back-office automation.   As an example, our HS Freeform product enables our customers to have machines read and understand any kind of text, considerably speeding up their response and processing times.

Sometimes that text is handwritten.  For all of the extraordinary march towards digitization over the last few decades, there is still a tremendous amount of handwritten information in today’s large enterprises. While a lot of attention in AI has gone towards speech recognition, handwriting recognition — a closely related challenge — hasn’t experienced the same level of progress, and we have been working hard on this extraordinarily complex problem. With this HyperScience capability, when banks process mortgage applications, insurance companies ingest medical information, or governments interact with their citizens, they’re able to serve their customers orders of magnitude faster.

In other cases, the challenge will involve understanding and parsing text to extract meaning — life insurance underwriters (and their back office teams) often have to read through long applications, which contain a lot of information, for example about one’s medical history.  This creates massive pain and inefficiency – beyond the considerable amount of time (and caffeine) required, the insurance company can’t offer their customers a policy any faster than the underwriters can read through the backlog.  With HyperScience, it’s not just that insurers can do the same job better, faster, and cheaper than before, it’s that insurers can offer their customers a quote almost instantly. Faster isn’t just “faster.” it means our customer is doing business several days in the future relative to their competitors. They can offer their customers a quote while their competitors offer a long wait.

Over the last few years, we have been working on developing and deploying AI technology to solve these types of vexing problems.  Until today, we have not spoken much about what we do but the time has now come to be more public.  We are launching a website today that will provide more information about our products (demos are available!) and the terrific group of top notch machine learning engineers we have built in both New York and Sofia, Bulgaria.

We are also excited to announce that we have raised $18M in Series A financing from FirstMark Capital and Felicis Ventures.   We had previously raised a seed round from Third Kind Venture Partners (Shana Fisher), AME Cloud Ventures, Slow Ventures, SV Angel, Acequia Capital and Box Group.  Our investors’ faith in us and our mission means a lot to us and we’re grateful for their active involvement in helping us build the team and build the business.

<>

Please ensure your neuro chip is activated, and nod to continue...