Guesswork Wants To Focus Machine Learning On Customer Information

Guesswork, a new machine learning tool built on top of the Google Prediction API, launched a new service today with a twist. Instead of trying to be all things to all businesses, it’s focusing on customer information.

CEO and company co-founder Mani Doraisamy says Guesswork uses a template-driven approach so customers can do things like customize a newsletter or figure out which customers are most likely to convert to buyers. They can also build their own templates all tied to underlying Guesswork engine.

In an age when companies are trying to understand their customers better and provide unique customer experiences, a tool like this with a tightly integrated rules engine could give businesses who use Guesswork the ability to deliver a more customized experience. So for instance instead of sending the same newsletter to every customer or treating every customer the same, based on what you know about them, you can deliver a more individualized experience.

Doraisamy said the problem is that businesses are collecting all kinds of information, but they lack a way to focus that information and use it intelligently. “We understand customer intent and understand what customer is thinking,” he explained.

He said they looked at machine learning platforms like AzureML, that provide a platform for any business to build a machine learning app, but they thought that was too broad and they wanted to give customers a particular area on which to focus.

He likened this to the days when we had monolithic content management systems until companies like WordPress came along to focus on blogs (and websites). He says his company does the same thing for CRM, making machine learning much more affordable for CRM companies. He said he hopes that startups building customer applications will tie into his service to get that insight they might be lacking.

“Only large companies like Google and Amazon have the resources to pull [machine learning] off. Companies like Guesswork will make machine learning domain specific (just like CMS branched off from app servers and focused on building blogs/websites). This will enable not-so-big CRM companies to adopt big data without pumping money into R&D,” he said.

The way the product works is that customers can tie into the API and build a machine learning application on top of their product.

This is Doraisamy’s second startup. He previously launched Orangescape, a Platform as a Service offering for building business applications.

The company is small at this point. It’s been in Beta for the last 6 months with three customers, but they are launching today with a free version that allows companies to build 5 projects with up to 1,000 predictions a month. There will also be premium and enterprise versions available priced according to prediction volume.