Want to solve a problem quickly? Have a “chat” with DataChat’s robots
When any firm comes to a road block during its project, it’s often very difficult to quickly solve the issue before the deadline.
Jignesh Patel, founder of DataChat, wants to change that.
DataChat is a spin-out from the UW-Madison’s Department of Computer Sciences, and is based on advanced data analytics and machine-learning technologies that have been developed at the university in Patel’s artificial intelligence research group. The company was profiled as part of a series written by UW-Madison students.
Patel and his team are working on a software that allows firms to quickly and efficiently gather the necessary data and information to create effective problem solving.
“We want to minimize problem solving time for companies that need to meet short deadlines for complex projects,” said Patel, who’s had a successful track record of entrepreneurship, DataChat being his fourth startup.
The company trains and deploys custom chatbots in the cloud that allow decision makers to get insights from their data by simply conversing with those chatbots. In response to the conversation, the bots search through a vast library of code patterns, assembling the appropriate pipeline of data analytics and machine-learning methods for the task at hand. Results are seen in the form of charts and pictures, making it easier to see the patterns that are hidden in the data.
“Conversation logs across the enterprise can be fed back to the DataChat platform, which can then be used to retrain the bots empowering them look for deeper and richer patterns,” Patel said in describing DataChat’s proprietary technology. “Thus, the bots become even more powerful as they are deployed over time. Also, we can customize our bots to fit any firm’s specific needs.”
As an example, consider a marketing department that needs to analyze data from various marketing channels, such as Google analytics and social media analytics, to understand, in real-time, which campaigns are working and why.
DataChat can create customized chatbots for that enterprise to allow decision makers to interrogate the data by simply asking questions about campaign performance, for example. The chatbots can identify the under-performing campaigns of interest, and visualize the data in ways that can inform better, faster decisions.
“DataChat is currently in stealth mode, and is available for beta trials with a few select customers,” Patel said. Thus far, he said, clients have been satisfied with the work.
“Medium and large enterprise organizations often struggle to obtain insights from large datasets. Current software solutions involve complex set-up and offer limited reproducibility, meaning insights gathered from data often can take days or even weeks to obtain,” Patel said. “DataChat aims to change this process by training chatbots to perform complex analysis tasks on customers’ data. Now, instead of dedicating hours or even days to answer a single question, large data sets can be queried multiple times in minutes, enabling businesses to make rapid and informed decisions in real-time simply by chatting with our chatbots in natural language.”
Patel added: “With DataChat, decision makers can leverage the power of data analytics, machine learning and AI, and in real time, without having to write, or learn how to write, code. With DataChat, natural language conversation is the “programming language.”
Once the service is fully launched, firms may sign up using a subscription model in which they are chargedchargedk a monthly fee to use the chatbots. The chatbots can be deployed in public or private clouds. The bulk of the company’s revenue is expected to come from the monthly subscription model. Some customers may require specific customizations to our bots, and the company plans to use a consulting rate for that task.
DataChat’s vice president, Aaron Rolich, presented to potential investors earlier this month during the Tech Council Investor Networks’ track of the Wisconsin Early Stage Symposium.
By Kyle James
James is a student in the UW-Madison Department of Life Sciences Communication.