In this Q&A interview we discuss My-Take’s latest innovation TextCoderAI, with our founders, Todd Hoskins and Rich Armstrong. TextCoderAI is an artificial intelligence based solution that is simplifying the coding and reporting process of insight community analytics. It utilizes natural language processing, historical data, and statistical tools to analyze community member comments with lightning speed. In turn, greatly optimizing the coding process overall.
What prompted its creation?
Hoskins: With research as a whole and insight communities in particular, most professionals would agree that understanding the ‘why’ can be as important as the ‘what’. To that end, open-ended questions are extremely helpful for companies trying to make business decisions. However, coding these responses is incredibly time intensive and demands skill, discipline, and a bit of art. So, if you combine the importance of open-end coding with the time demands - it presents a big opportunity for innovation.
Why not use technology already available?
Armstrong: We started My-Take almost nine years ago, so we have a large body of experience when it comes to data analysis and coding. We found the solutions available in the marketplace weren’t suited for the demands of insight communities and our customers’ needs. So, going back more than three years ago we began experimenting with different approaches that would provide automation, but also maintain a high level of accuracy.
When did TextCoderAI launch?
Hoskins: We filed a provisional patent in 2018 and began utilizing it in the marketplace in the back half of the same year. The performance of the technology has exceeded our expectations.
How do the My-Take text analytics function?
Armstrong: One of the important aspects of our approach is that it is multi-faceted. We utilize natural language processing, historical information within the respective community, and a pilot set of human coded data. We also have a statistical tool to help validate the results. The two greatest aspects of the technology are: the researcher experiences a subset of member comments directly through the pilot without having to individually code all responses, and it also is a system that is continually improving as more studies are completed within a community.
What can a researcher expect in actual use?
Hoskins: There will be big differences in speed. Studies with 400 responses are coded in less than half the time. The impact increases even more with larger sized communities. Secondly, you can expect accuracy. We understand that there is a high bar in terms of client expectations in the analysis of their data. We’ve built that accuracy into the system and provide a means to check the AI results.
Who is most likely to benefit from this technology?
Armstrong: Any My-Take customer will benefit at some level. Clients utilizing us for full service will continue to benefit from our overall solution which is an excellent value, as we seek to automate many components of the process. There is also a benefit for clients of receiving even better reporting as our team will be able to focus on completing more high-end analysis as well as delivering more compelling stories and visuals around the data. SaaS, or DIY, customers will see all the speed and accuracy benefits that our team enjoys on the full-service side. With some quick up front training clients will be able to master this technology in short order.
Armstrong: Our innovation team never stops. First, we will continue to enhance TextCoderAI. In addition, we have a lot of amazing technology in development which will continue to expand the definition of an insight community and help make our client researchers more and more valuable within their organizations.