Obviously AI introduces natural language predictive analytics service
A few days back, Joe McKendrick wrote about an IBM study showing seven business areas that are ripe for AI. It's just part of the onrush of developments that are making AI mainstream. And so it's easy to get jaded when you hear an announcement of yet another AI-enhanced tool.
So when we saw an announcement from the tiny startup Obviously AI, we were expecting to see yet another refinement in what we term guided analytics. That's analytics where machine learning is employed to help you choose the best data sets, ask the right questions, and frame the narrative with the best visualizations. Obviously AI is a new cloud-based service that enters general release today that delivers predictive analytics without requiring users – which in this case are business analysts – to code.
Obviously AI brings no-code predictive analytics to conventional structured data, which can include standard database sources or file formats such as CSV.
The key is a combination of natural language processing and AutoML-like capabilities that let you write a query down in standard English (you can skip the SQL), choose the data set and column(s) you want to target. Then the service performs some data cleansing to normalize the data, internally runs a handful of competing algorithms to select the best one, and then spin out the results. You can have the system provide an additional assist in recommending the best columns of data to use. As part of the process, the service can recommend the best columns to use based on which ones have the most impact on the results. They have benchmarked performance to crunch 10 million records and produce a prediction report within 30 seconds.
Clearly, Obviously AI is not the only provider that is offering natural language query. For instance, Tableau's Ask Data provides a similar natural language query service. So does Power BI. But neither of them go as far as Obviously AI and extend the process to optimizing and running predictive machine learning models.
Obviously AI is still a tiny operation (for now the staff is less than 10) operating with seed and angel money – it has not yet completed its first major funding round. They are targeting small-midsize businesses, but for now are counting primarily on word of mouth and content through channels such as Medium. They have not yet formulated a partnering process for go to market, but they are likely to pinpoint cloud-based data warehousing services as their likely target.
With the emergence of natural language query, guided analytics, and AutoML services, it was bound to be a matter of time before someone put the pieces together to deliver an automated service that went from query to prediction.