Adobe will today announce the introduction of Virtual Analyst, powered by its Sensei AI.
The analyst runs 24/7 in the background to monitor data and detect and find the root cause of anomalies in online activity. This replaces the painstaking process of an engineer or data team manually searching analytics reports for insights, which can diminish in value over time.
“Insights we do believe have a shelf life and to have a system be automated and can handle these on its own is really key, I think,” Adobe marketing manager Nate Smith told VentureBeat in a phone interview.
Sensei was first introduced last fall as an artificial intelligence service trained by massive amounts of data gathered from Adobe Creative, Marketing, and Analytics cloud software.
Sensei can do things like auto-caption images, deliver data insights, or talk people through how to use Adobe software. Adobe ultimately wants the AI to also train novice creatives how to use Photoshop and other software like a professional.
In time, as future versions of the analyst are released, Sensei will remove humans from the loop to automate the completion of actionable insights, Smith said.
“That is the direction we’re going now. We’re in the initial stages of that with integrations with other products,” he said. “We’ve been working closely with the Adobe Target team so, for things like basic A/B testing, those are things we’re exploring right now, and we’re going to continue down that path.”
While in beta for the past six months, the virtual analyst has been able to unearth things like intellectual property theft as well as human or machine errors that cost companies money.
The alert system for Virtual Analyst solves a problem that existed in the previous version of Analytics Cloud, Smith said. By learning a company’s seasonality and patterns, the Virtual Analyst is able to reduce the number of false flag push notifications the team that manages company marketing data receives.
Also on the way, the analyst will incorporate Sensei’s natural language processing in order to function more like Alexa or Cortana, in much the way Sensei now works with Photoshop.
Other enterprise players in the bot-AI-team collaboration space, like Hipchat and Cisco Spark, are also working to develop or have already developed voice assistants to deliver answers to questions about company (or competitor) data in a meeting or office setting.