Tango Technologies Inc., a workflow intelligence startup, today announced the launch of an artificial intelligence-powered service designed to capture and automate work through its browser-based software.

The new feature builds upon the company’s existing platform, which features a Chrome extension that automatically captures and documents employee workflows as they perform them. This capability provides valuable insights for standardizing best practices, training new staff and improving overall employee performance.

“For most organizations you want your sales reps to be selling and building relationships with customers,” Babcock told SiliconANGLE in an interview. “You don’t want them mired in internal software, because that’s just not where their expertise lies.”

Babcock said sales representatives spend 65% of their time on administrative tasks, which Tango aims to reduce by automating those processes.

After sales representatives complete a call and review emails, they need to input information into data entry software using transcripts. This process can be tedious, and they often have to do it either during the call or afterward while going through the transcript and accessing their customer relationship management software, such as HubSpot or Salesforce.

Using Tango, sales reps can have the AI begin recording the data entry process for a deal opportunity into whatever software they happen to use, even if it’s bespoke. It will grab information from transcripts and populate CRM fields, or other software, based on context and let the sales reps choose how to locate the information and set parameters on how to process the call.

Users can also tell Tango custom workflow rules in plain English so it can provide guidance in real time to fix errors and offer correct information. This can include specific criteria and guardrails for sales opportunities and moving sales from one stage to the next.

Reps can also set thresholds for when it should bring the human back into the loop, such as if a deal might be lower or higher than is expected – for example if a supervisor or an executive needs to sign off on it.

“We’re not just trying to go end to end, linear automation on a simple consistently like the same workflow,” Babcock said. “We’re inviting the human into the loop to make those decisions.”

In some cases, users can also set a toggle during a step if the AI should automate it or not and that will turn it into a manual step. Babcock explained that this is where he believes this makes Tango’s approach differ from other automation tools.

He emphasized that Tango works transparently and openly when a process is running, so users get to see and validate the outcomes. Everything that the platform does is animated by the Tango agent as a way to bring the users into the process so that they can see and understand what’s happening.

“One of the core principles in what we’re building is transparency,” Babcock said. “We think a lot of agents today are sort of in this black-box category where users will try to run something and it’ll happen in the background and they don’t know what’s happening.”

A benefit of using an AI agent instead of strict automated rules, Babcock explained, is that the company is working on a capability that will cause the AI to “fall back” when something fails in the process execution — for example, if a button disappears or a name has been modified in the user interface. He said that once this is implemented, the agent will go into a “thinking mode” and address how the UI may have changed and how to heal the process before it contacts users on how to proceed.

Image: Tango