No Jitter is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Agent Assist, Powered by Talkdesk iQ

At ICMI Contact Center Expo, being held this week in Ft. Lauderdale, Fla., CCaaS provider Talkdesk announced the availability of Agent Assist, which it describes as an intelligent advisory tool that supplies data-driven real-time suggested actions to aid and guide agents in delivering quality customer interactions and outcome excellence.
Readers with a good memory may recall an earlier Talkdesk offering, SalesAssist. During a November 2017 Salesforce Dreamforce session, Gadi Shamia, then-COO at Talkdesk, talked about a next step in Talkdesk AI, AgentAssist. Similar to SalesAssist, this tool would use word transcription, Salesforce Einstein for sentiment analysis, and current context to suggest next best actions an agent should take with a customer.
With a new head of product at Talkdesk, Charanya Kannan, and Jafar Adibi as head of AI and data science, the “next step” has come to life in a bigger and better way than initially discussed.
With Agent Assist (now two words, not one), the customer relationship management (CRM) software the customer is using doesn’t matter. A customer can be using a custom CRM, Salesforce, Zendesk, or anything else. “It doesn’t matter because we're using our own Talkdesk iQ AI versus Salesforce Einstein AI. And we are surfacing this within our own agent interface versus showing it within the Salesforce agent interface,” Kannan explained.
Proving that point, the first Agent Assist customer -- EmployBridge, a staffing company with 600 contact center agents -- is using the tool with a Microsoft Dynamics CRM system.
Like other agent-assist solutions, Talkdesk’s Agent Assist tool tags company, vertical, or application-specific words as it transcribes them, while the customer is speaking. During the pre-briefing Adibi discussed the individual tasks that Agent Assist performs based on the tagged information:
  • Real-time actions and shortcuts -- Adibi distinguished the two by saying that real-time actions happen automatically, and shortcuts are hyperlinks that take the agent out of the agent interface, to another system, to find an answer or perform a task.
  • Smart Notes -- This feature automatically places the tagged elements of a conversation in the call notes. This means that agents don’t have to type as much, if at all. An agent can still edit the notes, but if they’re sufficient, he or she could just hit submit.
  • Automatic data entry -- Certain pieces of data in the conversation can be automatically plugged into the CRM or elsewhere. As with the other features here, Adibi explained, the maturity of automatic data entry will grow over time. Initially it will be tier one types of entries, but that will grow and expand using machine learning.
  • Escalation assistance -- Similar to Smart Notes, escalation assistance is used when an agent transfers a call to another agent or to a supervisor. Pre-populated notes from tagged data are sent along with the call, eliminating the need for a “warm transfer” conversation. Ideally, you’re getting the benefits of a warm transfer without the time it usually takes to wait for the second party and verbally present call notes.
  • Real-time analytics and error detection -- Adibi explained that this is most applicable when there is any part of the call or conversation that requires some sort of scripting, whether it's a mini-Miranda or if agents are required to open or close the call in a certain way. Error detection can catch the misses or half-attempts, alerting the agent to do what’s required.
The graphic below demonstrates how these capabilities might be combined in a customer scenario.
Talkdesk Agent Assist example
The final Agent Assist attribute Talkdesk stresses is data security. As Adibi explained, “Because we're not a data collection company, like a Google or someone else, all of the lessons learned in these conversations stays within Talkdesk for Talkdesk use, or partner or customer use, of the information.”
As greater numbers of contact centers start deploying artificial intelligence, expect company lawyers to be reviewing terms of service from CCaaS vendors a little more closely then we’d do when wanting to install the next version of iOS or Android.