Bots have come a long way since their inception, but consumer trust still lags behind. In our State of CX Trends 2021 55% of consumers didn’t think bots could provide better CX and 54% have had a bad experience when interacting with a bot. The legacy of 2016’s ‘Year of the Bot’ was too many badly designed bots, making “I’m sorry, I don’t understand” a far too regular occurrence. The result; consumer perception still being bots aren’t all that useful.
Automation doesn’t instantly translate into bad CX, but there are two main problems that cause the most issues:
- Bad conversational design means customer are stuck in one-directional flows with no ability to navigate beyond a pre-defined path.
- No mechanism for bots and agents working in tandem and handoff customer conversation to each other when applicable.
How does Agent Assist Work?
For brands that use private messaging channels to have conversations with their customers, it’s almost like a personal assistant for contact center agents. Agent Assist means they can benefit from a two-way collaboration between agents and bots, when the collection of personally identifiable information is required, or simply to run through a fixed sequence of steps with a customer. It allows agents to focus on other, more complex conversations while Agent Assist is working in the background, with the bot handing back when the task is complete.
Three Ways Brands can Deploy Agent Assist
1. User Authentication
When dealing with sensitive information, it’s paramount that an agent can verify the customer in question (we’ve all spoken to someone at our bank and needed to confirm our mother’s maiden name or our last transaction :) ). It’s a time-consuming but necessary process and one that Agent Assist is able to resolve quicker, more efficiently and securely. Agent Assist works with backend API integrations, allowing a brand’s verification software to identify a customer. The customer signs in and is handed back to the agent conversation, all within the same channel.
Agent Assist verifying a customer's details
To give brands and agents the confidence to provide a more personalized customer experience, we have applied a layer of AI to our Conversocial Bot platform to identify and remove personally identifiable information (PII) from conversational history after the conversation has been resolved.
This “PII Scrubber” works behind-the-scenes to purge PII from the customer conversations we enable with our Chatbots. All messages are analyzed via a customized NLP engine before a conversation is marked as “resolved”, ensuring personally sensitive data is removed.
2. Collecting Information for Quicker Resolution
Having every possible piece of information available in order to deal with an issue is the holy grail of customer service. Private messaging channels should allow for most queries to be resolved in one session, so gathering all the relevant details is key.
For instance, the majority of queries for a Conversocial partner in the health insurance sector are plan specific, so once an agent is aware of the issue the customer is addressing, handing off to a bot that collects name, date of birth and their policy number. Automation helps to gather all the information that an agent requires and the conversation is handled quicker and more efficiently thanks to Agent Assist.
Building great brand/customer relationships is about not only what the brand understands but what is remembered. Any bot built by the Conversocial team uses data from previous interactions to enrich the customer profile, and ultimately the experience.
3. Completing a Purchase or Transaction
Agent Assist gives agents the ability to help customers make secure automated payments within a private messaging channel. For instance, airline passengers wanting to add extra baggage to their flight booking ordinarily would need to go through a carrier’s website/app or call the contact center.
Agent Assist being deployed to add baggage to an existing flight booking
Agents can pass the conversation over to a purchase bot, that can take information like booking reference and number of bags required, then complete the transaction - either via Apple Pay on ABC or through a payment platform such as Stripe, for conversations taking place in WhatsApp. Payment platforms like these are designed for safe and secure transactions, so there’s peace of mind for customers that agents aren’t handling their card details, but by using Agent Assist the whole process stays in a messaging channel.
The time that agents save by having common scenarios automated means they are able to focus on more complex issues with other customers in Conversocial’s Agent Workspace. Using bots to handle the scenarios described above not only helps agents but actually improves the experience for the customer as well. There are certain situations where automation is preferable and a purpose-built bot is designed to make the process more efficient than a human is able to do. Agent Assist allows humans and bots to work together, handing back and forth, sharing the jobs that they are best equipped to handle.