August 10, 20215 min read
With the rising popularity of conversational chatbots, simple customer queries are now handled automatically. That’s great for customers because they get faster answers, and it’s great for companies because it lets them scale to meet fluctuations in demand.
But it also means that the kinds of queries that service agents handle get more complex. When conversational chatbots are working effectively, agents need to be trained and equipped like never before, because if they fail, the brand reputation is at stake and employee morale drops.
Companies create lots of content and processes to help agents succeed, but these assets are constantly growing and evolving, making it difficult for even the best agents to keep track of it all. And the answer an agent needs is in several different systems — CRM, auto-management, inventory, delivery, ticketing, etc. — so agents need to learn all of these systems to do their job.
It’s not surprising that agent attrition is such a problem in customer service departments. According to IBM, the cost of replacing a dissatisfied agent who leaves can be 33% of the exiting agent’s salary. If you aren’t equipping agents for success, then you’re spending a fortune on hiring and onboarding new agents.
Check out How to Resolve Issues Faster with the Right AI-Powered Routing Strategy.
Why should agents travel through different systems to fetch data when technology can automate this and surface the relevant information in real time? Automation of manual tasks, bringing processes into the agent desktop, and augmenting them with further AI capabilities, can help brands set up their agents for success.
In the current scheme of things, agents get stuck and often ask their colleagues or supervisors for help; but, with good AI on your side, you can serve up your best agent’s answers proactively without asking.
You can train AI to learn which agent responses result in the highest customer satisfaction, and then AI listens to conversations in progress and suggests the best responses based on the context in real time. Not only does this save time for agents and customers — we’ve seen brands reduce handle time by 55% this way — but it also leverages the skills of the most talented agents to train those who are new or require more help.
AI can also show agents cases that are similar to the one they’re handling right now. When a good agent resolves a customer issue with a high happiness score, that case becomes a go-to example for guiding other agents toward the same result. But if you’re handling hundreds or thousands of cases a week, the right one is hard to find by searching through knowledge bases manually. So let AI do that.
Onboarding and training new agents is a constant struggle for many contact centers. Having the right tools to get agents up and running is critical. A great application of AI is delivering step-by-step guidance to agents using Guided Workflows. Because AI is analyzing each conversation in real-time, a guided workflow can be presented to agents that is relevant to the conversation topic. For example, if the customer is asking how to return a damaged product, a set of step-by-step instructions will be presented to the agent that allows them to walk the customer through the return process. Guided workflows can also be used to detect cross-sell and up-sell opportunities and offer the agents the right steps to close the sale.
Another useful application is understanding how happy the customer is at any moment. AI won’t ever understand what happiness feels like, but it can analyze combinations of factors to detect when a human is expressing it. Or its opposite. Giving agents a live prediction of the CSAT score is useful for training them to change tactics based on company policies. When the predicted score drops below a certain point, offer them something for their trouble. When the predicted score rises above a certain point, refer them to your advocacy program.
CSAT prediction can also automate escalation to a supervisor when the score looks like it will be very low, saving the agent from performing the manual steps of (1) deciding when escalation is appropriate, and (2) sending the cry for help.
These tools are available right now, but most agents don’t have them, exposing a flaw in your strategy.
Going forward, I see AI taking a bigger role in helping agents resolve customer issues, such as in the assignment of cases. Getting this right is crucial for the success of the whole interaction, and AI is currently underutilized in making that perfect match.
For example, AI can detect that a query is coming from a premium customer who spent a lot of money on your brand. It’s not enough to route them to the next available agent. Maybe that agent was onboarded yesterday, and that wouldn’t be a great experience for someone who built up a history with your brand. AI should match that premium customer with a premium agent to ensure they’re handled the right way. Assign cases based on priority, history, age, location, revenue, whatever combination of factors make sense for your business and your customers.
This is just one of the AI assistance features we’re working on adding to Modern Care so agents can deliver better customer service experiences. Learn more about what we’ve done so far to help agents with creative use of AI.
Learn more about AI-powered routing and Smart Pairing from Sprinklr Modern Care with our new eBook.