Where does your enterprise stand on the AI adoption curve? Take our AI survey to find out.
We’re only as good as the moment, that fragile moment when we please or hopefully don’t disappoint the customer.
— Howard Schultz, CEO, Starbucks
That quote comes from an interview with Starbucks CEO Howard Schultz back in 2011. Retail has always been an industry where service quality informs success, where “the customer is always right.” But in the seven years since that interview, consumers have raised the stakes higher and have created considerably more complex challenges for retailers.
According to eMarketer, 70 percent of people now expect personalized experiences when they engage with a retailer. A recent Salesforce Research report suggests that 64 percent of consumers — and 80 percent of business users — expect companies to respond to and interact with them in real time. And with 75 percent of shoppers using their mobile devices in-store, it’s clear consumers expect retailers to be omnichannel.
Building out a retail operation that can deliver personalized, immediate, and omnichannel experiences is an enormous challenge. And yet, in a world where retailers win or lose, the battle for the shopper on the level of customer experience they provide is table stakes for retailers.
That’s why so many retailers choose to invest heavily in technology — 60 percent increased their tech spend in 2016, according to Gartner. A strategic deployment of technology like chatbots can meaningfully help deliver more personalization and more immediacy. It’s what the customer expects, too; IBM recently found that 65 percent of millennials prefer going online to get support and that they didn’t want to interact with a live agent at all.
Bot vs. human
The holiday season recently ended, and we’ve all just made our New Year’s Resolutions. For many people, that means lofty goals of going to the gym more, running a mile a day, or finally taking up yoga. According to Gold’s Gym, its traffic jumps 40 percent between December and January. And that means bulk orders of fitness gear.
Let’s imagine for a second you’ve recently made a big purchase of new workout gear — but you got the wrong size for one of the t-shirts.
Typically, to resolve your issue you would have to call up customer service and wait on hold for a few minutes (for some reason, the sports store is facing unprecedentedly high call volume right now). Eventually, you’re connected to an agent, and you begin the tortuous process of spelling out your last name and repeating a 12-digit order number three times.
From there, the agent on the other end of the phone takes a deep breath and begins to recite the full list of six new workout shirts before asking which one you want to change. The call takes 15 minutes.
With a chatbot, things can be more efficient. First off, there is no wait time. You click the “chat” button on the email or site, and you’re instantly connected to a bot.
What’s more, with a bot connected to your CRM data, the conversation starts with a contextually aware question, like “I notice you have just completed an order of six workout shirts. Is that what you’d like to talk about?”
That contextual awareness makes for a far quicker and more pleasant experience for the customer. A quick “yes” and the bot shows all six shirts on screen, and you easily pick out the wrong-sized shirt and make a change.
On a basic level, bots are good at those routine, repetitive tasks like processing returns and exchanges or changing shipping addresses. Our research has found that bots are able to handle 20-30 percent of incoming customer queries on their own.
Bots + humans
For those other 70 percent to 80 percent of queries, bots should work in partnership with human agents. In these scenarios, bots triage more complex issues and rapidly allocate them to the best possible agent to solve the customer’s problem.
The bots collect specific information from the customer in a text-based interaction before handoff to an agent. That cuts down on misheard order numbers, but more importantly, it means agents don’t have to waste time asking repetitive questions of customers. That cuts down on handling time, and it gets rid of some pretty boring tasks from an agent’s point of view, giving them more time to spend resolving the more challenging customer service cases.
I’ve seen this hybrid model rolled out across several companies now, and in every case, it led to a considerable boost in customer satisfaction (CSAT) ratings. Humans and bots working together provides a better service experience for customers than either humans or bots working in isolation.
At a time in retail when service is the key battleground, that means the measure of success will increasingly be the quality of the partnership between bots and human agents.
How to add bots to retail
So how do retailers decide where bots fit into the retail experience — and where they don’t?
Bots are a good choice to handle:
- Repetitive, simple tasks: Bots are better on transactions and simple Q&A interactions — “When does my order arrive?,” “How much is shipping?,” and the like.
- Multi-system interactions: They’re also good for interactions where a human agent would have to access multiple systems. Let’s say a customer wants to change an order and change their address. Typically, those two transactions take place in different systems for a retailer. That means agents need to log in to system one, ask for a customer’s details, and change an order. Then they must log in to system two, ask for the same customer’s details a second time, and amend the shipping address. This process is frustrating for customers and agents alike. A bot, connected to all the relevant systems, can complete both tasks at the same time — from the customer’s point of view, in one simple interaction.
- Multi-item orders: Bots are also useful in situations when text is more efficient than voice. It’s considerably easier to navigate lists via text rather than voice, for instance. It’s a somewhat onerous task for an agent to read out a list of credit card numbers for you to pick the right one to charge. It’s far quicker, far easier, and with far less potential for error if those cards were displayed in a list for the customer to click on.
However, bots are not at all ready to resolve all incoming customer queries, and there is trouble ahead for companies that expect bots to do too much.
Bots are a poor choice to handle:
- High touch, complex requests where the answer is not obvious or when a question can lead to multiple answers — for example, if a customer calls to say that a refund they requested was processed, but that the amount refunded was wrong. That’s a tough issue for a bot to resolve because there could be multiple reasons why the amount is different (shipping costs, taxes, return fees), and customers are thus expecting more detail than a bot is able to provide.
- Requests where emotion is involved: Let’s say you have a medical insurance claim dispute. That’s somewhere that you want sensitivity and a human touch, not a bot.
In such situations, bots should triage requests, but rely on human partners to solve them. Those bots can run through the basic questions every agent must ask — things like order number, name, and address — then collect some basic information on the nature of the problem before identifying the best human agent to resolve it.
For better customer service, just add bots
From the company’s point of view, adding bots to the customer service experience expands capacity, cuts wait times, and increases customer satisfaction — helping businesses measure up to the elevated expectations of today’s retail customer.
From the agent’s point of view, repetitive and simplistic tasks are off the table, which gives them time to focus on the more challenging issues that a human agent is far better suited to solving.
And from a customer’s point of view, incorporating bots into the service experience means instant responses to simple queries and more dedication from human agents when there are more complex issues to resolve. Ultimately — and most importantly — that means a higher level of service.
Clement Tussiot is senior director of product management for Salesforce Service Cloud, a customer service software that delivers in the cloud.
VentureBeatVentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative technology and transact. Our site delivers essential information on data technologies and strategies to guide you as you lead your organizations. We invite you to become a member of our community, to access:
- up-to-date information on the subjects of interest to you
- our newsletters
- gated thought-leader content and discounted access to our prized events, such as Transform 2021: Learn More
- networking features, and more