Salesforce has announced a swathe of updates to its Service Cloud, which it believes will help companies connect their customer service operations - from DevOps teams, through the contact center, all the way to the customer engagement layer.
Customer service teams have been under increased pressure during the COVID-19 pandemic and companies have had to adapt, particularly as they deal with more demand and have had to shift contact centers towards distributed environments. Broken workflows and departmental silos have never been more apparent in organizations that are trying to proactively deal with customer requests.
These demands coupled with the trend towards customers wanting to be serviced in a channel that suits their needs at any given moment in time, means that companies need to rethink their approach to customer service entirely. Agents reading from scripts and dealing with individual requests in isolation isn't going to cut it for the digitally connected buyer.
The Salesforce Service Cloud updates this week seek to reduce the burden on agents dealing with repeat requests, through automation, while pulling in DevOps teams to deal with emergency situations, and using AI-powered workflows to ensure that work is processed quickly and easily.
Clara Shih, CEO of Service Cloud, Salesforce, says:
With the breadth of Salesforce technologies and the new innovations that we're pushing out, plus the addition of Slack, we're really able to transform service end-to-end. With the customer experience layer, we're making sure that companies are in all the digital channels that their customers want to be in, whether it's messaging or chat or self service.
And then we're going all the way through to the automation of business processes. In the past, an agent might have to look up an order and then let the customer know, but now we can connect all these systems with MuleSoft, and the customer can look it up themselves. She can just chat with a bot, and the bot can tell her that her order has been delayed by six days and here's the tracking number. She doesn't even have to call in.
Salesforce's recent acquisition of Slack is also helping the vendor bring teams together to break down departmental silos and get them working on problems of importance for customers in real-time. Shih adds:
And then for harder issues, like if there's a storm, for example, and lots of people are affected, it's not something that is easily done through self-service. We've found that Slack is the digital HQ for collaboration. And for these bigger issues, your frontline customer service agent, she can't solve it on her own, but she also can't, raise her baton and talk to her supervisor like she did when everybody was in person.
So now all of this is happening on Slack and instead of taking days to email another department to figure out how to compensate the customer whose order has been lost in the storm, this can happen within minutes because Slack is real time and it just connects everybody with this rapid messaging.
Throughout the course of the COVID-19 pandemic a deeper understanding of the significance of workflows has emerged. With distributed teams, organizations have noticed how work is (or isn't) flowing easily through their companies and are looking for ways to fix these broken processes.
Central to the Service Cloud updates this week is using AI-powered workflows to get agents working in contact centers the information they need at the point that it matters to the customer. Equally, serving customers the data they need when and where they need it. Salesforce believes that these workflow improvements, in conjunction with Slack and conversation mining, will change the way that contact centers operate.
The announcements this week include:
Customer Service Incident Management - to accelerate resolution for major incidents, by helping companies detect, diagnose, and respond to service disruptions. This is aimed at helping service teams to proactively notify customers of a problem and triage increases in cases, as well as provide transparency for customers and the operations teams working to resolve the root problem. Swarming automatically brings together the right internal and external experts in Slack to collaborate and solve major incidents and escalations.
Omni-Channel Flow - built on Salesforce's workflow platform omni-channel flow allows service teams to create complex rules based on CRM data for routing cases, calls, messages, and chats across the service team and other departments. With Omni-Channel, Einstein Case Classification, and Einstein Article Recommendation in Flow, Service Cloud can now analyze incoming cases and automate routing to the best queue, agent, or process - including processes that interact with external systems - and auto-respond to customers with relevant articles to drive efficiency and reduce support costs.
Robotic process automation capabilities (RPA) for Service Cloud - off the back of the recent Servicetrace acquisition, service teams can now automate repetitive tasks such as look-ups and write-ins across legacy systems that lack APIs.
Einstein Conversation Mining - using Natural Language Processing (NLP) to identify the most common types of interactions with customers. For example, Einstein Conversation Mining can determine which use cases to prioritize for a service bot rollout or prompt the creation of a new knowledge article to address a frequent customer concern.
Messaging for In-app & Web - enables customers to start a persistent messaging experience, like SMS and WhatsApp, directly in a mobile application or on a website and pick the conversation up where it left off.
Visual Remote Assistant - allows for two-way video and audio for face-to-face conversations between agents or field technicians and customers
Workforce Engagement Intraday Management - allows companies to improve customer experience and agent morale by adjusting employee schedules when things don't go as planned, like if there is a severe weather incident and call volumes spike.
Service Cloud Voice - brings together phone, digital channels, and CRM data in one central workspace for service agents
Shih provides a clear example of how some of these tie together to change the operations of service throughout an organization, which directly improves the experience for customers (some of which may not even be aware that they need support).
Take a media streaming company, she says, that has an outage in a certain area. This is a proble, as every minute that they can't stream their content is a negative experience. Shih explains that by having a connected platform, automated workflows and using Slack for collaboration, the streaming company can not only solve the problem in real-time but also service customers requiring an explanation. She says:
We can see a flood of calls coming in. Everyone's complaining about not being able to access their favourite show, we know that these are related cases to each other.
And so rather than treating them as one offs, we cluster them as an incident. Your customer service agent isn't going to get the instance back up, they're not a developer or DevOps engineer - but with a button click, they can start a swarm with that DevOps and security team. They're no longer in the same office, they could be several states away, and now they're collaborating.
That DevOps team starts swarming on it, they start trying to debug what the issue is, and while they're working, they're providing real time updates,
And maybe you've had 500 customers call in, but actually the number of affected customers is bigger than that because not everybody calls in. Maybe it affects 3,000 customers. And now you can actually proactively message or email those 3,000 customers before they even have to pick up the phone just to let them know there's an issue and it is being worked on. You could then offer them $50 credit towards their subscription next month. And then once it's been resolved be able to close the loop.
The role of conversation mining
The use of conversation mining and applying intelligence and AI to service delivery is particularly interesting. As noted in the example above, this can be used to identify problems when calls flood the contact center in real-time, which can directly feed into incident resolution processes.
However, Shih highlights a couple of other interesting use cases too. For instance, using conversation mining to highlight where companies should be providing self-service opportunities to their customers to reduce the burden on contact center agents. She explains:
One of the most valuable applications of our conversation mining AI is to understand: what are the top reasons that customers are calling in? These contact centres that get tens of millions of calls every year, it's really helpful to know why people are calling. It changes by day, by week, by month, so you have to continuously run machine learning and AI on this.
What we're finding is that there's a number of reasons, when looking at top call intents, that customers really shouldn't be calling in for. A reasonably large insurance company that we are working with that we're applying this to - it turns out customers were calling to change their address and reset the password on their account, neither of which are good reasons for customers who want to call in. Agents don't want to deal with that. It's just manual data entry.
Those are self service processes that any company within five minutes, with just a few clicks, using the Salesforce platform part of Service Cloud, are able to expose as ‘do it in the mobile app, do it on the website, share it with a bot, save yourself a call'. And that act alone was able to deflect 20% of the calls that were coming in.
What does that mean for the rest of the customers? It means customers calling in for legitimate reasons, like their house just burned out, don't have to wait on hold because there's less of those automatable intents clogging up the system. And for the agents, they get to focus on the meaningful interactions that really require their higher order thinking, problem solving and empathy skills.
Another interesting example Shih mentioned is to use Service Cloud Voice to analyze where agents need additional training. She says:
We have to help agents continuously upskill and that's what we're doing with our Workforce Engagement Management solution. We're providing push personalized training to each person. And this works because we have Service Cloud Voice. We're not sampling 1 in 1,000 calls, we're listening to all of them.
We then know that one agent needs training in this specific area and another agent needs training in this other specific area. And instead of forcing both of them to go through 20 hours of training, we're just pushing each agent the 20 minutes in the area that they need.
There are so many moving parts to effective customer service that tying them all together can seem like an insurmountable challenge for organizations. But there are some key lessons here. Identify common complaints and automate/implement self service where possible. Break down silos by connecting data and giving service agents the information they need. Bring teams together in real time with collaboration tools to solve problems quickly. Be agile enough to change workflows when the circumstances change (e.g. COVID). Ensure that when service agents are dealing with customers, those interactions are meaningful. Focusing on these outcomes should help to ease the pain, both for companies and customers.