Customers have spoken. They no longer want to interact with a chatbot that only gives them 4 options, none of which answer their problem. The future of customer experience is in human-centered AI, which continuously learns how to interact with your customers. How can you prepare your organization for this new wave of human-centered AI? We’ll be covering this topic and more in this live panel discussion.
Want to learn more about Forethought?
Webinar Summary
In this webinar, one of Forethought’s marketing team members sat down with Head of Sales Engineering, Zia Syed, and Enterprise Customer Success Manager, Lindsay Fifield, to discuss what preparing a customer support organization for AI entails. They discuss what putting a team through a digital transformation is like and how CX leaders can prepare themselves for it.
Customers have spoken and they expect more than what current orgs can offer them when it comes to customer support. As CS has evolved over the years, we’ve come to see that we expect more out of the companies we purchase goods and services from. If we have questions about a product, we want answers quickly and we want them to be as accurate as possible.
Customers have zero desire to interact with a chatbot that only provides a handful of options – often leaving out anything related to what a customer may be asking.
And with people being more online than ever these days, they know how quickly they should be able to find the information they need. Unfortunately, as we all know, it can actually be pretty difficult to get fast, accurate information in a timely manner. In fact, 71% of customers say customer support has remained the same or gotten worse across all channels.
So how do we prepare? What do we do?
Bluntly put, the future of customer service is in human-centered AI, which continuously learns from customers’ inquiries and conversation how to interact with customers and provide accurate information. And that’s where Forethought comes in, we’re an AI platform that takes customers through the entire support process, helping prevent inquiries from becoming tickets in the first place.
What is “AI of the Future”?
To begin this discussion, Zia and Lindsay were asked to tell us more about AI, or better, “When people hear ‘artificial intelligence’ they have no idea what to expect. What is AI after all”
Zia went on to explain that AI is not what we’re seeing in the movies. It’s not a sentient being like Ultron that is going to over the world. He says, “AI is a system that can make autonomous decisions and emulate the behavior of the human mind. AI shouldn’t require human maintenance and it should work silently in the background – learning through its own data experience rather than through human intervention and manual programming.”
AI that works in this manner enables it to make real time decisions across use cases and scenarios since the AI evaluates on its own the best and most efficient way to meet an end objective.
As cool as that sounds, that’s still pretty futuristic. What Zia describes is pretty different from what we see AI being marketed as today. As much as all support organizations would love for an AI customer support tool to continue learning and grow with their support, what they’re getting and seeing now is not that at all.
Zia mentions, “Rather than making autonomous decisions based on historical data or live context, these AI bots, systems, and infrastructures rely heavily on hard-coded triggers that indicate a behavior or action should be followed.” What he means is that what is presented as “AI” today is really just a chatbot that runs off of if/then statements that have been created and maintained by an individual within the support organization.
This means that no one really trusts artificial intelligence as it stands today. Chatbots look for keywords in customer inquiries and will only be able to reply if that keyword is found within the query. So, if someone sends in a question asking, “Can I get a refund?” The only times it will send back accurate information is if the chatbot sees the word “refund.”
This just doesn’t really work for customer support. What if someone asks “Can I get my money back?” They likely would get a message that says “Please rephrase your question.” and then never actually get the required response. Zia ends his thoughts by talking about how this is not real AI. Chatbots that run off decision trees are difficult for scaling and don’t actually help customers as you grow.
Real AI vs Faux AI
So if chatbots based on decision trees aren’t real AI, what is “real AI?”
Zia gives us a great answer by talking about how AI should not rely on any sort of hard-coded mechanisms. Real AI should emulate human behavior, it should be able to take in historical data and learn from that data to be able to meet customer needs. It should also continue to learn from customer data.
By taking into account historical data and trillions of data points across industries to train both general and hyper-specific models, true AI can go way beyond what hard-coded automation frameworks can do.
In the world of support, for example, an AI system can be designed to automatically route concerns to specific, relevant support departments — depending on the AI driven analysis of the customer’s inquiry and an understanding of historical alignment. The AI system should also be able to understand when and if certain inquiries need escalation depending on the customer’s sentiment, as well as other factors — and should be able to accordingly escalate or prioritize the issue, which is actually what Forethought’s AI is capable of doing.
Faux AI is the chatbots we’ve been describing and it’s actually been a detriment to support leaders seeking out new tools. Lindsay talks a little bit about customer perception of AI and it’s not great news.
In regards to AI, “it’s not the best customer experience. It’s hard to manage and it doesn’t deliver the best, most impactful results.” AI has a bad reputation and what’s given is that, “are these chatbots that are built with really complex and hidden decision trees, these if/then flows that result in an output that is usually the wrong output.”
Customers expect more, and with the help of AI support organizations can partner with the true AI tools they need in order to help them meet their customers where they are at.
Using AI to Improve CX
Throughout the webinar we mention a “digital transformation” within support organizations. But what is that?
Lindsay says, “Digital transformation is the integration of digital technology into all areas of a business, fundamentally changing how you operate and deliver value to customers. It’s also a cultural change that requires organizations to continually challenge the status quo, experiment, and get comfortable with failure.”
She goes on to say that a digital transformation is not a one-size fits all. Lindsay talks about how it’s all meant to digitize and modernize customer support. A digital transformation is the doorway to an overall CX transformation which is the end goal of the companies who partner with Forethought to lead their transformations.
Lindsay talks about one particular transformation with a Spoonflower, one of Forethought’s customers. Spoonflower had put a lot of thought into how they’re transforming their support org and partnered with Forethought to have the tools required to empower their agents and their customers in their customer support journeys.
She says, “Since Spoonflower’s partnership, we’ve been able to help them reduce ticket volume by deflecting and resolving over 50% of their customer queries through a true-AI powered chatbot. This has allowed for a more pleasant busy season for their agents and we’re helping satisfy customer needs.”
AI can do so much for support orgs, they just need to partner with the right people in order to transform their CX and their agent experience in the process.
The process for implementing the AI necessary for your teams is also not as scary as people might think.
AI implementation is also not a one-size-fits all. Implementation will look different depending on the desired outcomes of the particular organization looking to implement. What does stay the same is the five phases of implementation that Lindsay goes over in the webinar. These phases include:
- Kick-off Call
- Here we meet with a potential partner and evaluate their goals, align on success metrics, timelines, and expectations
- We establish a connection between customer’s helpdesk and Forethought
- Schedule weekly syncs to track progress
- Solutions Implementation
- In phase two, we train models baked on customer idas and provide a testing environment for Forethought tools
- Coordinate agent training for other tools
- Analytics Dashboard
- Set up analytics dashboard for teams to track progress
- Review progress
- Continuous improvements
- Identify opportunities to further improve AI model performance
- ROI Evaluation
- Present ROI based on current metrics
- Discuss next steps for partnership
This might seem like a lot, but in reality new Forethought customers require little lift to get things going. Zia ends the webinar by emphasizing that new customers don’t need a lot to get things rolling with AI. Teams need their data, their helpdesk admin, and a couple agents to get new content rolling for the AI to get started.
Companies need to empower their agents and empower their customers to help themselves. They can do this with the help of customer support AI.
