Apr. 13, 2021
Every day, millions of people submit support requests for the products and services they use. As our products and services become more complex, so do the support requests. For most companies, the number of customer support tickets is steadily increasing as the number of channels where customers require support increases as well. Today, a user may reach out to a company over the phone, on e-mail, SMS, or on social media. To make customers feel heard, support teams need to leverage artificial intelligence to supercharge their workflows.
Let’s start by defining artificial intelligence to better understand what it can and cannot do. In computer science, artificial intelligence is the simulation of human intelligence by machines. In popular usage, AI refers to the ability of a computer or machine to mimic the capabilities of the human mind, recognizing objects, learning from examples and experience, understanding and responding to language, making decisions, and solving problems.
Popular media often portrays artificial intelligence as a substitute for human expertise, when in most cases, AI serves as an assistant to skilled knowledge workers, like an intelligent sidekick. In the case of customer service, it’s used as an aid to help frontline teams prioritize tickets, find trends, and to suggest knowledge articles to improve response times.
According to a 2020 Customer Experience Trends Report by Zendesk, “Companies that leverage the most customer data — those in the top 25% of managing data relative to other similar companies — see 36% faster resolutions and a 79% reduction in wait times. They also solve four times the number of customer requests.”
Our team, at AptEdge, is building the toolkit companies need to supercharge their customer support function by leveraging their customer data.
Training machine learning models is often time consuming and expensive. AptEdge has developed an AI solution that mitigates much of the overhead associated with Machine Learning and can deliver value in real time.
For example, suppose a customer support agent views a new case that has just been reported by an important customer. She recognizes that the case is related to another open case she’s working on and links the two for reference using an Edge.
AptEdge immediately learns about the connection and can suggest additional similar cases, one of which can help solve the problem and decrease time-to-resolution.
Conventional AI requires hundreds of examples in order to learn from past events. Often these examples need to be prepared manually. AptEdge AI takes a different approach.
For example, when a support agent links a case to a newly discovered product issue, AptEdge immediately begins learning from this first example with no additional effort required. As more examples trickle in, AptEdge will begin suggesting the new product issue and will continue to learn over time.
AI is notoriously finicky about the language people use. Just ask Siri! At AptEdge, we are focused on Customer Support engagements. We respect the dialogue between customers and agents.
Our AI solution is designed to be sensitive to the language and topics that arise in these conversations. We have curated a unique data set in conjunction with the development of AptEdge to ensure that our AI learns to hone in on the most relevant content. This means AptEdge will perform better on your support tickets than AI solutions that have not been adapted specifically for Customer Support.
Recommendation systems can be handy, but they have a limited context. AptEdge can help you take a recommendation for a given ticket and relate it to other tickets in your CRM. For example, imagine you are using Zendesk Knowledge Capture to beef up your knowledge base and find good articles for customers. On one ticket, it recommends a great article, so you hit send. Next, you open the AptEdge app, which suggests several similar tickets that could all be resolved by the same article. It’s a great way to expand the impact of a tool you’re already using.
Business leaders often view customer support as a cost center, but support teams own one of the most valuable information assets of the company—the record of customer engagements in the CRM. AptEdge helps you extract both strategic and tactical information from your customer support requests, turning your CRM into a gold mine for your company.
AptEdge AI helps make sense out of diverse sources of customer information from structured sales data to unstructured emails from customers.
We can help you answer questions like:
What customer support cases put the most value at risk by sales volume?
Who’s likely to churn based on frustration expressed in support requests?
Are smaller customers expressing frustration with a product issue that could eventually impact your larger customers?
Consider how this kind of information can improve the product development strategy at your organization. Equipped with insights from AptEdge, product teams can respond more quickly to changes in customers' perception of the product, align their priorities with business interests, and make data-driven decisions based by listening to the voice of customers.
At AptEdge, we love AI platforms like Google Cloud AI, Microsoft Azure AI, and Salesforce Einstein. These platforms provide building blocks that your team can use to develop solutions to your own unique challenges if you have the time and expertise.
AptEdge, on the other hand, is a Customer Support platform that can complement AI platforms and deliver additional value out-of-the-box. For example, suppose you would like to know which recent cases are related to a new product issue, say, JIRA-1234. In some situations, we could use Einstein Case Classification for this. But not always. Einstein requires at least 400 closed cases to begin learning.
Do most businesses have 400 closed cases related to the new Jira? Probably not. In addition, Einstein can only learn to classify by field value. Did all the agents on our team consistently assign a field value for the new Jira? If not, Einstein can’t help. AptEdge is designed to solve these sorts of problems, which are common in Customer Support organizations. AptEdge can begin classifying cases related to a product issue given a single example, and its performance only improves with more.