How To Get Started With AI-Powered Customer Service Automation
The age of AI-powered customer service automation is upon us. Research analysts Gartner predict that by 2020, a mere 15% of customer interactions will be fully handled by humans, which points to a sea change in how we support the customers who reach out to us with questions about our services and products.
Yet lack of understanding about AI-powered automation often freezes customer service leaders in their tracks, leading to inaction. Will it work for my specific needs? Will it displace my team? How do I actually position AI internally, both to my company leadership and to the service agents who’ll have to work with it?
Here are some of the questions we’ve worked with our 50+ customers on addressing over the past several years, and some of the key things we’ve learned along the way.
How Do I Know if AI is Right for My Support Team?
Many customer service teams struggle with a high volume of repetitive and time-consuming tickets. These are often questions about refunds or cancellations; order status inquiries; shipping lookups and so on. These often require the use of more than one system in order to resolve them, focusing agents to repeatedly “swivel chair” from system to system in
order to answer the same questions, over and over again.
If your support team has a high volume of repetitive, text-based tickets – tickets that require more than one system to resolve them – your team is an excellent fit for AI-powered customer service automation.
Moreover, if you’re looking to make customer service a competitive strength, and/or are looking to reduce costs, AIpowered automation is an excellent means of getting you there. It not only helps achieve these objectives, it also empowers your existing service agents to focus their time and activity toward more complex queries – leaving the repetitive, time-consuming work to now be fully automated and resolved by AI.
What’s Real in AI, and What’s Not?
Make no mistake: practical, real-world, non-Hollywood AI really exists, and it’s rapidly becoming a competitive advantage in support organizations and contact centers worldwide.
That said, it’s best to remain skeptical of results and numbers that look too good to be true. While one day AI may be truly autonomous, today’s practical AI solutions seek to enhance human performance, not replace it. Take a “narrow” AI approach that’s truly focused on solving your particular customer service problems, and choose vendors with established AI research teams, a strong engineering culture, investor backing and a strong market share.

How Do I Position an AI Project Internally?
New technologies, particularly AI and machine learning, often create buzz and excitement, yet if the terms themselves end up being the main draw for a
business, they’ll likely to lead to disappointment. AI should be approached with the sobriety and methodology of an ERP purchase.
It’s important for the entire organization to know just what problem it is you’re trying to solve. Keep in mind that conversations are just 30% of the average customer service experience. The other 70% is the series of actions and processes that customer support agents have to perform manually each time a customer asks them for a refund, or the status of an order. This has a huge impact on costs, customer experience and on long-term agent morale.
Once you’ve established and communicated what you’re trying to solve for, the key to effective decision-making is determining the ROI of the project. Every buyer or signatory on the project will ask you, “How much are we saving?” or “How much are we making?”.
Push your AI provider to provide a reasonable projection of ROI, along with their analysis of how they’ll help you accomplish it. This will illustrate to management the type of project (for instance, productivity improvement) and the level of returns (for instance, 10% automation that reduces costs by 20% and increases CSAT by 5%). An analytical ROI approach demonstrates to management that the project is not experimental, but rather a critical requirement to achieve specific targets.

Should I Buy Now, or in 12 Months?
There are always pros and cons for innovating early, yet if you’ve established clear ROI and a fit with your objectives, there are many excellent reasons to begin right now, and not wait a year.
First-mover advantage provides your company with the experience of practical, real-world experience with AI before your competitors and peers gain this experience, while also allowing you to meet your customers’ rapidly increasing expectations for what a great customer experience looks like. It shows them that you’re investing in technology to make their experience better, and it helps you reduce internal system costs and slow headcount growth that might otherwise consume a portion of this year’s budgets.
What Else Should I Be Considering?
At the end of the day, customers don’t reach out to your customer support team because they have no one else to talk to. They reach out because they have specific issues such as refund requests, subscription upgrades, or a problem with the product. In this problem-focused domain, the two most important factors are speed and quality.
Being able to provide a fast and accurate resolution, whether automated or not, goes far further than providing a “real human rep” level of comfort that can require that the customer wait for one, two or three days for resolution of their problem.
