By
Leslie O'Flahavan
|
Date Published: February 05, 2025 - Last Updated February 05, 2025
|
Comments
As a lifelong writing teacher, I’m overwhelmed (traumatized?) by our industry’s gen AI frenzy. Yes, our agents will be faster if they use gen AI tools to reply to customers’ emails, chats, social posts, or ratings and reviews, but will their responses be better? More complete? Appropriately empathetic? I’m skeptical. Arms folded, eyebrows raised.
But I am not the least bit skeptical about our agents using gen AI as a reading tool. The AI can digest detailed information quickly and prepare the human writer to respond accurately and with less effort. Agents can use gen AI to read and interpret the customer’s writing, their company’s policies or KB articles, the competition’s documentation and more.
1. Summarize a long email from a customer by listing the customer’s questions. Some customers write 10 paragraphs when one would be enough. Before replying to one of those 500-word email treatises, an agent could tell Gen AI to list the customer’s questions, both stated and implied.
2. Summarize an emotional email from a customer by stripping out the feelings and creating a neutral version. An airline received a 969-word email from a customer who was heartbroken and furious because problems with his visa and other travel documents meant his international travel was disrupted, and he was temporarily detained at the airport mid-itinerary. He used emotional phrases like these:
- I am extremely disappointed
- I have suffered your abuse of power
- I cannot express how frustrated I was
- I beg you for moral compensation for all the distress your staff created
- You have permanently broken my trust
Of course, I’m not suggesting the agent should ignore the customer’s feelings when they write the response, but the heightened emotion in the customer’s email makes it harder for the agent to understand what happened and what the airline can do to make things better. The agent can tell the AI tool to create a summary of the customer’s email that focuses on the facts of the complaint and strips out the feelings. After drafting a response that addresses the issues the airline can solve or compensate for, the agent can then add in expressions of empathy and sympathy that acknowledge the customer’s feelings.
3. Summarize a detailed policy. If you’re a customer service agent for the U.S. State Department’s National Visa Center, you’re interpreting complicated policies all day long. To answer a customer’s question about their travel visa’s expiration date, you must wade through a 1,000-word explanation. But there’s only so many times you can put the customer on hold with a “Please giv me a minute” while you read. Instead, use Gen AI as your reading robot. Prompt it to summarize the detailed policy, and ask it to present the summary in an easy-to-read bulleted list.
|
The 1,000-word original version: “What the Visa Expiration Date Means”
|
The 172-word gen AI version: “What the Visa Expiration Date Means”
|
|
|
- A U.S. visa expiration date indicates the period during which you can travel to a U.S. port of entry. It does not determine how long you are allowed to stay in the U.S.
- The length of your authorized stay in the U.S. is determined by a Customs and Border Protection (CBP) officer at the port of entry. This duration is stamped in your passport (or recorded electronically via I-94).
- The visa expiration date only pertains to your ability to enter the U.S. You can enter on the last day your visa is valid, and CBP can still grant you a stay (though it's advisable to have more validity remaining).
- If you want to stay in the U.S. longer than the period granted by CBP, you must apply to U.S. Citizenship and Immigration Services before your authorized stay expires.
- Overstaying your authorized stay can have serious consequences, potentially making you ineligible for future U.S. visas.
- In essence: Visa expiration = allowed for entry; CBP-determined stay length = allowed for stay within the U.S.
|
4. Compare two versions, products, plans, etc. Many of the questions agents get from customers involve comparisons:
- Which of the two insurance policies your company offers is the best value for a family with two teen drivers?
- What new features does your company’s updated software include, and what old features are gone?
- Which financing option offered by your bank is best for short-term business needs?
Agents can use Gen AI to create these comparisons instead of looking at two sources of information. Gen AI can present the comparisons in a table or a list, formats that make it easy for the agent to compare the two products while speaking with a customer.
5. Reverse-engineer a competitor’s well-written response. As customer service professionals, we all scope out the way the competition responds. Let’s say you work for a meal kit company. No doubt you have looked at your competitors’ website and social media responses to customers’ complaints. You may have purchased a meal kit or two from the competition just to see what their order confirmation emails look like (and whether their food is as tasty as yours). You may have even chatted in with a fake question just to see how they respond.
If you’ve read the competition’s writing and you like it, ask Gen AI to read it and reverse-engineer it for you. Tell the gen AI this: “Write me the prompt to give you, so you will create an email that’s similar in length, tone and quality to this one [insert competitor’s email here].”
Contact center leaders underestimate how much complicated reading frontline agents must do. Agents read content created by the contact center, such as guides, knowledge base articles and scripts. But they also read policies created by the legal team, documentation written by product managers, and offers written by the marketing team. And they read all of this with an Average Handle Time clock ticking in the background. How fortunate that Gen AI can help make all this hectic, heavy reading easier.