Can generative AI do support?
Many support professionals find it hard to find the right tone when communicating with customers—so why not let generative AI do the work for us? I tried with two common situations and the results were… interesting! Judge for yourself.
Situation 1: I’m not ignoring you
Initial Message: I’m sorry I have not made any progress on your case because I’m just swamped. I’ll get back to you as soon as I can.
A wimpy excuse and there’s no timeframe for the next step.
ChatGPT draft 1: I apologize for the delay in making progress on your case. My current workload has been quite heavy. Please rest assured that I’m committed to addressing your concerns, and I will reach out to you with updates as soon as I am able to. Thank you for your understanding.
The wimpy excuse has not changed. There’s still no next step. And the excuses sound pretty robotic, right? Let’s ask ChatGPT to make a commitment.
Chat GPT draft 2: I sincerely apologize for the delay in your case’s progress. I want to assure you that I am fully dedicated to resolving this matter. Despite my current workload, I will prioritize your case and provide you with a substantial update by [specify a date or timeframe]. Your patience is greatly appreciated, and I’m thankful for your understanding as I work through this.
There’s a next step (to be filled in by the user)–but it sounds like it’s a grave inconvenience to the speaker. And see how the addition of the adverb “sincerly” to the apology make sit weaker, not stronger.
Situation 2: We won’t fix your bug
Initial message: I found bug 123 in our code that causes your issue. There are no plans to fix it at this time since it impacts few customers. Have a nice day.
Pretty curt. And the sendoff is too cheerful for the message.
ChatGPT draft 1: I discovered that bug 123 in our code is the root cause of the issue you’re facing. I truly understand how frustrating this can be. While I know this is causing you inconvenience, I want to be transparent and share that currently, there are no immediate plans to address it since its impact is affecting only a small number of customers. Please know that your feedback matters to us, and we’re continuously working to enhance our services. If you have any further questions or concerns, please don’t hesitate to reach out. Wishing you a pleasant day ahead!
A good start where the bot takes ownership but the language is verbose and non-authentic. And that bad ending is still there. Let’s ask for something shorter.
ChatGPT draft 2: I’ve identified bug 123 as the reason behind the issue you’re encountering. I understand your frustration and while we currently don’t have plans to fix it due to its limited impact, please know we value your feedback. Feel free to reach out if you have more questions. Have a great day!
It’s less verbose and slightly less robotic-sounding but it’s patronizing (what frustration is it talking about?) and the terrible ending is still there. Let’s specifically ask for its removal.
Chat GPT draft 3: I’ve pinpointed bug 123 as the cause of your issue. While we’re not currently planning to fix it due to its minimal impact, we appreciate your input. If you have further questions, feel free to ask.
Finally, no more cheerful ending. But a little short on empathy, and where’s the workaround?
So what do I do?
If you are looking for inspiration on how to word a customer note, ChatGPT may be helpful–but it piles on the platitudes and the fake affect of standard customer service scripts, as it was trained to do, no doubt. Use it for inspiration, but be the boss: review the suggestions critically before you press Send.
If you are a support manager, don’t just ban generative AI (it won’t work, anyway). Instead, share examples and use them as coaching opportunities. Yes, you can use this post!