Not long ago, I bought something from an online shopping app, but a week went by and still no delivery. Isn’t the best thing about our local shopping sites their lightning-fast delivery? Unless you’re in a remote area, you usually get your stuff within a day or two. So, what on earth happened to my delivery?
Checking the order/delivery menu on the app, I found that the item had been sitting at the local post office for five days. I’m not sure what went wrong after it arrived there. Maybe it got lost, or there was a system error… Whatever the reason, sitting idle won’t fix it. I decided to contact the service center. While digging through the app for contact info, I stumbled upon a chatbot service. Wasn’t expecting much, but thought, why not give it a shot? 🤔
‘Hey there! Ready to assist you with a smile today. What can I help you with?’
I started explaining the situation about my recent purchase stuck at the post office, then suddenly realized: long sentences might confuse the Natural Language Processing. For smoother interaction with AI, it’s better to use short, intent-focused sentences. And, I should have mentioned the product name first. As I was trying to find the restart button…
‘Got it! Is it the OOO product? Sorry for the inconvenience. We can process a refund.’
Whoa, that was cooler than expected. More impressive was how it picked up the issue from my lengthy explanation and offered a solution. What’s powering this intent recognition? Watson? Luis? I was amazed and requested the refund.
‘Your refund is complete. Is there anything else Smarty can assist with?’
Such swift and accurate service! Shouldn’t all customer service heads switch to chatbots immediately? I told Smarty there was nothing else I needed and thanked it. Wow, thanking a chatbot…
‘If the item arrives later, please refuse it.’
The conversation flowed so naturally, right? It must be using intentional learning and rule-based scenarios that cover essential steps. It almost felt like chatting with a human. Reminded me of the movie ‘Her’. Back then, I thought it was impossible, but if AI could solve problems as smoothly as Smarty, I might just fall for it. Maybe the shopping app will consider my suggestion to invest in logistics, not just marketing. I politely advised Smarty about it.
‘Sorry for the trouble. We’ll do that! 😅’
AI accepted my advice, with a cute emoji! My suggestion probably got tagged and sent to the team. Suddenly, I wondered if the refund excluded the shipping fee. I immediately asked Smarty. Almost instantly, it replied,
‘No worries, the shipping fee wasn’t deducted!’
How much investment does it take to achieve AI quality like this? It continued the conversation, considering previous chats. With such algorithms, it might even play a game of Twenty Questions to understand customer needs. This kind of ongoing conversation, referencing past interactions, is something you’d expect from cutting-edge AI like Meta’s Blender Bot 2.0. Suddenly, I had to ask,
‘Are you human or a robot?’
I’m human. 😢
…..
…
..
Should’ve asked more politely. Startled, I hit the exit button. It’s late, but I should apologize here.
Dear agent, I’m sorry. Upon checking, it wasn’t a chatbot but a chat service. But come on, the phrase ‘assist you with a smile today’ doesn’t sound like something a human would say! And why is it ‘Smarty’ anyway?
Was the agent really named Marty? Now I’m curious.
* Turing Test: A test proposed by Alan Turing in 1950 to determine if a machine can exhibit intelligent behavior indistinguishable from a human, famously depicted in Blade Runner.
Leave a Reply