Our next and final fail is avoided by the Giant Otter feature that no one else offers: true, lasting memory — the kind that lets a chatbot get back to work with a user after an extended break or a change in scenery.
Our last post looked at chatbots that pay off their sci-fi stereotype by sounding robotic. A worse problem than a tin ear, though, is clunky mechanics. That’s fail territory.
Our last two posts profiled fails that top-down chatbots too often commit. Both involved saying the wrong things, like being a bit too cheerful and ignoring food allergies.
Comparatively, our third fail might not even look like a fail. But it’s the opposite experience you want for your customers, and as close to “bleep bloop” as a bot gets.
Here are a few examples of chatbots that sound robotic:
If you have a chatbot, odds are that your reasons are specific and your expectations are clear. After all, chatbots excel with focused objectives that address real customer needs.
But even chatbots with one job to do have moments when they trip up — often due to top-down design that misses obvious intents (what a bot needs to understand) and flows (where the subsequent conversation needs to go).
Our last article touched on five frequent fails committed by chatbots and caused by top-down design. (You shouldn’t throw the bots under the bus.)
Exhibit #1? Bots that seem like they don’t get it. When a chatbot doesn’t understand basic inputs from users, you’ve got a fail on your hands.
At Google’s I/O conference on May 8, they introduced a new extension of their AI Assistant that generated a little bit of buzz.
When it comes to going viral, chatbots can’t compete with cats and newscasters. After all, no one’s ever filmed a chatbot falling into a swimming pool.
Still, chatbots have had their fair share of embarrassments. Search for “chatbot fails,” and you’ll find lots of articles serving up blunders and then claiming that chatbots are trending downward. But are chatbots failing, or is it their designers?
In our last post, we looked at the essentials of automated live chat, as well as the all-too-common “top-down” chatbot development method most platforms perpetuate.
We at Giant Otter hate to see good bots go bad because of this flawed approach — one that starts them from scratch and leads them to the scrap heap.
AI-based chatbots can stand toe-to-toe with human help desk technicians and customer service agents. Their construction just has to start from the bottom up.
You’ve met Siri, chatted with Alexa, and said okay to Google. And you’re sure: Chatbot technology is right for your business. (It’s a conclusion a lot of companies are coming to.)
R2 got Luke out of a jam or two, and you’re expecting your bot to do the same for you. But you should know that, out of the box, your bot can’t even say hello, much less help you blow up the Death Star.