Ahh, AI. The promised future coming to fruition within the little screens we interact with on a daily basis. Though we assumed artificial intelligence would come in the form of apron wearing robots like the ones in the Jetsons, AI comes to us in more practical forms with the same promise: to make the daily lives of humans better. And what better way to incorporate artificial intelligence into your daily life, and content creation process, than an AI virtual assistant?
Using an AI virtual assistant includes many benefits.
-
- Reduces human error
- Infinite availability for 24/7 assistance with no downtime
- Saves time and human effort through process automation and smart decision making
- Increases productivity by taking over repetitive and mundane tasks
When it comes to business, utilizing artificial intelligence seems like a no brainer to help scale. But what happens when those helpful AI virtual assistant products don’t function according to your expectations?
When AI goes bad
I recently had an interesting experience using an AI scheduling assistant to schedule a meeting with a potential new business partner. I was intrigued by the prospect of an AI assistant attempting to set up a call for us, handling the details like timezones, schedule conflicts and meeting locations automatically.
It seemed like a no brainer. All I had to do was give the assistant access to my calendar, set it and forget it. However, when I arrived at the website, I was disheartened to find a distinctly low-quality customer experience. For example:
- Multiple spelling and punctuation errors in the copy
- Confusing naming conventions to differentiate products and services
- Misused or improperly phrased words
Having just gone through the calendering process, I know that the AI works well for what it does. However, I can’t help but lose trust in the product when the content that supports the product does not match the product quality. It made me wonder how far AI really has to go in order for these groundbreaking products to be usable.
The journey to AI depends in part on standardized, structured content. That’s how we train the AI to recognize patterns and make better choices for what content to put together for each person.
I’ve seen NLG (natural language generation) software that does a pretty nice job of creating new content components out of existing content. I’ve also seen NLG fail miserably, producing many words that make no sense to a human being. And if the content is created for humans, then it needs to be written in a way that humans can fully understand.
So how do you avoid an AI-generated alphabet soup?
When you create content according to the principles laid out in The Personalization Paradox – standardize in all five dimensions, tag with metadata, enable content reuse – then you can achieve success with NLG or other AI-created and AI-delivered content. The results can be a useful part of a personalized customer experience, building trust in your product and your company.
The alternative – an obviously automatically generated mishmash of words – undermines everything an AI company is trying to do. Especially when its product depends on engaging with humans.
So where does that leave us and our fine AI friends? Well, in order for AI to work for us the way we want it to, we first need to serve the machine structured content. By “feeding” the system consistent, standardized content, the machine will be able to learn, and in return simulate, high quality content that mirrors human speech.
I’m not suggesting a robot uprising. But perhaps as content creators, there is an outcome even worse. Badly written and unusable content.
- Content Strategy and Taxonomy: Don’t Make Your Audience Play 20 Questions - August 6, 2024
- When in Doubt, Componentize - February 9, 2024
- When AI Goes Bad - September 27, 2021
