As we often say, the current stage of AI’s lifecycle is all about excitement and experimentation. The former may be in overdrive, with a correction still to come, while the latter involves a feeling-out process. Based on AI’s buzz and gravitas, execs throughout the land are trying to figure out where it fits.
There are many answers to that last question, some of which are clear today and some of which require more time in the oven. As it often goes with emerging tech, it takes a few years for “native thinking” to develop. This is when applications and deployments aren’t framed within, or guided by, legacy thinking.
As a historical example, it took about three years for iPhone apps to tap into the native strengths of the device. We’re talking GPS, accelerometer, capacitive touch, etc.. These were eventually unlocked in the circa-2010 timeframe when apps evolved from small websites to the likes of Foursquare and Snapchat.
Back to AI, that same cycle will play out. Meanwhile, some success factors are already evident. For example, enterprise AI deployments – including for SMBs – resonate most when they offer tangible benefits like time savings or fewer headaches by automating dreaded tasks (e;g, marketing copy).
“Most SMBs run very lean businesses and have limited resources,” Scorpion CRO Jamie Adams told Localogy Insider. “Our focus is deploying AI solutions to help them optimize their sales & marketing funnel, from converting website users to leads via AI Chat, answering phone calls and booking appointments, and analyzing and scoring leads to be leveraged as high quality first-party data that can be used to optimize Google, Bing, and Meta campaigns.”
Ensuing Mishaps
All the above overhangs the current environment and colors our daily analysis. And this thinking was prompted when the latest example of AI experimentation crossed our desks. Of course, these studies can include best practices and worst practices – both valuable to inch closer to AI enlightenment.
In this case, it was the latter. Specifically, Taco Bell experienced some growing pains and oscillations in its AI deployment. Part of those challenges came from the where, in addition to the what and how. In Taco Bell’s case, its AI deployment and ensuing mishaps resided in its drive-through ordering.
This involved a chatbot-style AI agent to take orders at drive-thrus of 500 trial Taco Bell locations. The results were predictably uneven, embarrassing, and even comical at times. In one instance, a drive-thru customer ordered 18,000 cups of water, just to bypass the system and get a human to intervene.
This sparked an “active conversation” about where to use and not use AI, Taco Bell Chief Digital Officer Dane Matthews told The Wall Street Journal. That brings us back to all the experimental-stage dynamics outlined above, and begs the question of why that “active conversation” didn’t happen pre-deployment.
Moving Target
So what can we learn from this? Though AI’s applicability will be a moving target and a native learning process, there are a few extractable lessons from Taco Bell. For one, AI’s trial deployments should probably be internal and business-facing, before they’re unleashed on customers.
Also consider situational variables. A QSR drive-thru is already degraded in terms of customer agency. Even when talking to a human, sound quality and lack of face-to-face interaction compromise the ability to communicate. Downgrading the comms even further by removing the human seems ill-advised.
Lastly, what’s the cost/benefit of this deployment? By degrading the customer experience in all the above ways – not to mention a potentially-lasting reputational hit – what was to be gained? It would seem those gains are limited to the financial impact of redeploying or eliminating minimum-wage positions.
To be fair, Taco Bell has been honest in treating this as a learning opportunity. And it gets credit for experimenting. Going forward, Matthews expresses his goals to strike the right balance. That could involve first-line-of-defense AI interfaces, monitored closely by personnel who are ready to jump in.
Header image credit: Andrew Valdivia on Unsplash
