Insight
AI and Marketing: What You Should Know
March 1, 2024
NextAccess recently attended a roundtable event held in NYC which brought together speakers from some of the most well-known companies out there: Ogilvy, Zappos, Meta, and Microsoft, among others. Designed for a marketing audience, the event sought to untangle some of the confusion around AI and to discuss how marketers should be thinking about incorporating AI tools into their workstreams and processes in a realistic and methodical way.
The key takeaways were perhaps more thought-provoking than groundbreaking, but certainly worth consideration. Let’s take a look at some of the overarching insights:
AI will take the place of grunt work to free people up for higher-value tasks
There’s been a great deal of discussion —deservedly so—about the jobs that AI will be taking away across industries and functional areas. Everyone from writers to coders are likely to be impacted. What these industry leaders are suggesting is that AI tools will automate the laborious parts of people’s jobs, e.g., the dreaded note-taking, the inventory auditing, the spreadsheet populating. No longer will Joe have to take notes on an endless meeting and have to debate whether to include vacation plan discussion in said notes. Just record that call, have Joe upload the transcript to Claude, and voila, a beautiful summary with key points and next steps. Joe is free! To do what? That brings us to our next point.
AI still needs a human being to be successful
Getting back to Joe, who was hired to do office manager tasks. Do you even need him now that you’ve got your nifty Claude and other AI? Yes, you do. Not only does Joe still have to do the tasks that take the generated information and share it appropriately, but he also needs to check the output. Because yes, even Claude is not infallible. Your summary will be mostly accurate, but the names might be slightly off or attributions not quite right. Your human needs to go through it and make any adjustments needed to make sure it’s accurate. While a few errors here and there might not be a big deal in many cases, sometimes they can be and companies need to watch out for that.
For an amusing example, we tried out an image generating tool, Ideogram, to see how it would handle detailed prompts. We loved the image it came up with of our Thai Ridgeback in a future dystopian hellscape:
…until we noticed that he had five legs and maybe an extra ear growing out of one of his ears. As we say, it’s all fun and games until someone grows an extra leg.
Safety and accuracy in AI are paramount
Companies love chatbots, and we imagine that the proliferation of AI-generated chatbots will reach new heights in 2024. According to the event panel discussion on “Intelligent Bots for Better Customer Service,” a Forrester study with Meta showed that 70% of their customer queries were resolved on chat, the rest by agents. This in turn reduced agent turnover, as bots relieved agents of the need to answer the same questions a hundred times. And yet. Caution is needed here, particularly in cases where consumers not just expect but demand accurate information. Case in point: H&R Block and TurboTax unveiling chatbot assistance for income tax filing questions. As the Washington Post noted in an article on March 4th, the results are less than stellar, with TurboTax responding with wrong answers over 50% of the time and H&R Block not faring much better with over 30% incorrect. In the rush to boast of implementing new AI tools, companies should not skimp on proper testing.
The bottom line for marketers wanting to incorporate AI into their toolkit is this: stop thinking of AI as a tool and start thinking of it as a collaborator. It’s not a plug-and-play monolith that a company can integrate and then step back to let AI handle everything. Or at least it shouldn’t be. Will some companies do that very thing regardless? Of course they will, and we’ve seen plenty of examples of that already. Smart businesses, however, don’t want to be roasted by their customers on social media. They’ll focus on improving customer service and outreach by implementing a well-thought-out AI plan that increases internal productivity—while not rushing out external AI solutions before they’re ready for prime time. Those are the ones who’ll succeed in both the short—and long—run.