Our relationship to work is in the early stages of changing forever as our capabilities are drastically enhanced by the technologies associated with AI. I’ve been working in this space for decades, but the general public got a massive dose of what I’m talking about with OpenAI’s release of ChatGPT. Suddenly, machines can interact conversationally, responding to written prompts with stunning levels of sophistication.
The concerns about AI replacing jobs are valid (if not completely founded). Many companies are taking a short-term approach to implementation centered around cost-cutting and immediate profit. This kind of activity is bad for jobs—not just in the roles it eliminates, but also in terms of the quality of the jobs that it doesn’t render obsolete.
In his contact center automation predictions for 2023, Gadi Shamia notes that past solutions designed to get agents to answer more calls faster are missing the mark and that the focus should be on freeing agents from repetitive tasks. He points to an automobile association that uses automation to handle standard roadside emergency requests.
“By automating these non-hazardous calls,” he wrote, “the company reserved agents to deal with complex issues, such as a passenger stuck on a snowy highway at night.”
This points to the real power of AI. When used to improve both customer and employee experience, it’s a total game changer. This requires more of a long-term mindset, but the end result can be jobs that include significantly less tedium and significantly more human interaction.
There are lots of jobs that become more rewarding—both to those performing them and the customers they serve—when humans get to spend more time interacting with one another. The human element is something that AI can’t replace: our instincts, our ability to contextualize information, and our creative problem-solving are essential for AI to thrive. In business settings where humans and machines are collaborators, blending their strengths, AI can improve experiences for employees and customers.
HOSPITALITY, TRAVEL, AND SERVICE
Hospitality and service industry jobs are all about the human touch. As a guest at a hotel or restaurant, people generally want to feel cared for the way they might in someone’s home. Personal attention and an array of options will keep most customers feeling rewarded. Unfortunately, most current automations are designed around cutting costs, not creating better experiences for customers or employees. I recently had a conversation with award-winning travel journalist Peter Greenberg, who was very passionate about this point.
“Anytime I get to a hotel and they steer me toward a kiosk, I leave the hotel. Anytime I get to the airport and they don’t want to do customer service and they steer me toward a kiosk, it drives me nuts.” Greenberg said. “When they ask me ‘Why don’t you want to use the kiosk?’ I explain to them, ‘Because I’m trying to keep your job.’”
Efforts to cut costs by offloading check-in tasks have created a technology barrier, the kiosk, creating subpar experiences for customers and employees. The solution is to put conversational AI into the background. Kiosks might still be useful, but the experience of using them would be augmented by humans who know how to smooth out each step in the process and who have access to relevant information about their surroundings and their travel path.
Scenarios where automation isn’t all front-and-center or on-screen create opportunities for deeper connections and more meaningful connections between customers and employees. In these scenarios, most automation does best by dissolving into the background and augmenting a more meaningful experience with another human.
SALES
When people are making high-value long-term investments, like buying real estate and cars, a human touch does a lot of the heavy lifting. Real estate transactions in particular involve large amounts of money and long-term investments. There are opportunities for AI to improve all 360 degrees of these experiences in ways that don’t threaten real estate agents.
In a recent episode of his “Keepin’ it Real” podcast, RE/MAX CEO Nick Bailey asked Kai Larsen, associate professor of Information Management at the University of Colorado in Boulder, about how AI might affect the real estate industry. Larsen points out that this technology is in its nascent stages and is hard to predict. His recipe for implementation is sound, “You start with something that’s been trained to do a good job in this area and then you add the human being who’s an expert to improve it further.”
This fits well alongside Bailey’s mantra, “Technology exists to support real estate agents, not replace them,” and in the same episode, he’s joined by a RE/MAX agent, Michael Thorne, who has adopted ChatGPT to augment his team. “This tech shouldn’t come between you and your clients,” Thorne says. “It should allow you to spend more time [with them].” Thorne describes using generative AI to review existing blog posts, identify knowledge gaps, and craft new blog posts in a similar style and envisions a future where automation can move tasks like document signing further into the background, so agents and buyers can establish deeper relationships.
Even something like bookselling can be seen through this lens. Reading a book is an investment and there are buyers who want nuanced recommendations that aren’t fueled purely by algorithms. Other retail settings share a similar dynamic. Aspects of buying things like clothes and groceries will become easier to automate, but a human touch will continue to enhance the experience.
MARKETING AND PROMOTIONS
It stands to reason that lower-level marketing jobs won’t involve a lot of copywriting, but might instead include lots of copy curation and cleanup. Higher-level marketing jobs might become more about finding ways to connect with consumers in more personalized (and, paradoxically, less creepy) ways, which is something that humans can do with AI tools at the ready.
It’s hard to predict the shape of marketing to come, as marketing departments tend to operate in pursuit of short-term gains, but there are plenty of opportunities for promotions to become more contextualized for increasingly granular segmentations, ideally in ways that lean more on human-to-human interaction.
This will require a shift in thinking about the practices behind marketing, which often involve misdirection. Depending on what’s being sold, that might be a combination of things, like exaggerated claims, confusing language, or false scarcity. It’s easy to imagine scenarios where companies are developing their own large language models that can interact with customers on the company’s behalf. It’s likely that these experiences will involve marketing endeavors that leverage all kinds of customer data to create more personalized messaging. It’s also likely that consumers will have access to their own LLMs (either models that are trained to interact with company LLMs on behalf of customers collectively, or bespoke models that act more like personal gatekeepers).
Marketing and promotions are a bit of a cat-and-mouse game and humans will continue to be involved, whether that’s trying to be more sophisticated in outwitting consumers or, more hopefully, working to make products and services more transparent for the benefit of everyone.
HIGH-LEVEL WRITING AND THOUGHT LEADERSHIP
Large language models will continue to get better at interpreting prompts and writing human-quality text. This isn’t great news for copywriters who churn out copycat LinkedIn ads, but for writers who have something valuable to say—and particularly those who’ve already written or said a lot—AI lets them broaden their capabilities while reimagining their brand.
My colleague, Josh Tyson, has been writing for 25 years and we’ve discussed the possibilities that might emerge if he trained an LLM on his collected works. He could feed it Google Docs, URLs, and scans of old magazine and newspaper articles. He could have it listen to microcassette recordings of interviews with directors and musicians, plus the hundreds of hours of podcast material he’s created. If he was embarking on a new writing project, Josh could ask the LLM to summarize everything he’s written or said in the past about the troubadour Donovan or Chicago’s art scene in the early ‘00s or interaction design. He could even ask an LLM to seek out connections between those three disparate topics. Josh might want to see data visualizations showing adverb usage over the past 15 years. The possibilities are endless.
Authors and thought leaders with their own personal brands might be compelled to create digital twins to do all sorts of things. I had a recent conversation with bestselling author Charlene Li about this idea and she lit up, saying she’d think of her digital twin like a second brain, “I could ask myself questions!” She gave the example of prompting her LLM for thoughts about hospitals’ relationship to failure, noting that she’s written about hospitals’ and businesses’ relationships with failure but not about that topic specifically.
“I can scale myself,” she said. “It’s not just another me, it’s a better me… 70% of what I do I can now do with GPT.” Li doesn’t see this as a threat. She sees an opportunity. “I can do more of the 30% that’s unique to me and keep working with the technology to keep adding on to that so I can do other things—higher-order more value-added things.”
It’s not hard to imagine that authors like Josh and Charlene might decide to sell their learnings and expertise in the form of access to their digital twins, rather than as a book. A digital twin can answer questions with information tailored to individual users based on their industry and level of experience. Again, the possibilities here are endless.
DESIGN
There’s never been more of a need for designers. We’re in the throes of unleashing radically enabled experiences on the public and only designers can make sure they are safe, reliable, and sustainable. Traditional design skills like attention to detail and expertise with layout and composition will always be valuable, but technical training won’t be a barrier to entry. There are already code-free tools in the marketplace that allow people with no computing expertise to design conversational experiences.
Plenty of people are already refining their skills as prompt engineers in their daily interactions with ChatGPT. As LLMs like GPT become part of ecosystems of interconnected technology and data, designers will be the ones who create the experiences that sequence these elements together.
For instance, if I want to book a flight in the not-to-distant future, my primary interface might be an LLM communicating as a rich web chat that lets me converse with one of an airline’s many intelligent digital workers (IDWs). I’m a frequent traveler and this IDW knows that I’m willing to fly at off-peak hours if I can get a discounted seat in business class. I might also get discounted rates, not because I’m part of an obtuse frequent flier program, but because the airline frequently runs complex simulations to figure out the best ways to reward individual customers based on their purchase history and browsing tendencies.
Every part of an experience this rich has to be designed. This work is done by many designers who are accounting for many things beyond the scope of the design we’re accustomed to thinking about. It’s been optimized to give me a rewarding experience, but it’s also been reviewed by a team of designers focused on the ethical considerations that go into something this personalized. Designers who are looking for ways to factor sustainability into each step had their iterations reviewed as well.
Designing never ends with technology that’s this powerful and applicable in so many situations. This is work that should keep designers busy for the foreseeable future.
The most effective automations across industries like travel, hospitality, marketing, and sales will almost certainly be designed by the people who have the most familiarity with the tasks that are being automated. This means people who find ways to combine their knowledge and experience with AI-enabled tools will always have work waiting for them.
