Dispelling the all-or-nothing Generative AI myth

Why complete generative AI automation is a myth – and at this point, a mistake

Don’t get me wrong – I work in AI for an AI company (Loris, Hi – you may have heard of us). And I also understand the massive benefits AI can have (including writing the cringy first draft of this blog). But assuming you need to flip this “AI On” switch overnight is not only wrong, it’s a recipe for failure. 

Part of the issue with the recent pace of Generative AI adoption is that it has created this blinding sense of urgency. Not a false sense of urgency necessarily, but such “AI FOMO” that businesses are adopting first and figuring out what it can do second. 

This has perpetuated this myth or misconception that Generative AI not only can do everything, but also that it should. Like most new, but exciting things, the reality is less compelling, yet that’s why it’s important to understand the key areas where AI is coming up short. 

1. Complex problem solving: If you’ve ever interacted with a chatbot, it’s often painfully obvious that complex issues are not its strong suit. Unique, off script, or unexpected customer scenarios often require creative problem-solving skills that today people are great at and AI has yet to learn. While AI is fantastic at understanding complex data, and extracting data-driven insights, it can’t match the creativity and adaptability of human agents when addressing novel challenges.

2. Ethical and moral dilemmas: Given that humans still struggle with these, expecting AI to correctly navigate customer interactions that involve ethical or moral dilemmas is asking too much. It may be objective, but understanding when to ignore policy and act in the interest of the customer, for instance, deciding on exceptions to a policy given an emergency, isor handling a sensitive issue with nuance and empathy. AI doesn’t possess a moral compass and might make decisions solely based on algorithms.

3. Emotionally-charged issues: Generative AI lacks the emotional intelligence and empathy to respond to emotional issues in the same capacity that human agents can. Part of this is just experience – humans have all experienced emotions while AI hasn’t. Much like the complex problem solving issue above, AI can analyze and surface emotional attributes like frustration, confusion, and sentiment, but they can’t replicate a highly nuanced emotional response that a customer would expect.

4. The desire to answer even if the AI doesn’t know: If you’ve been following the news about the creative approach ChatGPT took to legal citations, we all quickly learned that with generation sometimes comes invention – or even fiction. When this is expected or encouraged, it’s called “creativity”. When it is not, we call it “hallucination”. Understanding how to put safeguards in place and insulate customers from potential hallucinations have become a necessary caveat with Generative AI, but not found in all areas of AI like machine learning.

Combining AI and human expertise can automate tasks, but not teams

Rather than a high-risk, all-or-nothing approach, the smartest path to AI in CX today lies in blending the best of AI and human capabilities. AI working in tandem with customer service professionals like agents, QA analysts, and CX leaders can have a number of benefits that create better experiences with less work and in less time. Here’s how this collaboration can benefit both businesses and customers:

1. Automate routine tasks -> More time to be human: Unlike people, AI doesn’t get bored with repetition, like answering the same questions and performing the same tasks. If you can remove these mundane requests, you free up human agents to focus on providing more attentive service. This also has an impact on agent experience, since you’ve given them fewer mundane assignments to get frustrated or bored with. 

2. Making sense of complex data -> Act on those insights: One of the areas where technology excels over people is data analysis. AI can analyze vast amounts of customer data to identify trends, preferences, and pain points far more easily than people. Using those insights, whether responding to customer requests, incorporating feedback regarding products, or suggesting changes to company strategy still needs human insight to put them into use. 
3. Highlight interesting things -> Focus attention where it’s needed: Much like in data analysis, AI can scrutinize a large set of interactions and processes to spot outliers and points of interest that an analyst or manager can delve into more closely. These are currently the processes that are too complex or nuanced to be automated entirely, but can tasks within those processes can.

Where do we go from here?

Ten years from today, this article might seem so outdated the perspective on AI that it would be laughable. But today, Generative AI and other forms of AI have key limitations that make the idea of replacing human agents and CX teams not worth betting your whole company over.  

The foreseeable future of customer experience lies in a collaborative approach, where AI enhances efficiency, consistency, and data-driven insights while human agents provide emotional intelligence, creativity, and ethical judgment.

This is not a bad thing. Businesses that recognize the value of this hybrid approach will be better equipped to deliver exceptional customer experiences, adapt to changing customer needs, and stay ahead in a competitive market. As technology continues to advance, more of those routine tasks can be automated and CX professionals can connect more with their customers. It’s not about humans versus AI; it’s about humans working alongside AI to create a customer-centric future.

Photo credit: superimages123