Customer service leaders are not prepared for how quickly consumer behavior is about to change. ChatGPT has ushered in a new era in terms of how customers will interact with brands, and it will fundamentally overhaul the entire customer experience, especially service and support interactions. Let’s take a look at what’s going to happen over the next 12 months that will precipitate this change, and what customer service leaders can do today to meet the new standards of customer expectations.
The amount of attention and investment in generative AI since the launch of ChatGPT is paradigm shifting. Microsoft and Meta are betting a significant part of their overall corporate future on the technologies. But perhaps the most important event just occurred in March, with Google’s launch of their own ChatGPT competitor, Bard.
Shortly after Microsoft announced a $10 billion investment in Open AI to help power Bing, I predicted that the amount of attention the investment received would force Google to play catch up and get into the chat game themselves. This prediction is playing out with Bard, which is a direct response to ChatGPT and one that is, by all accounts, already addressing some of ChatGPT’s shortcomings.
But what happens now? The answer is that we are going to see an AI arms race between the two companies with the battlefront being their respective search engines. As they both lean further into generative AI and chatbots, the user experience on their sites will shift rapidly from search queries to chat. And by “rapidly” I mean at a breakneck pace unlike anything we’ve seen from either company. My prediction is that by this time next year Google, which in large part hasn’t changed their homepage in nearly 20 years, will completely eliminate their iconic search bar and replace it with a chatbot.
This will precipitate a profound transformation in online consumer behavior. Conversational search will become the new standard. Potentially, we could see the entire online shopping experience flipped on its head… rather than brands putting products in front of customers, what if customers come to a brand and dictated the entire experience on a conversational basis?
For example, rather than sifting through a brand’s entire selection of bathing suits, or a bank’s suite of accounts, a customer could tell the brand what they’re looking for or that they just simply want to see something new.
Savvy customer service leaders should already understand what this means for their organizations. Customer service chat, which has been a steadily growing if not wholly impressive part of the customer service experience, is going to be catapulted to the number one channel consumers expect to use. And on top of that, consumers are going to have very high expectations of what a chat experience can be due to the billions of dollars being invested in it by Google.
Understandably, the vast majority of customer service organizations who have been used to relying on outsourced BPOs and elementary chatbot providers are going to feel like they’ve been hit by a truck. Expect to see precipitous declines in CSAT, CES, and customer lifetime value by those who don’t start putting in the time and effort to envision what their organizations should look like in the chat era.
Here are three ways that customer service leaders can start making the right preparations today:
1.) Ensure that chat is live 24/7 365 and can cover times of high demand. This should be basic blocking and tackling by now. Consumers expect to engage via live chat whenever they want. But companies have struggled with the simple act of keeping chat on. In fact, we found that nearly 30% of brands turned off chat over Black Friday weekend last year. Brands that find themselves in a similar position over 2023 Black Friday weekend can expect to lose significant–perhaps irredeemable– holiday revenue.
2.) Really optimize for digital. I have had discussions with many customer service leaders who see “optimized for digital” as a box to be checkmarked: If they have chat on their website, then their bases are covered. That, to me, is an incredibly insufficient definition of digital optimization. Do you have chat on every page, including pages that correlate with conversion rates like the checkout page? Is the chat experience designed to pull information about the customer and their relationship with your brand such as whether they’re a repeat purchaser, if they have any outstanding issues unresolved by customer service, or information on what products they might be interested in? Digital optimization is not a one and done effort. Consumer expectations keep changing. So should yours.
3.) Begin the process of future proofing your organization. I believe that by 2030, 90% of customer service inquiries will be automated. Does that sound outlandish? Well consider the fact that Forrester thinks that 70% will be automated… this year. This will upend the concept of what a customer service organization will look like. Gone are the days of relying on huge volumes of customer service agents provided by a BPO to essentially solve customer service by brute force. The customer service organization for the future will be vastly smaller, much more reliant on technology, and led by data scientists and conversational UX writers. Perhaps no other business unit will look more unfamiliar by the end of the decade than customer service. Plan accordingly.
I’ve heard this moment in time being described as the AI tipping point, in which finally the promise of artificial intelligence will begin to yield incalculable changes on society. I fully subscribe to that notion. But even if you’re a skeptic, you have to ask yourself: Are the old ways of doing things tenable? Does it make rational sense to have the same organizational layout and budget allocations that have been the status quo for 20+ years? Regardless of where specifically you envision technology and its impact on consumer behavior heading, it’s heading somewhere fast, and the future belongs to those who take the time to think about it.
Today, automated chat is getting a lot of attention – and investment. Clearly, it is causing a great deal of rethinking and recalibration for CX leaders as to what ‘relationship’, ‘service’, and ‘content’ will mean going forward, especially in identifying what is most influencing customer decision-making. As sophistication and dependence on automated chat build, the high levels of penetration predicted are almost an inevitability.
From my perspective, the greatest impacts are likely to be in customer journey disruptions, which will descend on many organizations and their marketing, support, and communications planning. The author’s statement – “Expect to see precipitous declines in CSAT, CES, and customer lifetime value by those who don’t start putting in the time and effort to envision what their organizations should look like in the chat era.” – is a pretty safe bet, as is the influence on brand advocacy.
There are two significant observations here:
The first is that there will be massive change in the customer experience function based on these tools.
The second is that customer experience executives and customer service organizations are not prepared for this massive change.
Nor are the technology vendors correctly prepared for this.
Since every technology vendor has added AI to its roster of technologies – whether or not they are actually incorporating machine learning in some form or another – and since the amazing capabilities of generative AI have been exposed to the public, now, every provider will be using (or attempting to use) generative AI and large language models such as GPT3, GPT4, BERT, LaMDA, or others.
The fundamental challenge is that generative AI is not a retrieval mechanism and will hallucinate if it does not have the answer. That is especially problematic for regulated industries
But it’s really a problem for every organization. The organization competes on knowledge and if every organization used a publicly available large language model, they would have efficiencies, but they would not have competitive advantage, because competitive advantage comes from differentiation.
This means that generalize language models need to be specialized for an industry, but more importantly, specialized for the organization. Even industry specific language morals can be inappropriate for some applications.
In order to correctly leverage the capabilities of a chatGPT type of experience, the organization has to map and understand the very specific needs of customer segments and integrate generative AI with retrieval mechanisms using organization, specific language, terminology and content.
While there are some organizations beginning this process, there are still many unknowns and many moving parts.
Knowledge is fragmented, siloed and distributed throughout the business, it can be out of date and poorly organized. Retrieval uses metadata to contextualize a query. In most cases, knowledge is not well tagged or structured.
At the end of the day, the organization’s ability to absorb change is fundamentally misaligned with the speed of change of technology (this is not new, but is being amplified by the acceleration of advances)
The panic caused by this massive breakthrough in technical capabilities is rippling throughout the enterprise and the marketplace. But rather than dive headlong into a generative AI project and potentially wasting hundreds of thousands and millions or tens of millions of dollars, the organization needs to take a step back and spend time truly understanding the needs of customers and map the sources and owners of knowledge needed to meet those needs.
My company is working on a large scale, global, multi year knowledge management initiative for an organization whose leadership asked “how do we prepare for chat GPT?“.
The answer was “exactly what we’re doing now”. Because we still need our knowledge house in order to feed these types of tools.
We still need the basic blocking and tackling of good content, knowledge, customer data and product data hygiene.
Without that foundation, generative AI will not be able to solve problems around customer experience and customer support
Organizations need to wary of simple answers and turnkey solutions that will miraculously solve customer experience gaps and deficiencies without human intervention. These are not easy problems to solve and while we have new toolkits that will help get us there, we can’t forget that all of AI runs on data, and in many (or most) cases the data is more important than the algorithm.
You can’t outsource your competitive advantage, and the competitive advantage comes from knowledge of the marketplace, customer needs, solutions, routes to market, operations, manufacturing; of every aspect of the organization and marketplace.
That knowledge is spread throughout the business. Tapping into it requires context and structure.
There’s still no (large language model, generative AI, machine learning magic pixie dust) free lunch.
Over 50 years ago, many of us were entertained AND disturbed by the Kubrick movie: 2001: A Space Odyssey. The climax of the film was the confrontation between the two astronaut navigators and Hall 9000, the AI-driven computer that went rogue based on 18 month old data in its memory bank.
As this excellent, thought-provoking piece implies, not only are leaders not prepared for the ramifications of ChatGPT and its cousins, most of us are ill prepared for this likely revolution. I found Seth’s comment, “…generative AI is not a retrieval mechanism and will hallucinate if it does not have the answer” to be downright scary. What does this do for complex medical procedures, intuitive defense situations, and decisions requiring more creativity than history?
We humans like the core of our relationships to be authentic and compassionate, not “Just the facts, mam,” to quote Sergeant Friday on Dragnet. How is the distinctively human side of customer experience to be maintained? Hal was ultimately controlled through the ingenuity of the astronauts, the type that defied logic and data storage. What does that imply for where we are going? What is the role of leaders in promoting the quest for innovation, not just the search for information?
The increased use of generative AI and conversational search will have a profound impact on business as we know it… as consumers, vendors, employers, and employees. As a consumer, the prospect of interacting with a chatbot is nothing new, but the nature of that interaction is going to change significantly in the near future.
As AI systems continue to develop, we will see a greater emphasis on a more personalized experience that includes instant reference to purchase history and recommendations. Response times in addressing issues will also improve. This has the potential to profoundly shift an organization’s competitive advantage, depending on the speed and quality of their implementation.
The challenge, of course, will be in organizing the sheer mass of information that will be required to feed and “train” the systems to operate properly and represent the desired brand-promoting interaction.
While our access to information and support features will increase exponentially, it’s important to realize that there is significant (and often commercial) value in the human connection. People often make purchase decisions based on emotion. Repeat business (a significant contributor to higher profitability) is significantly influenced by the quality of the relational experience that a customer has with a brand through its employees. And the best sales and service representatives are often those that master the art of emotional intelligence (rather than data). That being said, the relational nature of the AI interaction will also improve over time.
We also need to think about the employee experience (picture an inward-facing customer experience). The launch of these platforms will also have a profound impact on a company’s Human Resource practices as they will require a new set of practices for the appropriate use of the technology. Their use in serving the internal customer will also increase from their first interaction with the company to their eventual exit. This will be a significant shift with massive benefits and risks of its own.
One word of caution for organizations who are on this journey… Employees are increasingly using AI to innovate and solve problems more quickly. That’s wonderful. But clear guidelines need to be established around the loading of proprietary and/or confidential data. Just food for thought.