In 2021, I can ask a speaker to order nearly any product I want, and have it delivered in under two hours to my doorstep.
In that very same world, I can also spend six plus hours on a single telephone call to my bank — using a technology whose core stretches back to 1876 — to report and begin to resolve an urgent instance of identity theft.
On one hand, an indulgence is serviced with such speed and convenience it’s vectoring toward “send it to me before I want it,” as comedian Ronny Chieng quipped about Amazon Prime.
On the other hand, service is so slow and complex that it makes a cruel mockery of the stakes and urgency at the core of the issue.
How can both of these realities co-exist in a functioning consumer-centric economy?
To understand the scope of the disconnect, we need to take a brief detour into the implications of our technology habits and how we interact today with people and machines alike.
The key factor separating great service experiences from the rest is the balance between technology and human support, and critically — the interplay between the two.
We love our phones, to a scary degree (spending 4 hours per day on them on average in 2020, up a full hour per day from 2019). More than half of that time is communications-related: texting, messaging, consuming social media content, creating social media content, checking our social media content’s level of engagement.
But we famously hate using our mobile devices’ eponymous function: talking on the phone (61% of millennials would avoid calls entirely if they could). With the communications asynchrony that comes from messaging, we can be both “always-on” and never truly “on,” which reduces the pressure on us — because, after all, the live nature of conversations means they can be hard, and people tend to avoid hard things.
With the exception of video conferencing and its buzzy-but-questionably-durable cousin social audio — the Internet is still largely an asynchronous place. We browse, post and engage at our own pace. As Bo Burnham put it in Inside, this self-service “buffet” of content delivers us “a little bit of everything, all of the time” (comedians seem to have the pulse on this topic). It’s gotten to the point where we struggle mightily to commit to a single piece of well-crafted, well-portioned media that dares to be as long as a traditional movie.
If committing to an enticing movie is hard, how about committing to a customer service phone call? And if initiating one of these activities proves challenging, how about seeing them through to completion?
Amidst this backdrop, we’ve embraced a surprising degree of autonomy and resourcefulness in servicing ourselves. When we want to buy, book, or barter, we can usually conduct business without having to actually speak to a single soul. When we’re curious about a topic, product or service, our first instinct is to Google it, or maybe, if we don’t feel like typing, we ask the nearest voice assistant. We’ve grown so accustomed to helping ourselves that we even eschew customer service representatives in retail stores in favor of hunting for information ourselves on our pocket computers.
This trend is part of a broader reconfiguration of how we seek and consume information and services — not from people in conversation, but from technology. To keep the buffet metaphor going, we don’t want the proverbial digital waiter to tell us about the specials and come back what feels like three hours later to take our order. We want free reign to fill up our plates with what we want, in the quantities we desire. But we also want our waters full, a clean table, and perhaps most importantly — a real pair of ears to complain to when we are dissatisfied with any aspect of our experience.
In other words, we don’t want service to get in the way of the product. But we don’t want the product to go entirely unattended. Call it “Self-service+”.
Despite this tendency, there’s a strange truth we can’t overlook — we still appreciate the indelible human touch that accompanies the synchronous personal service interactions that go well. Perhaps it’s a feeling that’s accentuated by its modern rarity. Whatever it is, there’s no arguing with the value of expertise when one party clearly has it, and the other does not. Genuine human exchanges build trust and affinity, delivering on a higher psychological level than even the best-designed digital products, not dissimilar to the way that receiving (or writing) a handwritten letter makes us feel inside. Good human service is received so well because we know it isn’t easy, and there’s a difference in knowing that whoever served you actually understood your complex individual needs. Even the most basic human needs are never consciously grasped by an automated system acting coldly on logic, no matter how sophisticated.
This trend is part of a broader reconfiguration of how we seek and consume information and services — not from people in conversation, but from technology.
So, while we are becoming instinctually averse to seeking human-powered service, when we find ourselves in true need of it, the stakes are incredibly high. The times when we have to dial that 800 number, having exhausted all other options. When we have to flag that sales associate down for a non-obvious recommendation. When we need to locate our missing suitcase in baggage claim. These are the instances where our experiences are likely to fall into one of two extremes — incredible frustration, or immense gratitude — and an emotional creature, not a soulless machine, will likely end up on the receiving end.
Every mature service model in 2021 is going to embrace a hybrid approach between humans and technology. The key factor separating great service experiences from the rest is the balance between technology and human support, and critically — the interplay between the two.
So what will it take for humans and machines to work better together to bridge the service gap we suffer through day-in, day-out?
Let technology lead on upper-funnel customer experience (search & discovery)
In the early stages of the customer journey — generating awareness, consideration and even making a purchase decision — well-designed digital products are rightfully winning today. Most sales don’t require salespeople.
We feel empowered when we peruse multiple suppliers to find the best deal on our flight, and pick just the right AirBnB for our needs at our destination. We have fine-grained tools that deliver us detailed, well-presented results, further validated with badges, reviews and comments. When everyone’s their own travel agent, gone is the business of travel agencies.
We also suffer from the paradox of choice and desperately crave curation and simplified decision-making. You see this evidenced in Spotify’s daily playlists or the introduction of a ‘shuffle’ button upon logging into Netflix before titles are even presented. Do it for me, as long as you know me. And at this point, great services have algorithmically earned that degree of trust. That’s why an easy way to take friction out of a customer experience is to remove our sense of agency entirely, or to give us the illusion of choice when strongly guiding us toward a small set of options predetermined to be sure bets.
Of course, there is still a role for humans in helping us find what we want. High-ticket items like cars and homes have become accessible to consumers in rich detail in ways that inform decision-making without human intervention, but more often than not, information alone is not sufficient to convert a sale. This is where a smaller fleet of highly knowledgeable sales and service professionals come in. Building on the highly-informed consumer’s baseline to instill confidence, methodically isolate decision-related criteria and consider multiple options objectively. That means fewer people will be needed, but that their expertise and acumen cannot fail — their exchanges may be brief, but they’re incredibly high-value.
For customer service requests, embrace a hybrid model that isn’t unidirectional
The cost structures of human-powered service, paired with passable-to-impressive quality of speech technology and natural language understanding (NLU), dictate that every company should look to some form of customer service automation. But the limitations are such that humans cannot be fully extracted from the equation.
The way that most service operations are setup today emphasize containment and triage. The lower the volume to the agents, the less hassle to consumers and the more cost is saved by the company (many of our clients estimate each call received represents $7-12 in expense).
This way of thinking is broadly correct from a business standpoint, but it’s overly simplistic. Certain queries may be escalated for a variety of reasons. Most obviously, because they’re unachievable using self-service systems — meaning, humans are the last resort to problem-solving. But there may be more subtle reasons to get a human involved earlier. The customer may not know the problem they’re looking to solve, and disambiguation is a challenge to which humans are uniquely well-suited. It’s a ‘killer app’ for synchronous communication.
Once a consumer understands what problem they’re solving — in the nomenclature of the brand (which likely differs from how they’d describe it themselves) — they're much more likely to be happy with a self-service experience. Another reason might be that a query carries a very high risk profile — where the reputational risks of getting it wrong are such that the dollars-spent on customer service go much farther in preserving customer lifetime value than the average service expense. Brands should segment their customer need states not only in terms of attributes like time-to-resolution or likelihood of chatbot containment, but also the stakes these interactions represent.
Good conversational service design is not purely about escalation triggers, but rather about how human and automated support can gracefully play together to address a specific need. This also means that customer service professionals can and should be users of their own conversational AI tools — for instances where their expertise is limited, why shouldn’t they similarly be able to engage a system for an answer, rather than relying on a specialist colleague? Augmented service professionals can be as useful as automated service experiences.
Brands should segment their customer need states not only in terms of attributes like time-to-resolution or likelihood of chatbot containment, but also the stakes these interactions represent.
Flip service psychology on its head taking a page from social media
The most addictive digital products today make clever use of variable rewards to keep us coming back for more. That means it's unpredictable when we are going to get notifications, such that we get them frequently enough but can’t know with certainty when they’ll come. The pernicious effect of this has been chronicled at length, but what might happen if variable rewards were introduced into customer service experiences? Every company understands that “surprise and delight” moments are golden for building lasting loyalty. Why aren’t more of these injected at moments of service, where loyalty to a brand may be at its most tenuous? Why not offer more deals across more channels (chat, web, IVR, you name it) such that not only are you heading customer frustration off at the pass, but also reinforcing the notion that good things happen when you engage with our brand, no matter how? Whether it's a discount, a free trial, or a simple perceived skipping of the line because of past brand loyalty, a little can go a long way when proactively offered.
Offer premium service, on-demand.
Providing deals and incentives is a sure way to soften the blow that comes from a negative service experience, but could service become an incremental revenue stream rather than a nagging expense line? Certain high-priced products have an expectation of a high touch service level baked in, from Amex to Apple. But what about brands whose base of customers may be willing to pay that much more to hop the line on-demand? The calculus for consumers may be fairly simple — pay a dollar to speak to an agent immediately, versus wait 45 minutes for the next available “free” agent. It may not be something I’m willing to pay for at the outset, but once push comes to shove, I might just open that wallet — and then potentially even convert into paying for that level of service ongoing.
Embrace the affordances of visual IVR & visual conversational aids
When we use visual interfaces for so much of our self-service experiences to explore and order things, why haven’t the visual components of these interfaces made their way more prominently into our customer service experiences? This is not to say voice technology exits the picture, but rather than instead of hearing options, I can visually scan them, and use my voice (if I so choose) as my input modality? A visually responsive system that shows me it hears me — rather than my hope and a prayer that I’m properly heard by an IVR that prompts me to free-from describe something — is a major boost to customer confidence and efficiency. Insofar as everyone has access to a smartphone to dial a hotline, they have a web browser to easily work through similar flows using the input and output modalities of their choice. Voice-led input and visually-led outputs optimize for speed and maximize the affordances of our senses.
What about when talking with someone live — why must visual exhibits be absent? If someone is walking me through options, a list, or data that informs my decision (say, for example, my past service consumption history, or multiple tiers in a plan), why not bake a simple companion web experience into the call, accessible via a one-time link?
Voice-led input and visually-led outputs optimize for speed and maximize the affordances of our senses.
Nail the data & personalization value exchange
“Help me help you.” We often hear this phrase as a way of compelling someone to open up, to give us more information that will allow us to better advise them. The point holds for technology: the more data a system can have about me, the faster and more tailored my experience will be.
A customer profile is often considered primarily as a singular longitudinal view of a person and their track record interactions with a company across channels — ideal in forming a profile that can fuel upsell recommendations, tailored marketing offers, channel-specific communications preferences, and more. But equally important might be the synchronous understanding of that same customer on the phone — the actions they’ve taken to remedy the issue on their own, the amount of time they’ve waited on hold, etc. A bias for recency data to improve a single service interaction can be as important as nailing the long data game.
And there are new tools in the service arsenal to do so — including analyzing our most personal data, our voices themselves, which can reveal much about our mental and emotional states. My colleague has written extensively about the upsides and risks of using biometric voice data, but suffice to stay that if your customer service operation neglects to use these data entirely — across either automated or live exchanges — you’re abandoning your ability to tailor in-the-moment service as well as to glean important patterns over time. You may decide that using voice biometrics to tailor service is not the right step for your customers or operations — out of regard for privacy policies or your expectations of how consumers want their data handled. That’s entirely up to you, but it should be communicated to consumers proactively as a service differentiator that reflects a deeply thoughtful approach to respecting their privacy and security.
What about when multiple CRM systems come into play, across multiple departments or companies? As much as consumers value privacy and security, their expectations dictate that a company’s systems not only talk to one another, but also play nice with other companies (as anyone who’s booked a flight through an aggregator will bemoan). Data-sharing practices and policies that both protect consumers but also enable speedy resolution of inter-departmental, inter-company challenges are key. And as much as possible, this circle of trust should be inclusive of both humans and responsibly programmed AI agents.
Conclusion
As chatbots, voicebots and IVR systems grow more sophisticated, and people grow better at using them and appreciating their strengths and shortcomings, the big question for brands becomes: where does the human connection to your brand have the greatest impact, and how can your technological toolset best manage the rest while teeing up those vital human-to-human interactions?
The future of service is not about replacement but thoughtful augmentation and handoff, transforming synchronous service from a last resort to a breath of fresh air that reminds us that there is still value in talking to each other.
This is all far “easier said than done.” But really, shouldn’t the same be true of service itself?
.jpg)
.jpeg)
