Artificial Intelligence (AI) is hot. One breathless press release predicted that by 2025, 95% of all customer interactions will be powered by AI.
AI is not new. It’s not just about “bots” for self-service. Or self-driving cars. In general usage it means the usage of advanced analytics — more than process automation based on rules. Can include the processing of “natural language” (e.g. Alexa, Siri, Watson), decision making using complex algorithms, and “machine learning” where the algorithms get better over time.
Here’s one definition from AlanTuring.net:
Artificial Intelligence (AI) is usually defined as the science of making computers do things that require intelligence when done by humans. AI has had some success in limited, or simplified, domains. However, the five decades since the inception of AI have brought only very slow progress, and early optimism concerning the attainment of human-level intelligence has given way to an appreciation of the profound difficulty of the problem.
And another from Wikipedia:
Artificial intelligence (AI) is intelligence exhibited by machines. In computer science, the field of AI research defines itself as the study of “intelligent agents”: any device that perceives its environment and takes actions that maximize its chance of success at some goal. Colloquially, the term “artificial intelligence” is applied when a machine mimics “cognitive” functions that humans associate with other human minds, such as “learning” and “problem solving” (known as Machine Learning).
IBM has been pushing Watson (of Jeopardy fame), Salesforce.com launched Einstein last year, and my inbox is full of press releases and briefing requests this year from vendors big and small, all touting AI.
My question is: Can AI improve the Customer Experience? Please answer “yes” or “no” and explain in the comments below. Examples appreciated!
Yes, it can, but it is not only AI that is involved there – massive computing power is another one, perhaps even more important. And, of course, data – LOTS of it. AI is just a part. Still, many companies (and investors) are betting the farm on it, also by using a wide meaning of AI, often covering not much more than rule based systems.
A bit more concrete: A better experience starts where I get better kb or forum search results with the results that the system deems relevant actually being relevant. It is not feasible to have the corresponding text analysis and tagging done by humans, and star based rating systems are not the real thing, either.
One main concern is about usage: If companies use better algorithms based on an inside out thinking the customer just gets more focused spam and perhaps also feels annoyed by continued intrusion. Finding the (individual) boundary might be something for Dr. Watson, too 😉
2 ct from Down Under
But… Will it? No.
Can AI Improve experience? Eventually, it will have the capacity. Part of the reason it will be able to improve the experience is that the current human experience is still woefully inconsistent. Improvement shouldn’t be a challenge.
What will prevent it from actually improving experience is the reality of corporate dynamics.
Right now, early adopters are jumping on it because it is a new bright shiny object. It is being promoted passionately on social media and discussions by companies that have a vested interest in selling the technology. It is being consumed by companies with visions of lower operating expenses dancing in their heads.
For the last couple of decades, companies have relentlessly tried to push consumers to more cost-effective support options. They hide phone numbers, and make consumers jump through profoundly annoying self-serve hoops before allowing them to escalate issues.
Many still outsource to call centres that provide low-cost services, but are staffed with under-trained employees with substandard knowledge-bases to work with.
Why do we expect it will be any different with AI?
AI will still require human interventions for issues that are non-linear, and I suspect that companies will still do their best to prevent consumers from reaching those humans.
There are far too few organizations that think long-term when it comes to customer-focus, loyalty and profitability, and too many focused on quarterly P&L.
Sorry – I seem to have woken up cynical this morning….
The short answer is, “It already has been improving Customer Experience.” Unless you subscribe to a very futuristic definition of AI (such as the ability to fully achieve generalized human intelligence), there are countless examples today of where the loosely used term AI (which is the one used by all those inbox press releases you get) – has been improving CX. More and more may not even have much actual predictive intelligence, but just have forms of automation (such as rules management, BPM, and RPA).
But there has been predictive AI in use for CX for years. For example, every time you go on an electronic device, some form of AI is probably affecting your experience, from voice interfaces, look ahead search terms, site personalization, offer & price optimization, and so forth.
The question at this stage is “how much MORE can it improve Customer Experience?” I believe the answer is – quite a bit. I believe there is a 5-10 year lag between all the AI tech available, and the ability for brands to put these technologies to effective use. And we all know the innovations won’t sit still, as huge $ is still being poured into AI VC (although its slowing some).
So no matter what, we are in for an exciting next 5-10 years!
Yes, with massive caveats.
There are historic parallels, and also lessons, for the CX growth of AI; and, as you note, there is evidence of AI spreading like an unchecked virus. Application of AI, though, isn’t universal. It is situational, dependent on whether it improves, functionally and emotionally, the consumer (or employee) experience, circumstance to circumstance.
A lot of AI is proceeding like the Manhattan Project, building on superficial and assumptive results to move to the next level of application. The biggest component of artificial intelligence is human integration, how well it melds with day-to-day stakeholder use. Don’t see nearly so much evidence of longitudinal testing results here. A lot of pivotal things – including level of sophistication and stakeholder control (think Hal, in 2001: A Space Odyssey), enterprise culture and processes – can get in the way of AI effectiveness, so more study , in more industries, in more circumstances and situations, is needed for its specific application to CX..
Shaun, AI already now DOES improve customer experience. Not everything is about customer service, but even there we see improvements – while I agree that many companies are more looking into cost reduction than anything else. But then AIs (in the wider sense of the word) underpins knowledge base searches etc. On the marketing side we see adoption of technologies that reduce the real irrelevant stuff from advertisements, improve channel delivery and overall reduce some clutter. While we are not exactly customers there is a lot of AI behind Google search, or Bing, … I do think that we just take much for granted.
And, beyond the hype, vendors are already smart enough to package AI into their business applications – in a few years it is just a staple that works behind the scenes.
In short, yes! Like all smart programming AI requires a lot of care and feeding; and like all customer experience issues, AI needs to be directed carefully. I remember mis-applied AI in the early days using an IVR system, and customers got so flummoxed that they opted out of the IVR and left the company entirely! Still, when AI is used for challenges that can be modeled such as “how can we predict customer effort (and produce an automated Customer Effort Score, or CES)?”, and AI’s recommendations engines suggest course corrections or next actions (much like Amazon’s “Customers like you also bought ….”), and then AI “learns” how its recommendations worked and effects changes (like a course correction when sailing or piloting a boat) — when all of this happens, AI can (and is) most definitely affecting and improving customer experience.
There’s no doubt. Chatbots are all the rage and they are actually useful! AI is actually programming automation now. In home I of T are making people more and more dependent on AI. Its only going to get better. The real question is where wet computing (us) comes in. Stephan Hawking precicts our role will be superfluous….
Thomas, I agree with the examples you’ve given as having improved CX. But are they really AI, or are they just advanced algorithms?
Even if we do choose to define them as “Intelligence,” I think that human intervention still plays a critical role for those times that the problems are too complex or abstract. My point is that companies don’t have a great track record in compensating for technological shortcomings that can actually detract from customer experience. They are equally likely to just ignore them.
I remember many many years ago listening to a large telecom shouting the virtues of their new IVR system – refusing to listen to the 16% customer approval rating. They couldn’t help but try to justify the tens of millions of dollars they were saving. Customer Experience suddenly became irrelevant.
That’s why I believe that, although AI can eventually improve experience, there’s a good chance it won’t for a very long time – if at all.
As I mentioned before, I’m usually not this cynical. Must be the time change….
The short answer is an unequivocal yes – AI can improve customer experience, and is actively doing so today. “AI” is already both more intelligent and less artificial than many think.
Consider that voice assistant software is the No. 1 AI app today (think Lil Miss Red – type askairasia.com to talk to her; ‘she’ helps AirAsia customers by answering Twitter, Facebook and chat enquiries). Or the roboadvisors trend led by firms like Schwab and Vanguard. Or the prediction that 85% of all customer service interactions won’t require human reps by decade end.
AI will continue to accelerate, as its already impressive ability to do things like access and apply data, streamline processes, focus actions and model future scenarios gets ever better.
A “CX enabler,” AI increases the ability to personalize and customize interactions by making them more “human” – in many cases, without humans at all.
Thankfully, machines can’t think—yet. But they can do a decent job faking it, provided their application in the realm of customer experience aligns with customer wants, needs and perceptions.
Automation of all kinds can improve customer experience to the extent that what the customer is doing in their CX journey is tied to something electronic. This is the case primarily for information-seeking sections of the CX journey.
Keep in mind that many types of products are bought and used without any information-seeking sections of the journey, or anything tied to something electronic. This is especially true for some B2B industries.
And even for high-intensity electronics/info-seeking CX journeys, there will always be policies, processes, handoffs, attitudes, and strategic/tactical decision-making behind it all by the supplying firm — these are huge determinants for the quality of customer experience in its totality.
CRM, IVR, self-service, and enterprise feedback management are all examples of CXM budgets going overboard in favor of shiny objects that managers often suppose can be plugged in to rapidly reduce costs, solve headaches, and drive revenue. These budgets are typically overboard on the technology itself, leaving little investment opportunity in their wake for the cultural, ecosystem, people, and process investments that really must accompany the shiny objects for full success.
One of my favorite articles on this topic is “Don’t Confuse Customer Experience Technology with Customer Experience Management“
Yes, and it already is.
Artificial intelligence and machine learning are extraordinarily powerful, and there’s significant opportunity for businesses to leverage that technology to improve, enhance, and evolve customer service efforts. At Nuance, we focus on three types of intelligence that are key to providing effective customer care: conversational artificial intelligence, human-assisted artificial intelligence, and cognitive artificial intelligence.
Conversational AI is the ability for people to naturally engage with machines using everyday language. It requires the ability for machines to understand speech and text, understand context, and provide written and spoken responses back to the person. These technologies can be leveraged in automatic speech recognition, natural language understanding, text-to-speech, and biometrics technology in order to provide customers with a seamless, natural way to communicate with businesses. Human Assisted AI is the ability for a machine to recognize when it doesn’t understand something, and then engage a human agent to help, ensuring customers are always led to the most appropriate resource. In addition, the technology works in the other direction as well. When a human agent is dealing with a customer, the system empowers the operator (think Robocop) with context-suggested responses and predicted actions, so that agents can provide effective service to customers. Cognitive AI refers to the ability to leverage the data that is generated by the billions of conversations that occur in a system to power learning from human agent interactions. As a result, the machine can start to automate more common transactions, anticipate customer problems before they occur, and provide solutions that are smarter faster.
Yes it can! From my vantage point in the contact center various AI tools help cut down agent training time by helping them locate answers faster and externally it improves the quality of self-help. It also helps with other functions like case triage to the right department based on the content of the message, potentially reducing handle times. It’s not difficult to tie either of these examples back to improved customer experience. This stuff is fascinating and I’m interested to see what happens next.
I am in agreement with all of the thought leaders who have already commented. If AI can (and already is) have an effect on improving the accessibility of the customer experience, then it is a hugely positive advancement in the digital revolution. However, we must never forget that even where it can have a positive effect, it is in ADDITION to the undeniably important HUMAN interaction that is far more likely to have a positive effect on what customers REMEMBER about their experiences and what turns them in to loyal advocates of a brand.
Hi Ian, yes, the main thing is the human touch, and augmentation instead of replacement of human support. But let’s do a thought experiment, after I had a very similar discussion just the other day:
Apart from their ability to efficiently and effectively solve my problem one of the core traits that is expected from a service agent is empathy. Now, a lot of support happens via phone or chat, generally digital channels that do not offer an immediate visibility of the opposite party. You are calling in for support and perceive empathy and receive the result that resolves the issue in your sense. The question is: Does it matter whether the empathy is ‘artificial’ or not, if you cannot distinguish between the ’empathy’ brought forward by a machine and the empathy brought forward by a human?
As said, a thought experiment. I do not have a real answer myself. Yet.
A very interesting thought experiment indeed! I do not have an answer to that either, although I would need to be convinced that artificial empathy is not a contradiction in itself!
What came to mind with that thought experiment, Ian and Thomas, is the fact that “support” is not customer experience management. We must be much broader in our definition of CXM to address the end-to-end customer experience journey across the entire customer life cycle. Otherwise we are drawing conclusions that eclipse what customers really need.
I know that as a customer I need my suppliers to minimize the probability of situations where I need support! All customers want to buy products and services that work right the first time and every time (as interpreted by the customer’s expectations!), without interruption of contacting support. Application of AI toward this broader goal would be noble indeed.
Still, there will always be people in companies designing AI, managing all the functions of the company, establishing and executing policies, processes, handoffs, accountability, agility, suppliers, channels, alliances, business models, etc. And for products/services such as Alexa, ensuring privacy, safety, etc. CXM must also address all of this, or we are still in danger of mucking up the goal of being right the first time and every time for each customer’s expectations and the goal of maximizing business growth in all dimensions.
I think we need a balance of all of the above, an acknowledgement of their relative roles and their synergies.