The changing nature of AI in Customer Support

Something significant is happening in Customer Experience (CX), and the dominant narrative is missing several important points. That narrative goes like this: AI is getting better fast, so full automation is inevitable, so human support teams will be a thing of the past.

There's some truth in this, but the overall picture is dangerously misleading. AI is genuinely transforming how support works, but the conclusion that human agents will be phased out everywhere as AI capability improves is a mistake.

The future of support is not human or AI. It's both together, each one doing different things, and in the best-case scenario each one is doing those different things better than ever before, creating massive upside for companies and their customers as a result.

Let’s explore what’s happening right now, and what it means for the brands who truly value the experience of their customers.

How AI is being used in Customer Experience

AI is showing up in Customer Experience in two distinct ways, and it makes sense to discuss them separately.

The first is front-end AI.

Customer-facing AI means chatbots, automated resolution flows, intelligent routing and voice assistants. All these technologies handle interactions that could be said to follow less complex or predictable patterns.

Suppose a customer wants to track a delivery, understand a charge on their account or process a return, AI can resolve all these cases, and dozens more like them, inpredictable and repeatable ways. A well-modelled AI system will close the loop faster than a human agent stepping through the tickets backed up in a queue, and it will do so 24/7/365, with human quality control overseers continually monitoring the outputs and retraining the model wherever necessary.

The second way is in the back-end.

These tools don’t face the customer. Instead, they sit behind the agent, enabling everything from instant retrieval of essential context that might take a human ten minutes to fully process, to customer sentiment analysis, to lists of resolution steps that have already been tried or might be tried next.

These back-end elements, invisible to the customer on the phone or chat interface, supercharge the human-to-human conversation instead of replacing it. A human agent who can get to the right answer in five seconds instead of five minutes creates a powerful, brand-positive experience for the customer on the other end.

The benefits for both business and customer here are significant, building a value chain that can transform a frustrated customer into a loyal advocate. Other underappreciated benefits accrue to the agent too, leading to a further value unlock down the line.

With well-modelled AI support in their corner, the agent can be more confident, consistent and helpful — and much less likely to be burned out by the old frustrations of working on the Customer Experience front-line.

Both the front-end and back-end AI usage matters, but each one is pointing in a different direction. Front-end AI reduces the number of interactions that require a human. On the back-end AI, it’s raising the quality ceiling for the interactions that do. It’s win-win-win all round: customer, brand and support agent all benefit, creating a positive flywheel that compounds massively overtime.

“It’s  win-win-win all round: customer, brand and support agent all benefit,  creating a positive flywheel that compounds massively over time.”

The interactions that remain

Estimates about the number of customer support interactions now being handled by AI vary, but the direction is consistent. Some Otonomee customers are seeing in excess of 50% of enquiries fully resolved by AI without human involvement, while the likes of Intercom has reported that its Fin.ai service is hitting 80% in AI resolution.

But what about the remaining 20-50%?

This is where the opportunity lies for brands that take Customer Experience seriously. It’s uniformly the harder stuff, problems that require nuance, experience, and sound judgment to resolve. It could be a patient navigating a HealthTech platform who doesn't fully understand what their own data means, or where to turn to for help. It could be a fintech customer whose account has been temporarily suspended and is desperate for a resolution.

Two overriding hallmarks of these interactions are that (1) they’re often technically complex or (2) emotionally charged. In many cases, it’s both. The person on the other end, does not want more information. What they really want is a human being to hear them, understand the situation, and help them through it.

It’s important to consider here how human beings experience problems. When something important goes wrong, especially with health or financial technology, or important travel plans, people want to speak to someone whose good judgment and empathy can help bring the matter to a satisfactory conclusion for all concerned. In this type of situation, interacting with a bot is unlikely to appease them. If anything, it is more likely to add to their frustration.

The agent role is changing, but it’s not disappearing

These issues that are left behind in the rush to AI and automation are both more complex and consequential, and therefore they require more sensitive handling.

This makes the agent’s job more challenging, but for the right people, it is also now more interesting and rewarding. Agents who previously spent most of their time on straightforward or tedious tasks now get to make a real difference in interactions that require the “power skills” their career has given them: active listening, creative problem-solving, genuine empathy and the capacity to build trust with another human being in a pressured situation.

For companies building support teams, this means the ideal candidate profile is changing fast and moving in two directions at once. On the AI side, roles with titles such as Knowledge Manager, Conversation Analyst and Support Automation Specialist are increasingly common.

On the human side, however, the ability to handle complexity without a go-to script is now more important than ever, and agents will require an even greater range of emotional intelligence (EQ) to move from one complex issue to the next. Many of the best Customer Experience agents in the future will be part practitioner, part systems thinker, and all human. And acting autonomously.

The Otonomee view

None of this should be interpreted as an argument against AI in support. At both front-end and back-end, AI is rapidly playing a bigger and more important role, and this will continue to develop. However, we foresee that a large minority of customer support cases will still require expert human handling, and that these cases will have an outsized impact on brand reputation and business growth.

By asking questions such as “what does the best possible interaction look like, and what combination of people and technology gets us there?”, brands will be able to turn their existing customer base into one of the best marketing channels they will ever have.

Otonomee has been building support teams for brands in high-expectation sectors like HealthTech, FinTech and travel for more than five years, and our leadership team puts several decades of senior management experience in the CX industry into play for all our clients.

Our experience has taught us that there are at least two types of companies in the world.

The first treats customer support as a cost that must be minimised.

The second treats their customers as a vital relationship that’s integral to growth and success.

These two company types are playing completely different games, and we believe it’s obvious which one will win in the long-term.

Is it possible that AI will bring business to some utopia where customers will self-serve happily and endlessly on robot-assisted interfaces?

Maybe, but we believe it will be more nuanced than that. We believe that AI will increasingly enable exceptional people with experience and sound judgment to do excellent work every time a customer needs help with a complex problem that really matters.

With AI’s influence continuing to grow rapidly, the Customer Experience function is undoubtedly changing.

We don’t believe it will be replaced entirely by robots. We see the agents of the future being more skilled, more fluent in technology, and more capable of creatively solving hard problems.

And that future is already here, because it’s in action at Otonomee every day of the week.