So, it’s a strange truth that in our age of seamless digital interaction, customer service is often awful. We all have horror stories of wasted hours of terrible hold music, FAQs that require a degree in technology, and chatbots that are as clever as a potato.

On top of all this, since the COVID-19 lockdowns, many businesses have scaled back customer service and extended deadlines for complaint resolution. Some have removed helplines, leaving customers reliant on ineffective and frustrating chatbots. But cost-cutting can be a false economy. On average, employees are spending nearly 4.8 days a month dealing with problems – up from 3.8 days last year. As a result, companies are taking record amounts of time to resolve issues and taking record hits to their profits and reputation. Satisfaction is falling year-on-year, with nearly 17% of consumers reporting issues with a trader.

Despite the cost-of-living crisis, research shows that more than a third of customers are willing to pay more for a reliable service, and companies that top the satisfaction leagues are up to 10 percentage points more profitable.

There is a huge appetite for better customer service, but it’s not being delivered.

Could AI be the solution?

When it comes to customer service, most people still prefer the human touch. As advanced as AI chatbots have become, most of us prefer to chat to actual people. There are definitely drawbacks to chatbots, such as the lack of emotional intelligence and empathy – but the main problem is that they just can’t handle complex customer queries. When faced with complicated situations, they struggle to get the job done.

But could that be about to change? Recent advances, particularly in Large Language Models (LLM) have made AI much more powerful and intuitive.

And we’re close to reaching a new milestone: autonomous agents.

When it comes to personal assistants there are three stages of AI intelligence:

• Stage one: the Instructor Agent is AI that can quickly and capably perform routine tasks such as making a payment. But when requests deviate from the routine, the AI can struggle, and the customer needs to be passed over to someone on the customer support team.
• Stage two: the advent of Large Language Models like ChatGPT4 are transforming AI into something close to Reasoning Agents. They have access to huge amounts of information and can have conversations that are fluid and less rigid. This form of AI can answer a far greater range of questions and can scan and summarise information that paints a clearer picture of a customer’s needs.
• Stage three: the current LLMs can find information and reason over it, but they can only solve problems if the path to the solution is straightforward. They’re not yet capable of strategic planning, such as offering advice on growing a business or whether to spend or save in the months ahead. This advanced stage of AI is the domain of Autonomous Agents, AIs able to work with objectives rather than tasks. Their development has only just begun, but they’re likely to hit the mainstream very soon. Give an Autonomous Agent an objective and it can think strategically, creating and executing a number of tasks; it can then change tactics and reprioritise tasks if the situation changes. Agile and proactive, Autonomous Agents have the potential to be a true game-changer in creating services empowering people to take greater control of their lives and businesses.

We may not have reached the era of Autonomous Agents quite yet, but we are seeing a massive increase in the use of AI in customer service.

The financial admin app ANNA is focused on delivering the best possible support for its small business customers. Its CX is delivered via AI chatbot and a customer support team in Cardiff. ANNA’s teams of data scientists, designers, and product developers are working on the best way to combine the speed and efficiency of AI with the more nuanced skills that humans can bring. ANNA’s own language model has been learning since 2018, but it’s now being supercharged by LLMs like ChatGPT and BARD – so where does that leave its customer service?

AI should be an enabler, discreetly ticking away in the background, always there when a customer needs it to solve a problem or perform a task: the ‘artificial’ is what happens behind the scenes and the ‘intelligence’ is the bit that the customers should notice. In customer service terms that means that the AI can do what it does best – tirelessly retrieving the most up-to-date and accurate information (far quicker than a human could) while the customer service team do what humans do best, showing empathy, and using their experience and emotional intelligence.

The art of assistance

At the heart of any great business, no matter whether it’s new or established, is the relationships it forms with customers. Building trust and loyalty, essential when growing a business, happens through great quality services and experiences.

But it's increasingly looking like the art of assistance will become even more refined by harnessing the power of AI. When it's used in a sensitive, thoughtful, and sophisticated way, AI can arm businesses with better data and information. This in turn empowers them to become better service providers, clocking up five-star Trustpilot reviews and delighting customers and staff alike. So, when you’re a startup, why not consider combining human and machine? It might make your life easier, your business better and your customers happier.