Customer Support With AI Is Impressive, But We Still Need A Human In The Loop – For Now

AI’s Growing Role in Customer Experience (CX)

It is surprising how quickly artificial intelligence (AI) has started to dominate every discussion involving business strategy—including in areas such as sales, marketing, and customer service. Any manager planning a customer experience (CX) strategy today is first asking, ‘How can AI help me to do this better?’

The Rise of AI: From Research to Real-World Applications

Although AI has been the subject of research papers for at least 75 years, it has really only had strong public attention in the past two years. This was because of the launch of ChatGPT from OpenAI at the end of 2022. This app caught the imagination of the public because it was so easy to use - the app had over 100 million users just two months after it was launched. TikTok took nine months to reach this level, and Instagram took two and half years.

Understanding AI’s Evolution: From RPA to ASI

The founder and CEO of HfS Research, Phil Fersht, recently shared an infographic describing the development of AI. Commenting on this, he said:

“We started with RPA to automate tasks, then graduated to GenAI to create new content and workflow, before Agentic came along to do a lot of this for us autonomously and collaboratively. Eventually, AGI will supplant us until ASI outperforms us.”

Just in case you are not familiar with all the terms Phil was talking about:

• RPA (Robotic Process Automation) automates repetitive tasks.

• GenAI (Generative AI) refers to smart chatbots like ChatGPT.

• Agentic AI enables autonomous AI assistants.

• AGI (Artificial General Intelligence) aims to mimic human intelligence across multiple tasks.

• ASI (Artificial Super Intelligence) would surpass human intelligence.

You can tell from Phil’s statement that we are just about at the Agentic stage. It is now possible to create autonomous assistants that go off and perform the tasks we set them and then report back. This may soon become very interesting for customer experience designers because, as familiar tools like Siri and Alexa start embracing Agentic AI, we might find that people start asking their digital assistants to make purchases or follow up on customer service problems.

AI’s Current Limitations in Customer Support

But what is clear is that this is currently the limit. AI today looks very impressive compared to AI before the pandemic, but it is still focused on narrow tasks. A good way to understand this is to realise that if you train an AI system in how to play chess so well that it can beat grandmasters, and then ask it to program a delicate clothes cycle on your washing machine, it will have no idea what to do. It understands chess and nothing else.

Some AI experts claim that the era of AGI is imminent, but the vast majority of people working in AI believe that we are still many years away from creating a computer that resembles the general intelligence of a human brain.

AI and Human Collaboration: The Best CX Strategy

So what does this mean for anyone planning how their business needs to engage with customers in this era of AI?

AI is an extremely useful tool. It can help create a very powerful automated response to customers by using very smart chatbots—smart if they are trained well. It can help to support your customer service agents by automating many of their manual tasks and providing each of them with a digital assistant. It can also help increase productivity across the entire business—just look at how all Microsoft 365 business tools are now fully integrated with AI assistants.

However, as Phil Fersht suggested, we are still at the point where AI is a very smart tool that needs to be applied to very specific tasks. This means that humans are still very much required to remain in the loop. All those 2023 media headlines about replacing all customer service with AI were premature.

How AI and Human Agents Work Together in Customer Service

Any modern CX strategy will require AI integration. Businesses should:

1. Use AI as first-line support – Automated systems handle routine inquiries.

2. Enable seamless human hand-offs – Ensure an easy transition when AI cannot resolve an issue.

3. Support customer service agents with AI tools – AI should enhance human efficiency, not replace it.

The CRM company Salesforce is a good example of this. They claim that their own AI software is handling 8 out of 10 customer questions. This is impressive, but it still means that 2 out of 10 customers need to engage with a human—the system is not handling 100% of customer engagement yet.

The Future of AI-Powered Customer Support

No matter how smart your AI customer support system is, we are still in a situation where 20, 30, 40, or even 50% of customers may still need to engage with a human because the AI just can’t solve their problem.

The best way to design your AI-powered customer support system is to remember when and where you still need a human in the loop.

For more information about the agile yoummday approach to twenty-first-century sales and customer experience, please contact [email protected].