Avoiding the dangers of AI misuse

Reggie Scales at Vonage describes how misusing AI is undermining organisations’ communications strategies (and how to fix it) 

 

Businesses across the UK are betting big on AI. Research found UK businesses spent 449% more on AI solutions in 2024 than they did the previous year, and McKinsey data shows 78% of organisations are using AI in at least one business function.

 

What’s more, Vonage research finds customers are willing and able to use these AI tools: 83% have used an AI tool recently. Yet there’s a disconnect between AI investment and customer satisfaction. Despite huge advances and increasing uptake of AI technologies, the vast majority of customers still don’t feel customer service has improved in the past year.

 

Clearly there’s a gap between businesses’ lofty ambitions for AI, and the reality of customers’ everyday challenges.

 

Organisations need to strike a middle ground, finding a way to embrace innovation and trial new AI solutions while ensuring they have a real impact on customer experience. To do this, they need to start by identifying the pain points customers are facing, rather than trying to reinvent the wheel.

 

 

Customer appetite for AI solutions

As everyday AI tools like Gemini, ChatGPT, and Claude.ai become more commonplace, consumers are becoming more comfortable interacting with AI. 

 

A third (32%) of customers plan to increase chatbot usage in the next 6-12 months and 28% expect to increase AI-enhanced video chat usage. Around a quarter expect to increase their use of voice personal assistants, automated phone support, and in-app calling. Gone are the days where customers feel ‘fobbed off’ by automated service options.

 

Customers are increasingly willing to use AI, but - rightly so - they do have some criteria for AI interactions. Almost half (47%) say they want to see proactive issue resolution, and 45% want more personalised experiences.

 

 

Customer expectations vs reality

Despite organisations’ willingness to spend big on new AI solutions and customers’ openness to them, the reality is customers are still largely unhappy with the customer service they’re receiving from businesses. Something isn’t adding up.

 

65% of customers complain about long wait times to speak with agents, the same number are unhappy with having to contact customer service multiple times, and 63% cite having to describe their problem to multiple agents - leaving them stuck in a ‘doom loop’, being passed around from agent to agent without resolution. 

 

Negative customer experiences can be fatal for organisations, especially fledgling brands that don’t have the advantage of long-established brand credibility. Three quarters of customers will stop buying from a business after repeated communication problems: no matter how great a business’ product is, communication is king.

 

Poor experiences are contagious: 42% of unhappy customers will tell friends and family if they’ve had a negative experience with a business. A fifth will take their negative reviews to social media.

 

AI has the potential to bridge the gap between poor experiences and excellent ones. So where are businesses going wrong?

 

 

The right approach to AI: layering

The biggest mistake businesses are making is applying ‘AI for AI’s sake’, in other words creating new AI projects and solutions that aren’t based on the specific challenges the organisation (or their customer) is facing. A better approach is to take a comprehensive look at the organisation’s pain points - as well as the pain points being passed along to customers - and working backwards from there. 

 

For instance, one of the main complaints from customers is long wait times. This could be easily improved with the right AI implementation: organisations can use an AI agent to triage calls and direct non-urgent or simple queries to another channel, like message or email.

 

Another common complaint is being transferred across channels and multiple touchpoints in order to reach a resolution. The doom loop challenge can be caused by a heavy-handed approach to AI, where AI agents and chatbots have completely replaced human employees and haven’t been adequately trained in how to handle enquiries or reach a resolution.

 

This kind of pitfall doesn’t do AI agents justice. Agentic AI has the power to assist human employees with ‘next best action’ prompts and suggest relevant knowledge base articles, which can be particularly helpful if a new employee is still getting up to speed. But without humans in the loop, or sufficient training for the AI, it’s almost impossible for AI to provide the proactive resolutions customers are looking for.

 

Contrary to popular belief, AI can be used to make experiences much more tailored, meeting customer demand for personalised experiences. AI can pull up a customer’s entire history with the organisation to share while a human agent is interacting with the customer, enabling the agent to tailor their responses and outcome to the individual customer.

 

The key is layering AI over existing communication channels, and not removing human interaction entirely.

 

 

Putting layering into action

Organisations should start where customers already are: the vast majority of customers choose to interact with organisations via messaging or voice calls, so this is a good place to start.

 

Organisations can take a three-step approach to layering. First, they should assess the most common pain points on their high-volume channels, then determine how they can use AI to address these specific challenges without disrupting familiar touchpoints. Once they know it’s working, they can then roll out these AI capabilities across all channels.

 

With a strategic, layered approach to AI in communications, organisations can achieve the high-quality service that customers expect across every channel and convert transient customers to loyal, returning ones.

 

---

 

Reggie Scales is Head of Applications at Vonage

 

Main image courtesy of iStockPhoto.com and Vertigo3d