Back to basics: the secret to successful AI innovation in contact centres

Now is an exciting time to be working with contact centres, with so many new tools and technologies to support agents and improve the customer experience. And, while I love the energy people are bringing to discussions around AI, there are still plenty of misconceptions and missed opportunities for meaningful change.

As someone with a great appreciation of shiny, new tools (perhaps more than most), I also firmly believe that AI success depends on getting the basics right. In this post, I’ll share some practical ‘dos and don’ts’ for implementing AI effectively. We’ll start with the fundamentals and then explore more advanced ideas in my next blog, so keep an eye out for that.

1. Solve real business problems, don’t chase buzzwords

Old-school rule: start with the need, pain point or the opportunity you want to address. And don’t assume AI is the answer to everything. Buzzwords don’t solve business problems and often pre-empt meaningful discussion. Properly integrated tools working together help people solve problems.

For example, clients have come to us requesting a RAG (Retrieval Augmented Generation), which is basically AI applied to search your documents. That seems positive and insightful, as often clients wouldn’t use the technical terminology. Sounds like they know what they need. But if we let the ideation end there, we may miss the opportunity for a fuller solution that really helps save employees’ time. While RAG might be part of the answer, the broader solution could involve streamlining workflows, improving knowledge sharing, or automating related tasks. By digging deeper, we can design solutions that genuinely save employees’ time and deliver measurable value. 

2. Be specific

To deliver actionable insights, AI needs access to relevant, contextual information. Providing relevant documents to the RAG model mentioned above can provide company or departmental-specific information to be prioritised but try to respond to information relevant to the individual and their particular circumstances. This could include internal data or external sources like weather or maps. 

As with humans, the more information AI is given about your situation, factors impacting it and possible remedies, the more likely it is to provide the best solution or response. The better the data, the better the results.

We should also remember; AI is much more than ChatGPT. It is an umbrella term covering diverse technologies, from generative AI chatbots (like ChatGPT) to machine learning models used for data analysis, natural language processing for understanding and generating human language, and computer vision for interpreting visual data.

3. Focus on action, not just answers

Remember that a question is usually an early step in a decision to take action. AI should help people achieve goals, not just answer questions. In the world of generative AI, it’s easier than ever to shorten the time between insight and action by augmenting interactions.

What automations can you create, or business processes can you augment to solve the underlying problem or achieve their goal faster?

This is where we now introduce you to AI Agents. They are like digital assistants designed to complete tasks or solve problems using data and smart decision-making and they come in many forms. On a basic level:

• Interactive AI agents, such as chatbots, engage directly with people to answer questions or guide them through processes.
• Automation-focused agents work behind the scenes, performing tasks or making decisions before handing off to a human, similar to robotic process automation (RPA) but with a degree of decision-making and other capabilities.

In contact centres, AI agents can complement tools like RAG models by helping customer service staff take follow-up actions efficiently. For example, they might automate summarisation tasks, allowing human agents to quickly locate critical information, or handle repetitive processes, freeing up time for more complex interactions.

By reducing workload and improving response times, AI agents can enhance both customer satisfaction and employee experience.

4. Don’t wait for the perfect solution - experiment with purpose 

Incorporating AI into the way we work requires changes that take time to embed, so use pilot projects to identify what works. Learn from taking action, refine your approach, and don’t be afraid to adjust as needed.

5. Set clear KPIs and track success

Experimentation doesn’t mean being sloppy or scattergun. People seem to forget that KPIs (Key Performance Indicators) still apply to AI. Targeted experimentation often reveals insights that lead to better adoption and measurable results. 

Make sure your planned AI project supports the business strategy and that you are measuring the savings or impacts of the improvements you’re trying to make. Without clear metrics, projects risk being shelved or failing to deliver meaningful outcomes. 

Take calculated risks, evaluate the outcomes, and course-correct where necessary. 

Define what success looks like upfront. For example, a KPI could be reducing average handling time (AHT) by 20% within six months or achieving a 15% improvement in first-call resolution rates. 

6. Think beyond speeding up processes

While faster processes are valuable, AI’s true potential lies in differentiation. Ask what would set your business apart from the competition, then explore how AI can enable those changes.

For instance, generative AI can clean and integrate unstructured data, enabling insights from previously disconnected sources that just weren’t feasible not that long ago. The action on those insights could be proactive, personalised suggestions to customers written in their preferred language. Or it could mean AI recognising a possible system issue early on and prioritising that work for your engineers/developers to get ahead of the problem, preventing a wave of calls from frustrated customers, all while pushing proactive information to affected customers to smooth things, keep them informed, and prevent further calls and complaints. 

Turning insights into impact

The possibilities for AI in contact centres are expanding rapidly. From automating routine tasks to delivering hyper-personalised customer experiences, AI can transform how businesses operate.

Want to learn more? If you would like to discover how AI can help your contact centre deliver faster resolutions, improve customer satisfaction scores, and empower agents to perform at their best, please get in touch or connect on LinkedIn.