All articles

AI at the Unified Comms and Contact Centre Expo


Tim Russell

Modern Workspace

•  Jan 26, 2024

Did you spot the contact centre vendor without AI? 

Recently I attended the Unified Communications and Contact Centre Expo at the ExCeL centre in London. It was probably not surprising that everyone there had an AI play, and everyone’s version was, according to them, the right one for you. I’m not going to go down into the weeds about the differences, as this is a case of finding the best solution to your individual needs. However, there are several high-level things to consider when we look at the development of the contact centre.  

Let’s start with a few basics on AI integrations, not just to contact centres, but customer interactions in general and, most importantly, why. 

Customer Interactions

    1. Website customer interaction with a chat bot

This is the most obvious interaction. Having a chat now bot that website visitors can utilise allows a more directed approach to customers just browsing websites. There is a balance to be found in either using it as a lock in, or genuine reason to provide support to the end user. Think of the data that could be used by a correctly implemented chat bot: cookie data, pages browsed, referring URL etc. All of this can help a chat bot provide contextual information to a user helping them on their journey. 

    2. Voice Interaction 

At some point we have all interacted with a voice menu system; press 1 for this or use simple phrases to request a service. Over the years I have worked with many contact centres and seen the continuing development of these interactive menus. In some cases, massive spreadsheets were used to map out possible spoken scenarios and the resulting call route. With the advancements of AI and Natural language it is possible to take multi-lingual, free text that can be evaluated by an AI to provide a logical outcome, by knowing what services an organisation delivers and mapping it to the ask. AI even has the potential to ask clarification questions, again, in a natural language pattern and not using pre-formatted answers. The frustration of previous systems leaving you feeling unable to find the correct sentence to get what you need will become a thing of the past. 

Have you ever heard “We are experiencing more calls than usual” every time you call a contact centre? You should only hear that when there is an exceptional delay, not every time you call. With AI, these messages can be modified and played only in specific scenarios, not only when they could suggest better times to call back. This could be applied in much the same way cloud-based navigation solutions reroute users via multiple routes to avoid congestion. The same AI capability could look at information from external sources, such as news sites, digital employee experience platforms and any other relevant information to provide contextual menus to callers, who can then consider the information provided by these information feeds. 

Agent Enablement

Although the latest information coming out of the EU’s new AI Act will specify some notification requirements around the use of these features, they will make a significant difference to the ability of agents to be onboarded quicker and be provided with information. This will allow them to be far more informed and empowered during a customer interaction.  

  • This first one has been around for a while, but I believe AI in the contact centre will deliver an amazingly powerful, albeit potentially divisive function in the form of sentiment analysis. During a call the AI could be a silent partner, listening to voice tones, use of words and volume levels of both agent and customer. This sentiment analysis can lead to reducing an escalation in tensions during a call, or frustration on both sides of the interaction; we are after all humans and driven by emotions. 

Technically, this capability could also be used in AI triaging, allowing agents to provide a more personal service once the AI interaction has removed as many interactions as it is able to resolve from the queue. The spin-off from this capability has again been used for some time, but the capabilities are increasing exponentially, such as fraud and anomaly detection. Forget the “My Voice is my password” security measure, now just your tonal inflection and choice of vocabulary could identify you, even without your calling ID or geographical location. Imagine not having to remember your first dog’s name, half your postcode, and the inside leg measurement you gave when you signed up to a service. You are you, you know you are you, so why should you have to give specific details; being you should be enough shouldn’t it? 

  • The next feature is the ability of AI to not only present systems to an agent at the appropriate time, but to prompt suitable responses for the agent to use. Not pre-formatted text as such, but a generated response that reflects the agent’s tone, the ask of the customer and a myriad of other information. This prompting can also be fed back into the AI’s foundation model allowing it to respond in the triage stage of a call and release the agent back to the more complicated scenarios that we humans are best suited for. 
  • Agent onboarding and quality control are the last two scenarios where I see real benefit. Face-to-face or virtual onboarding through desk sharing has been used for some time. Shared vision through XR or even fully immersive AR are options, although I believe for basic contact centre use these could potentially not be financially viable. Where AI can help is knowing that when an agent starts interacting with either calls or textual interactions, such as chat, it can provide a guide and a control of the responses provided. This repetitive prompting and support of an agent can instil confidence and drive them to be more productive within the contact centre environment, far quicker than was previously expected. It could also be possible to use AI as a multi-system interface; instead of having to know how to operate several different internal systems, AI can prompt and provide relevant information to an agent, summarised and presented contextually, which allows them to support the customer better. 

MicrosoftTeams-image (57)-3

Interaction Analysis

Penultimately I’d like to look at the complete interaction analysis. The stats I have read recently indicate that just 2% of data globally is analysed. However, in contact centres I know this is higher. If all recorded calls, performance stats, and other relevant information was analysed, AI could spot anomalies and offer suggestions for performance improvements. I’ll share a few of the scenarios I can imagine, but it is quite honestly as varied as the scenarios you can possibly think of. 

  • QA Analysis of a call provides insight into agent training potential and what, if any, training the AI would advise.  
  • Looking at end-to-end customer experience, AI could propose process improvements that allow for a more efficient and productive customer engagement. 
  • Already in use in a few security packages and probably in a few insurances practices; anomaly and fraud detection. I don’t think I need to explain this too far, but loss prevention is very real, especially in the insurance business. 

Finally, I’d like to touch on the high-level concepts of where AI could further be used in the contact centre environment. 

  • Agent tasks such as post call processing, information logging and data updates. 
  • AI will be able to provide a general situational awareness to both staff, supervisors, and customers to ensure expectation, reporting and analysis match the entire scenario. 
  • There are already several tools to help monitor employee health and wellbeing in the workplace, AI in a contact centre, combined with digital employee experience can take this to the highest level of employee awareness and protection possible. 

With environmental monitors, application responsiveness, voice print analysis and all the other potential data feeds, supervisors will know an employee’s state of mind even without being able to see them. 

  • The last scenario is one I will be writing a sperate article on, but it is the ability to utilise AI personas to create a kiosk service for citizens, which can reflect preferences for language, gender interaction and environmental considerations. All of these can lead to more relaxed customer engagement. Watch out for that article coming out in the future.


In this article, I have explored some of the ways that AI can enhance the customer experience and the employee wellbeing in contact centres. I have also highlighted some of the challenges and opportunities that AI poses for service delivery and the human factors involved. I hope you have found this article informative and insightful, and I invite you to subscribe to our updates, and reach out to me if you want to have a further conversation about how CDW can help you leverage and plan for AI capabilities in your contact centre strategy.

Subscribe to email updates
In this article