The growth of digital services over the last 20yrs has been steadily changing the nature of contact centres, to the point they now tend to deal primarily with “failure demand” essentially where a customer has been unable or unwilling to engage with a process, obligation or need in the “digital” channel.

Many organisations are exploring ways to deal with this increasing demand, whilst coping with digital transformation benefits cases that reduce their resourcing. The most promising approach is obviously being able to deal with simple or common customer questions through automated systems. This has led us to intelligent virtual assistants (IVAs), AI powered software that can interact with customers through natural language, either by voice or text. Using technology like this helps to understand customer intent, provide relevant information, and perform tasks based on user inputs.

IVAs are not the same as chatbots, which are rule-based systems that handle predefined scenarios and inputs. IVAs use advanced technologies such as natural language processing (NLP), machine learning (ML), and speech recognition to learn from customer interactions and provide more personalised and accurate responses, this is the key differentiation that bestows the term “AI Powered” on them.

However, it’s important to think about the use of IVAs and associated technology as a continuum, in that the more they are used, by definition the more they are optimised and therefore the more they can grow into more sophisticated use cases.

The key as ever is to start small and grow, perhaps through discrete areas of business that can be optionally directed from call queues.

This has the potential to lead to a broader “Autonomous Customer Service”, which Gartner define as a type of customer service and support using AI and Automation to “provide proactive, personalised, and predictive service experiences across multiple channels without human intervention”.

This builds on self-service and assisted service, but goes further with a more accessible, intelligent, and dynamic approach, integrating and then blurring the boundary between digital first services and the pathways to get support when issues arise, e.g.: -

• Customers expect fast, convenient, and consistent multi-channel service. Meeting these expectations by providing 24/7, instant responses, and tailored outcomes is essential. Adopting IVA related technology can achieve this and reduce frustrations with long wait times, repetitive questions, and hand-off’s to other agents.

• Reducing headcount in contact centres calls for greater agent productivity and performance. IVAs can handle a large volume of customer queries without compromising quality or accuracy, which frees up human time and resources to focus on more complex or high-value tasks requiring empathy.

• Reducing operational costs by enhancing agent onboarding, training, and retention through more intuitive and digestible access to knowledge, as well as reducing infrastructure and maintenance costs associated with workplace provisioning for human agents.

However, don’t be Fooled into Thinking about IVA as Single Component

For a start, its essential to think about this in a few dimensions and in that regard, Gartner categories are always useful. When it comes to AI/ ML there is obviously a substantial list where those acronyms occur, but three categories stand out as particularly relevant: -

1. Enterprise Conversational AI Platforms:These are software platforms enabling the development and deployment of IVAs for various use cases across different channels.

2. Data Science and Machine Learning Platforms:These are software platforms that provide data scientists and other users with tools and capabilities to build, deploy, and manage data and machine learning projects.

3. AI Developer Services: These are cloud-based services that provide developers with access to AI capabilities such as computer vision, natural language processing, speech recognition, and generation.

The point is, for a large organisation, it’s important to think about supporting IVAs from the widest cross section of the business.

It’s no good just adopting AI/ML/NLP/ GenAI offerings that are part of narrow business applications, limited to specific areas.

You need to think at an enterprise level, choosing solutions that can encompass the full organisation providing a wide portfolio of tools and software components to solve cross-cutting business challenges across multiple use cases, with access and interoperability throughout the wider IT Landscape.

The choice of the right combination of AI/ML solutions to support IVAs and resolve complex problems remains an ongoing challenge, hampering successful application of AI in operational settings.

However, products are emerging that provide enterprise scope and capabilities that map back across the range of Gartner categories above, e.g.: -

1. - The ability to quickly generate and vary machine learning models to support rapid adoption into end-user context, including safe and economic trial and error approaches.

2. - Being able to execute workflow, customer and suspicious behaviour analysis, uncovering important causal relationships from large data sets to derive new insight into AI/ ML Models.

3. - Ensuring reliability and ethical development and use of AI, by being able to present reasons for AI Judgements and adopt consequent learning back into AI/ML Models.

4. - Providing a means to verify conversational generative AI output, using large business data sets, in a safe and secure testing environment.

Conclusion 

Intelligent virtual assistants are not just a trend or gimmick, they are a powerful tool that can transform contact centre operations and customer service outcomes. By adopting IVAs contact centres can better meet demand, enhance customer satisfaction, boost agent productivity and performance, and reduce operational costs.

However, choosing AI/ML products and the delivery models that support them solely grounded in Contact Centre use cases or in singular Contact Management Applications, fails to take advantage of data, process, and insight from the rest of the enterprise and conversely fails to turn contact demand into wider enterprise value.