Enterprise Times participated in a roundtable discussion facilitated by AnywhereNow at the Savoy Hotel in central London. William Blench, CEO at AnywhereNow, was joined by Jon Burghart, Chief Revenue Officer and Jurgen Hekkink, Head of Product Marketing at AnywhereNow. The company provides customer experience solutions driven by Enterprise Dialogue Management. AnywhereNow supports customer engagement with efficient workforce collaboration, AI-driven insights, and a full omnichannel service experience. It leverages Teams and the entire Microsoft ecosystem, as well as a multitude of CRM integrations.
The roundtable discussions flowed on the challenges and benefits of AI adoption, customer experience, and its impact on business processes, as well as AI as an assistant to human agents and the future of AI in business.
Blench started the discussion, outlining the historical context of AnywhereNow. Its origins with Microsoft, and its transition from traditional telephony to modern communication systems. He emphasised the importance of AI in customer service, particularly in contact centres and service desks, to make agents’ lives easier.
As an aside, AnywhereNow was previously known as Anywhere365.
What challenges do customer service teams face while attempting to embrace AI?
The AnywhereNow executives outlined a range of challenges for businesses.
Data quality and integration
Poor data quality can undermine AI effectiveness in addition to increased difficulty integrating data across different organisational silos. Poor quality data can also lead to a lack of a 360-degree view of customer interactions. It can increase the challenge in connecting front-office and back-office systems.
Technological adoption barriers
There’s been plenty of hype on AI and machine learning in the media, which can generate resistance from employees fearful of AI replacing their jobs. As a result, enterprises and organisations need to prepare comprehensive training for employees on new technologies. Businesses are also increasingly becoming aware of the issue of “Shadow AI”—the non-company-issued AI tools used by employees. According to Hekkink, “It may be better for businesses to teach their employees how to use AI responsibly than to let them just go it alone.”
Organisational readiness
Burghart believes there is a limited understanding of how Businesses can effectively implement AI. Furthermore, there was a lack of clear business processes to support new technologies and difficulty measuring ROI. Employees fear technology replacing their jobs, which can lead to cultural resistance to technological change.
Technical challenges
The AnywhereNow executives stressed the importance of businesses managing AI hallucinations and inaccurate information. Hallucinations refer to instances where a model generates inaccurate or misleading information. Due to limitations in training data or inherent biases, this information is often presented as fact. As a result, organisations need to select appropriate AI models that meet specific business needs.
Customer experience considerations
Maintaining human touch while introducing AI will be vital for organisations. Enterprises have to ensure AI can handle complex, nuanced interactions, balancing possible cost reduction with quality of customer service. AnywhereNow’s overarching recommendation is to start small, focus on clear use cases, and involve employees in the process. Businesses can then gradually expand technological adoption with clear, measurable outcomes.
As Blench suggests, “We have the relationship with customers. However, we need to be aligned with their business objectives and help them scope out the business process and workflows that support the technology. What are the proof points that demonstrate that the technology works?”
Are businesses equipped to manage the orchestration?
An observation that arose from the discussion was the need for orchestration. Orchestration involves the coordination of multiple computer systems, applications and technologies to ensure AI deployment, configuration management, and other processes are performed in the proper sequence. Today’s businesses will need to support orchestration of platform integration. Companies need to connect multiple systems (CRM, ERPs, back-office, communication channels) and ensure data consistency across those platforms.
Burghart raised the issue of the complexity of orchestrating different AI and communication platforms. According to Burghart, “It’s all about building the AI model relevant to the business environment, where the AI is being deployed. Customers in a contact centre, sales or design will have very different models despite the technology being the same.”
Businesses need to create an orchestration framework that links different AI components using specialised AI agents for specific tasks (Verification, knowledge search, routing, etc.). Moreover, organisations need to develop modular AI systems that can interact seamlessly.
Key Orchestration Components include:
- Identification agent for customer verification.
- Knowledge agent for information retrieval.
- Routing agent to direct interactions to appropriate channels.
- Sentiment analysis agent for real-time interaction monitoring.
- Maintaining security and compliance frameworks.
- Selecting compatible AI models.
How should businesses measure the impact of AI on key customer service metrics and investments?
Many businesses still consider the call centre as a cost burden. However, AnywhereNow execs believe operators need to consider them as an investment using different metrics. These could include:
- First call resolution rates
- Customer satisfaction scores (CSAT)
- Net promoter scores
- Agent productivity improvements
- Reduction in call handling times
Hekkink says businesses can measure the impact of their AI investment.
Business Outcome Measurements could include cost reductions in customer service operations and the increased efficiency in routing customer queries. The improved quality of customer interactions could also lead to upsell opportunities generated for the business.
What is the best way to start implementing AI?
Start with small, low-risk AI deployments and measure specific outcomes for each use case. AnywhereNow stresses the importance of building trust through demonstrable improvements, then gradually expanding AI implementation across more complex scenarios.
Businesses need to develop specific measurement approaches that track agent performance before and after AI implementation. They also need to monitor sentiment analysis of customer interactions and evaluate time saved on mundane tasks.
Hekkink says, “The success of AI deployment is dependent on the support provided to stakeholders, the human agents and supervisors. The human factor is important. It’s about how AI is supporting employees and enhancing their experience. During our experiences with customers such as DHL, we ensured employees were involved in those processes and actually improves their jobs. Agents are better served by AI and are now getting the information directly to their desktops. Previously, they had to waste time searching for that same information. As a result, AI can actually help agents to have a more fulfilling role.”
AnywhereNow advises businesses to have clear, measurable SMART objectives that align AI technology with specific business goals. This ensures that investments can be directly tied to tangible improvements in customer service performance.
Is AI an assistant to human agents or replacing those agents in the contact centre?
This question is related to how autonomous AI can be and its success factor, which can be 98-99% more efficient. Currently Hekkink says the best way to deploy AI, is as an assistant to the human agent. However, businesses need to understand conversational AI involving simple interactions, conversational-transactional environment, previously handled by IVR systems, could be AI-enabled.
A customer, in normal language, says what they need. AI can then interpret the request and intelligently direct it to the appropriate agent. This is very low risk because AI isn’t making any decisions but helping to optimise the process, getting the enquiry to the best possible agent straight away.
Hekkink expects AI models to mature further and get trained on more standard human interactions. Eventually, these interactions will be replaced with autonomous virtual agents. Possibly in the next 5-10 years, around 40-60% of all interactions will be autonomous and handled by AI virtual agents.
However Hekkink believes organisations will remove all human agents. “Humans like to speak to other humans because of the empathy factor, and unexpected questions that often arise. People do not entirely trust AI services, so I don’t think the human action will completely disappear.”
