AI and CX Strategy in 2024: Event recap

Published 09/04/24 under:

With a noticeable shift in customer propensity to self-service, how can contact centres make digital channels and AI-powered solutions count, while bringing employees along on the ride? Set within the historic Tower of London, the March Kerv event was standing-room only with an eager audience of CX thought leaders, practitioners, and technologists keen to solve this puzzle.  

The art of triage: Balancing AI efficiency with human empathy

Kerv Senior CX Consultant, Paul Cox, set the scene by examining latest research findings and emerging trends. With the UK minimum wage (and agent costs) continuing to rise, AI advances, open APIs and web platforms have made self-service far easier to deploy and enhanced the experience. 75% of customers see it as the most convenient option[1] with a whopping 91% happy to use an online knowledge base[2].  

The number two ‘customer service wants’ are quick resolution (91%) and answer speed (83%) closely followed by empathy (78%) in third[3]. The big takeaway is that automation must be carefully tempered with human understanding.  

As illustrated above, the answer lies in effective triage that accurately matches experience to customer needs – at each precise moment in time – reducing cost while also growing loyalty. 

Futureproofing your CX 

Standing by to help with triage and everything digital, Genesys VP UK and Ireland, EJ Cay, underlined Genesys’ commitment as long-time leaders, investors, and innovators in AI. With personalised experiences becoming harder to deliver, the Genesys product roadmap is geared around delighting ultra-demanding, time-starved customers. That means: 

• Knowing who they are, how they got here and what they need. 
• Not wasting their time and making them repeat information. 
• Understanding how they feel – and acting accordingly. 

AI-powered solutions address these issues, for instance, by automatically detecting customer intent, offering assistance if they are struggling, or re-routing them to agents highly-skilled in handling complex and emotional situations.  

Digital innovation in action

Skipton Building Society

Bringing things to life with practical hands-on experience, Stephen McNamara, Head of Digital Commercialisation shared his team’s transformational journey. Kerv Experience derisked and accelerated the cloud migration with automation tools, meeting the company’s aggressive nine-month go-live date.  

The new Genesys Cloud platform enables 1,700-plus users to manage phone, email, and chat conversations from a standard desktop, in the process rationalising 1,768 configured DDIs with 418 queues and 382 call flows. Now, with asynchronous Genesys Web Messaging available on its website, Skipton advisors can have three or four interactions on the go at the same time. And, unlike traditional webchat, busy customers can start a conversation, leave to take care of other tasks, and then return to pick up where they left off.  

Within a few months, Skipton began to see improved CX with higher first contact resolution and lower average handling time. Previously impossible innovations are now firmly within reach. Future plans include Genesys integration with Microsoft Dynamics, automating the ID&V process with voice biometrics, and building out consumer duty capabilities through generative AI. The full case study will shortly be available on the Kerv Experience website. 

Utility Warehouse

Next, Product Lead, Kevin Mythen, explained how Utility Warehouse (UW) was leveraging AI to drive better and faster CX. UW enables customers to save money by consolidating their energy, mobile, broadband and insurance services with one bill and one direct debit. However, complex skillsets are needed to manage bespoke product bundles and satisfy not just one, but multiple regulators.  

The company implemented Genesys Cloud back in 2019, initially to drive high quality growth and create a ‘never want to leave’ culture that optimised knowledge retention. It was also looking to reduce the need for customers to make contact and streamline processes through self-service. 

Since then, working with Kerv Experience, UW has gradually built out digital and AI capabilities (see timeline below). Now, agents handle 3.2 million queries annually – 28% less than before, with self-service channels accounting for around one million meter readings and 140,000 payments. 

Leveraging Genesys Generative AI and Large Language Models (LLMs), auto-summarisation means agents spend less time typing and editing notes from customer conversations, saving 70 seconds on average per call. In addition, UW has seen a: 

• 7% improvement in after call work. 
• 18% reduction in hold time. 
• 12-point improvement in customer effort score (CES). 
• 96% compliance rate. 
• 56% increase in account note entries (welcomed by UW Legal teams). 

Offering channel swaps, initially from calls to WhatsApp (and shortly to bots) has seen CES improve further to 90-plus. New initiatives this year include enabling ID&V within WhatsApp (in test). Also planned is a video assist service. Initiated by text, the idea is to enable agents to quickly visualise issues, instead of customers describing them over the phone, tricky conversations that can take 45 minutes.  

Tech demonstrations

Kerv specialists presented various demos showcasing the following Genesys, Microsoft and Kerv solutions: 

• Predictive Engagement for website journey tracking. 
• Answer-providing Knowledge Bots, reducing contacts and transfers. 
• Microsoft Dynamics integration, increasing personalisation while lowering handling time. 
• Agent Assist, improving speed, quality, and consistency of agent responses. 
• CX Translate, reducing wait time and costs hiring agents with specialist language skills. 
• Quality Assurance, improving targeted coaching and compliance adherence. 

It was clear that GenAI solutions are now ably supporting a host of use cases. From understanding intent and text (Natural Language Processing) to transcribing speech (both real time and post-call), language translation (CX Translate), and generating insights and next-best actions. 

Service excellence does not happen by accident 

Dan Wright, one of several Kerv MDs, underlined the value that a professional change management partner can bring to AI and digital engagements. 70% of transformation projects fall short of their objectives[4] with 62% classed as unsuccessful[5] often due to clients not being honest about inhouse resource and knowledge limitations. Other common pitfalls that Kerv encounters and helps with include unclear goals, resistance to change, accessing the right skills, and lack of ownership.  

Removing these barriers and achieving true service excellence requires a robust approach. One that enables organisations to get under the skin of customer journeys and embrace new tools and ways of working. Success might also be contingent upon cultural change and will almost certainly involve optimised and measurable processes. 

Assisted by Kerv, Harrow Council saw a 35% jump in customer satisfaction. Similarly, Serco increased self-service adoption by 67% and fulfilment by 19%. 

Panel discussion

Hosted by Julian Barrow, MD, Kerv Experience, our expert panel fielded a lively interactive Q&A session. With far too many points to list here, below are a brief selection: 

Q: Are UW running any initiatives to help reduce the knowledge/agent churn you talked about? 

A: From a technical perspective, we will try to nudge advisors in the right direction using solutions like co-pilot, agent assist, and so on. Our aim is to have shorter training cycles and more on the job learning, with the comfort of knowing they always have an AI assistant on hand.  

Q: How many different languages does CX translate support? 

A: 110 currently with new ones added all the time. A full list of languages can be found here. 

Q: How does MS Dynamics work when an advisor has sent an authorisation request to a manager who isn’t around/doesn’t reply and a customer is waiting? 

A: The approval process could be configured to go to a team of people where just one person would need to approve, or an option could be put in place to escalate further if the user is not available. 

Q: How can we contain hallucinations in Generative AI? 

A: Don’t use GenAI for chat responses. This sounds counter-intuitive, but LLMs are actually very good at the first part of any chatbot scenario, which is determining ‘intents’ (what does the customer want to do) and ‘entities’ (name, address, product, what’s the fault, etc). Just using LLMs for intent and entity extraction and continuing to use pre-authored responses could still lead to a more capable bot that can be more quickly deployed than using NLP tools. 

Q: What other advice would you offer? 

A: Use a CX specialist and specialised platform. Kerv has a wealth of experience implementing Genesys Conversational AI services and can be your guide to ensuring AI is a positive contributor to your customer and agent experience. Genesys also invests millions of dollars into training and fine-tuning AI models to assure accurate and high-quality responses.  

 

As the longest standing Genesys Cloud partner in EMEA with the most successful deployments, Kerv Experience continues to help clients accelerate improvements that boost customer and employee experience.  

To learn more, please get in touch. 

[1] 2023 Omnichannel Customer Service Index, Nuance.
[2] 2023 Coleman Parkes survey.
[3] 2023 Shep Hyken survey.
[4] Boston Consulting Group study
[5] Kerv Consult study of 100 IT and technology leaders (coming soon)