Deploying AI in 2023

Mark is a VP of contact Center operations who has had a tough beginning to 2023.  His fear stems from the impending budget cuts that he anticipates. As he manages the convergence of three key issues, namely: 1) postponing the migration to cloud infrastructure, 2) the persistent impact of the pandemic leading to agent turnover, and 3) a surge in call volume due to business growth. Consequently, his metrics for handling times, First Call Resolution (FCR), and Net Promoter Score (NPS) continue to drop.

Mark finds himself bombarded by various vendors seeking to aid him in integrating AI into the call center. Their proposals range from enhancing agent productivity to implementing a new AI-powered Cloud Contact Center as a Service (CCaaS) system. Hoping for a more streamlined solution for managing call volumes, Mark has eventually arrived at a decision for 2023 – focusing on deploying an AI virtual agent. This agent would handle multiple call types on the existing infrastructure, thus reducing the load on human agents and improving FCR.

Personally, I am excited about this development, as Mark chose to deploy Avaamo’s Contact Center AI to decrease call volumes. However, Mark’s predicament mirrors the challenges faced by many Customer Experience (CX) leaders during this year of uncertainty. The ongoing concerns about budget cuts, combined with escalating call volumes, have instigated a desire for an approach that leverages AI to boost capacity, lower call volumes, and ultimately enhance customer experiences.

The AI Dilemma – Where to Begin?

The integration of AI into contact centers has generated significant interest among CX leaders in 2023. As they grapple with cost-reduction amid economic uncertainties, traditional Contact Center vendors endeavor to protect their existing on-premise systems by promoting “bolt-on” tools that emphasize an AI narrative. These tools cover several fronts, including:

1) Reducing costs through Cloud Migration
2) Enhancing personalized services
3) Improving transcription and learning tools
4) Facilitating better agent assistance

For CX leaders, the dilemma is deciding where to start. What should be the priority?

Prioritizing the Challenges: Where Should We Focus?

1) Cloud – The CCaaS Challenge

CCaaS has promised so much. Until now, however, this has proven tricky. Indeed, only 29.5 percent of companies globally have switched to CCaaS according to 2023 Metrigy research. The tendency to compare large on-premise highly customized systems with functionality from cloud providers has created a challenge for CX leaders, Should I do this or wait? All the while, operational costs have increased. After all, the annual cloud payment typically adds to long-term contact center expenditure that affects extremely sensitive Opex budgets. CCaaS vendors charging in terms of headcount/seats, have little incentive to help businesses lower their contact volumes to counteract this. Other issues include handling tricky migration loads built on old systems, avoiding cloud vendor locking, and ensuring carrier autonomy.

2) Agent Turnover – Insufficient Staff?

A fundamental narrative of Contact Center Operations: we simply don’t enough agents online. The number of agents you have on staff is your magic number. You don’t need a huge team, but you do need to be smart with staffing & training to make them productive with all kinds of software. Only then can you can comfortably handle the volume of calls and messages coming through. Indeed, a veritable multi-billion dollar industry of vendors, consultants, and systems integrators push and promote this narrative. Vendors promise better Agent assist tools/workforce management, analytics dashboard, teaching and training tools all targeting agent productivity as the nirvana every CX leader should aspire for. Now AI and now Generative AI has muddied the mix.  The focus is usually to use AI to further optimize agent productivity in these general categories:

•  Transcription
•  Agent Assist
•  Scheduling
•  Coaching

Of course, it’s supposed be adding additional value to make agents more productive. Yet, from the outside looking in, not all that much has changed in the user experience.  It still your old IVR systems, Still, “press one for this, press two for that” IVRs confuse and disappoint customers.   Almost always, customers have to repeat themselves because customer context fails to follow them.  Then, they must wait on hold as agents shift between disparate systems, struggling to locate the necessary information to help them.

3) Self-Service Solutions

CX leaders have tentatively embraced self-service experiences by deploying simple FAQ bots or supplementary tools from chat and IVR vendors, typically aimed at addressing straightforward, repetitive queries. For more intricate workflow queries, assistance is routed to agents or humans. Recent advancements in AI and automation, however, provide an unprecedented boost in productivity through end-to-end automation of complex queries without human intervention. This is the essence of self-service automation.

Slaying the dragon with self-service automation

Mark explains this well to his team as he lets me in on an internal meeting where the debate continues: “The traditional way people think about AI now is to walk through a company’s workflow, identify tasks that a machine can help the agent, and automate them. But ultimately the upside is pretty limited because the best you can do is what the agent already doing but a little bit better; those typically don’t justify the huge expense adding several bolt on tools. The real money, then, is in totally blowing up the workflow and replacing it with AI, with end-to-end automation without human intervention.”  

I believe he is referring to sophisticated voice-based virtual agents that cut across all these problems: automate workflows and queries, allowing for the human element without the human touch. The explosive interest in ChatGPT underlines this. Whether it’s invoice processing, scheduling calls, or more complex prescription refills in healthcare, voice based IVAs (intelligent virtual assistants) from companies like Avaamo can slay the call volume dragon in one stroke. Here is a good example:

5 reasons Self Service automation works

1) Can scale to millions of calls solving the agent capacity problem
2) They can be trained in weeks using existing contact center data, as opposed to the time lag of months for human agents
3) They can be immediate an overlay over on premise/ cloud or custom systems without the need to upgrade
4) They can be integrated to existing CRM systems to extract data, inform, and resolve user questions in most cases without additional human intervention.
5) Offer a customers experience that delivers sophisticated self-service customer support rivalling human agent.

A recent Avaamo customer deploying a virtual agent on existing call center infrastructure reported a 60% deflection of support calls without human intervention. You do the math on the CPC (cost per call savings)

Agent productivity Vs No human Intervention.

The temptation when questioning the AI infusion is to seek out like-for-like systems and processes and how they can be optimized. However, this often adds layers of complexity. After all, legacy contact centers are highly customized, with many bolt-ons, making up an intricate patchwork quilt of systems that includes older capabilities, perhaps no longer necessary to the service experience.  A conversational AI approach given the advances in AI, offers the chance for a more streamlined approach, where leaders isolate the mission critical part of the customer experience — the actual customer journey — and by using a virtual agent can deliver that experience in weeks with no change in current infrastructure. Such an approach chops down contact volumes and frees up resources enabling a more proactive service operation. These capabilities are all out of scope for a traditional player who is obsessed with protecting their seat license. Even for the CRM players, it’s just not how they tend to think. CX leaders must take the blinders off: lowering call volumes is achieved best by using an hybrid approach of using the best AI technology to lower call volumes and freeing up existing agent headcount to focus on the “must have” human interactions only.