By
Andrew Neff
|
Date Published: November 15, 2024 - Last Updated November 15, 2024
|
Comments
Chapter 2: The Real Issues and Risks are Growing: 4 Key GenAI Weaknesses Tied Directly to Your Agents and these ‘Solves It All’ Solutions
Chapter 1 captured multiple, rapidly progressing changes and residual market potential. As the sector morphs quickly, it is critical to recognize the risks as well.
Four Weaknesses: The Risks and Possible Issues That Require Pre-emptive Planning Abound
#1: Increasing Costs vs. ROI
From Forbes in January 2024, “The cost of implementing AI in enterprise business processes may be more costly than we initially thought.” Gartner predicts “by 2025, growth in 90% of enterprise deployments of GenAI will slow as costs exceed value.” There are multiple reasons for costs increasing. New versions of GenAI and custom LLMs for business verticals are two. Another is that studies now show that GenAI is using more power and data centers are expanding exponentially. In July 2024, CNBC covered the massive needs for power and water, #concluding “the aging U.S. grid can’t handle the load.”
#2: Customers Are Losing Faith
Total automation eliminates the human agent option and lowers customer loyalty, NPS and CSAT. This is slowly but consistently progressing as GenAI dreams collide with CX reality. Customer retention is much les expensive than acquisition, and with no human agents available, loss of that key ‘human touch’ as an option can create a negative impression.
#3: Agents Must Remain in the Mix
Options with no live agents potentially place customers in a never-ending automation loop that is quickly frustrating, generating a lasting negative CX that is very memorable. As a customer, I can say first-hand that rare but recent customer issues with eBay, UPS, Hertz and others ended in extreme frustration as they had virtually or literally no choices for human agents. Getting caught in the endless automation loop with 5-10 options to choose from and no agent options left a serious negative impression. Now, multiply that 6-8 times with repeated attempts to reach an agent that all failed. My personal plummet in CSAT, NPS and loyalty to those brands is notable. It’s burnt in and will last. On the flip side, Amazon and Apple are so, so big now that I want a reason not to like them, but their ongoing live agent support keeps me trusting them.
Errors happen. It’s simply reality and always will be. Agents can fix those tricky ones while basic IVA’s alone cannot. Discover Cards (NYSE: DFS) now highlights that as its key selling point, running advertising touting live agent support! Just the right touch of humor reminds customers of the pain of ‘no agents ever’.
#4: Agent Training Needs
Think upskilling or reskilling. Agent training programs must be well designed on the new systems or even just for agent-assist GenAI tools. Agent turnover has leveled but remains a challenge as the global economy hovers and live agents aren’t as hot as they once were. Agent-learning systems must be comprehensive yet cost-effective, simple and not too time consuming. This area, where I repeatedly see contact centers already markedly weak on user-friendly training, now needs even more effort and simplicity if they expect agent usage. They are simply crucial, or the agents will just not use the GenAI tools. Just adding another boring chapter to already long training manuals means agents will instead just use traditional tactics. Agent retention, a long-standing challenge, emphasizes this glaring need. Agent usage impacts investment ROI.
Bringing It All Together
GenAI customer service presents additional challenges beyond the agents, including consistency and accuracy, NLU/NLG (natural language generation) quality, bias and fairness, CRM integration, data privacy and security… let’s just say there’s more! Any of these present notable risks when not well managed, but as they are outside of the agent’s world, I will leave those for another day. Adding all of these challenges makes it understandable that some are very hesitant after analyzing risk/reward. Customer service BPO’s are a good example, where several have announced plans for the future, but what will actually happen in reality there remains in question.
Next up will be chapter 3 and what is really needed here – how to succeed regardless of which customer service sector you are in, how quickly the GenAI market evolves and just what direction it takes.