Prepare Now for GenAI Advances
Two top research firms are recommending that organizations plan now for the increased capabilities that generative artificial intelligence (genAI) will have in the next few years.
Organizations must address a decline in tolerance for legacy self-service and focus on simplifying digital services through conversational genAI, according to Patrick Quinlan, a Gartner senior director analyst.
Gartner predicts that by 2028, 70 percent of customer service journeys will begin and be resolved in conversational, third-party assistants built into mobile devices. Quinlan expects the integration of genAI into mobile devices to disrupt traditional customer service channels by offering a more efficient and intuitive user experience.
“As customers increasingly ask their devices to help solve service issues, companies will need to decide whether to compete with these experiences or embrace third-party support,” Quinlan adds.
If they choose to compete, they will need to enhance their mobile app experiences to support conversational interfaces and ensure these tools can leverage existing knowledge to effectively resolve issues, he says.
Quinlan recommends that companies prioritize simplifying customer journeys by pivoting from omnichannel strategies, which have become convoluted, to omnimodal strategies that enable customers to communicate with companies through single digital channels using their preferred modes of communication. “That mode could be text, voice, video, image, or any combination. By doing so, businesses can maintain customer satisfaction and loyalty while adapting to the evolving digital landscape,” he states.
Companies can significantly enhance service delivery and customer satisfaction by strategically leveraging mobile app architecture and customer data, according to Quinlan. “As mobile devices become central to customer interactions, businesses should embed service workflows directly within their apps, creating a seamless and integrated user experience. This approach not only aligns service delivery with the customer’s mobile journey, but it also allows for personalized interactions based on the data held within the app.”
Quinlan adds that while current third-party genAI tools mainly handle simple queries using public information, future advancements in large language models (LLMs) and large action models (LAMs) could greatly expand their capabilities. These improvements might enable third-party AI agents to perform more complex tasks, such as navigating portals and completing transactions, autonomously.
Working with third parties has certain benefits, but also risks, Quinlan cautions. On the one hand, it allows companies to integrate into increasingly popular, low-effort experiences and leverage the vast resources and expertise of third-party platforms, potentially reducing operational costs and enhancing customer satisfaction. However, it also raises concerns about data security, analytics insights, and brand control. It is entirely possible, for example, that customers could defect from companies because they received incorrect answers from third parties acting on the companies’ behalf, without those companies ever knowing.
“To mitigate these risks, companies must establish robust knowledge management systems that ensure accurate and consistent information is available to third parties while implementing stringent security measures to protect data integrity,” Quinlan says. “By carefully managing these dynamics, businesses can capitalize on third-party capabilities to improve service efficiency without compromising their brand or customer trust.”
Improved Knowledge Capture Needed
But unless organizations harness the crucial tacit knowledge from their contact center agents, they won’t realize the full cost-cutting benefits and customer service improvement possible from automation, Forrester Research concludes in an unrelated report.
“Brands must transition to AI-led customer service interactions that human experts support in the background,” the research firm says. “Brands must push their vendor partners to develop next-gen agent workspaces that effectively harness and capture agents’ tacit knowledge, but the evolution will likely take years.”
To prepare for such far-reaching changes, Forrester recommends that companies make several changes.
First, they need to remove the following three barriers that hinder the capture of employee tacit knowledge:
- The standard approach to optimizing customer service AI solutions relies on analyzing conversation transcripts to train models. However, such solutions don’t capture many elements of conversations, like decisions that human employees make, according to Forrester.
- Variable transcription quality, particularly when multiple languages are involved, can introduce bad data into the AI. This situation is often further complicated by the need to have a well-rounded roster of agents, who all have different problem-solving approaches and communication styles. Such diversity makes it difficult to define and replicate a uniform standard for AI.
- The start and end of conversations are the easiest parts to replicate. Even if AI could effectively handle the rest of the conversation, it’s difficult to handle the middle, often resulting in handoffs to human agents. These handoffs are prone to loss of context, requiring customers to repeat information that they already provided to a chatbot or interactive voice response system.
Once the shift to AI-led customer service comes to pass, everything in customer service will change, Forrester predicts.
Metrics, for example, will change to be business outcome-focused, and companies will no longer be concerned about the cost of interactions but will instead consider the profitability of each interaction, according to Forrester.
Contact center-as-a-service (CCaaS) vendors will transform or perish, the firm adds, noting that as the shift to AI-led customer service unfolds over time, agent seat counts will decrease and vendors that rely on seat-based commercial models will have to adapt.
Workforces and the workforce management (WFM) systems that govern them will look different, it also expects. With AI leading customer interactions, companies will no longer need WFM solutions to forecast and schedule enormous pools of agents.
While some vendors have already started to explore options for next-generation workspaces, most are still only providing incremental updates that do little to challenge legacy systems. Nor do they fully leverage generative AI’s potential, according to Forrester.
The research firm expects next-gen workspaces to evolve in three phases.
The first will involve harvesting and codifying knowledge from the front office. To accomplish this, companies must allow AI to lead customer interactions from start to finish. AI needs to consult with human agents for exceptions. The agents will unblock the AI in real time. Legacy systems can’t do this, so organizations will need a new workspace optimized for enabling and handling human judgment calls. Vendors must also develop feedback loops to enable companies to capture the rationale of human decision making.
The second phase will involve accelerating customer service through AI. While human “unblockers” will be systematically integrated into AI systems so they can handle increasingly complex interactions, human expertise will still be needed for the most complex interactions.
So in phase two, AI will streamline back-office orchestration; recommend actions based on enterprise goals; embed into systems of work; and identify opportunities for human-led value, according to Forrester.
The third phase will enable broader orchestration and optimization through AI. According to Forrester, AI will enable organizations to deploy personalized customer service programs at scale, using real-time feedback loops to dynamically optimize service outcomes.
The research firm recommends that organizations start planning for this future now by co-creating, integrating, and building trust in genAI.