人工智能代理实用指南

Understand, deploy and refine AI agents for more personalized,efficient and human-centered customer experience INSIDE Foreword Artificial intelligence (AI) is rapidly transformingcustomer experience, and AI agents are at thecenter of that transformation. These intelligentapplications can autonomously perform tasks, The urgency is clear. First, the pace of AI development isaccelerating, making the technology more capable andaccessible. Second, customer expectations are at anall-time high. Consumers now expect real-time, effectivesolutions. Often, AI can meet these demands more What AI agents dopage 4 Tasks and teamworkpage 7 Personalization and pitfallspage 11 Conclusionpage 15 About Talkdeskpage 16 AboutCX Networkpage 17 Foreword Findings from the 2024 Global State of CX survey reflectthis momentum. Half of respondents strongly agreedthat automation is critical to delivering CX at scale. Morethan 40 percent named CX and service automation as empathy. Trust remains paramount. Customers needto know when they’re interacting with AI and that theirdata and experience are being handled responsibly. Equally important are feedback loops—the mechanismsthat allow continuous improvement. These include real-time performance monitoring, flags from human agents, Training is also vital. Effective AI agents are built on high-quality, representative data drawn from real interactions.Human input is key – during training and after CX leaders are now navigating a transformative momentin which technology doesn’t just support service butfundamentally reshapes it. The opportunity lies in Yet for all the promise AI agents hold, their adoptionbrings new challenges. Chief among them is maintaining ThisCX Networkreport, produced in partnership with Talkdesk, explores the practical deployment of AI agentsand copilots in customer service. Through real-worldcase studies fromTireHub, Humann, Michaels, andOuro, IN THE NEXT SECTION: Discover the potential of AI agents and thepossibilities they bring to CX. What AI agents do AI agents enable automation in CX. When appliedstrategically, it can do the heavy lifting, so humanscan focus on more complex problems and providing service needs. Additionally, organizations shouldassess their needs regarding the capture, retrieval,and delivery of in-house and third-party data systems. Generalist agents:They are designed to handle a widerange of tasks as opposed to being limited to a specific Copilots:These are AI tools designed to assist usersin completing complex tasks by offering suggestions,knowledge or step-by-step guidance in real-time. “If you want to start with AI, don’t start with AI,” saidDominik Olejko, one of the CX Network’s top leaders ofAI in CX to follow in 2025, when he was speaking at the The next step is understanding the various types of In fact, Toussaint Celestin, product marketing principal,AI at Talkdesk, suggests asking questions before Crafting impeccable CX in the age of AI requiresthoughtfulness about delegating work to the machinesversus the humans. Celestin suggests mapping out the Voice agents:Think Siri and Alexa! These agents usespoken language as their primary mode of interaction,and they answer questions and can have conversations “The first step in finding the right balance is to clearlydefine the goals and expectations of your customerservice operation. Organizations should consider notonly what they hope to achieve by incorporating AIautomation—such as reducing wait times, improving Text agents:These agents interact with users throughwritten input via chatbots or messaging apps. Hybrid agents:They combine both voice and textinteraction methods, offering flexibility in communication. Task-Specific agents:These agents are designed toperform tasks. Examples include agents that handle What AI agents do “Not all customer interactions are the same andtreating them as such can lead to poor experiences.Routine, repetitive tasks such as password resets,order status inquiries, appointment scheduling andfrequently asked questions are ideal candidatesfor AI-powered self-service and automation,” saysCelestin. “These tasks are straightforward and canbe handled efficiently by virtual agents or chatbots, outsourced call center and received negative feedbackfrom customers because they would end up talking toagents who knew nothing about the business or their phone number, but TireHub would also maintain acentralized 800 number that guarantees a responsewithin three seconds. Specialists are in the physical “Customers love having the choice, and the majorityare willing to wait a bit longer to speak with their localspecialist. This preference highlights the value customers “We wanted to focus on the entirety of a customer’sbusiness from their perspective,” says David Distel,national customer service leader at TireHub. “That To accomplish this renovation of CX, TireHub is applyingTalkdesk’s Copilot, an AI-powered assistant that detectstopics, o