Crescendo is throwing out the old customer service playbook. By combining human expertise with its expanded agentic AI platform, the company is turning customer experience (CX) from a cost center into a competitive edge, delivering faster, more accurate service and strategic insights that drive business improvement.
Its integrated platform aims to change the customer service culture with a new approach to implementing agentic AI. According to Crescendo CEO Matt Price, it improves results through a fully managed service that combines advanced AI and human CX experts. It automates tasks and introduces a conversational CX Data Assistant to provide strategic insights into CX operations with 99.8% accuracy in customer service responses.
Zack Urlocker, Crescendo’s COO, says CX technology has been all about deflection for the last twenty years. Businesses would do anything to prevent customers from engaging with customer service.
“FAQs, voice systems, and help desk systems were all designed around deflection. We think with AI, there’s an opportunity to increase customer engagement. Companies that figure that out will find that CX isn’t just a cost center. It’s a competitive advantage,” he told CRM Buyer.
Fixing CX Failures With a New AI Playbook
According to Price, most companies launching AI initiatives will struggle for months, and many of those projects will fail to achieve the level of quality customers expect.
He cited Harvard Business Review research showing that up to 80% of AI projects are unsuccessful. Crescendo counters this trend with its fully managed service, which can be implemented in as little as three weeks.
Urlocker noted that typical human-staffed customer service teams can generally get to a 90% level of accuracy.
He added that his company eliminates AI hallucination or false data results, has clear guardrails, and a handoff to human experts whenever necessary to ensure accurate answers.
“Using AI technology, you can get to 70% or 80% accuracy pretty easily. But getting to 90% or 95% is much harder,” Urlocker said.
How Agentic AI Solves Customer Service Tasks
Urlocker agreed that agentic AI has become a buzzword and the subject of many arcane discussions about how it works. However, in the context of customer service, agentic AI takes on mundane tasks that would have required teams of people.
Most customers seeking answers through a customer service channel do not care whether the answer comes from a human, an AI assistant, an article, an FAQ, or another source. He suggested they want an answer without navigating a complex system to get there.
Customers typically inquire about order status, shipping details, and changing delivery locations. Generative AI cannot answer those questions by itself. The information is not logged in a knowledge base but comes from a back-end system. Agentic AI, like a human support agent, can look up customer information in the appropriate system and take the necessary action, noted Urlocker.
“We’ve seen great results with agentic AI being applied in retail and e-commerce customers. It can respond rapidly to the most common issues and generally gets very high marks. When a question comes up that it’s unable to deal with, it hands that off to a human,” he added.
AI Turns CX Into a Business Asset
Urlocker explained that CX managers deliver more value to the business with Crescendo’s CX Data Assistant. Its conversational chat interface uncovers trends and insights that help managers optimize operations.
“Instead of just being in constant fire-fighting mode, dealing with the backlog of customer service requests, they are now able to identify root cause problems and provide detailed suggestions on how to improve not only customer service, but in some cases, identify product problems before they become widespread,” he explained.
CX managers get a seat at the table to improve the business by gaining better insights into manufacturing defects, product reliability issues, and quality assurance. The Data Assistant also provides customers with service after hours and on weekends without requiring overtime.
“This trend toward 24×7 service is going to accelerate, and companies should be investigating ways of putting these solutions in place before seasonal surges,” he added.
Fixing CX With Unified AI System Integration
Urlocker described customer service systems as an evolving process of adding new technologies. Some automation and some human processing hold pockets of technology together.
Poor system integration resulted in a recurring experience: customers call a company for help, enter their information, and are transferred to a live agent who asks for the same information again.
The more different channels a company adds, such as email, voice, and online chat, the more integration is required, leading to poorly integrated systems.
“If you add a separate call center, things get more complicated,” he insisted.
Urlocker observed that CX managers often find themselves with the unenviable task of being their own system integrators. They attempt to integrate disparate systems in ways they were not designed for.
“Buying from a single SaaS vendor isn’t usually much better, as they have often assembled their suite from various acquisitions anyway,” he lamented. “Crescendo’s goal is to provide a complete solution for CX managers that enables them to step away from the integration business.”
Crescendo’s Approach to CX Integration
Crescendo is expanding its integrations with various CX applications. As part of its ongoing strategy, the company has integrated its agentic AI platform with Salesforce, Zendesk, Kustomer, Freshdesk, Amazon Connect, Genesys Cloud, Gorgias, and Shopify. New integrations are delivered every quarter.
“Our approach is to integrate with whatever customers have, whether it’s a CRM, an AI point solution, or even a complex order management system. If there’s an API, we can integrate,” Urlocker said.
That process involves examining everything happening with its customers, such as companies using Zendesk or an online chatbot with either a human or AI agent. The goal is to ensure that we can coexist with the customer’s existing systems and processes.
He noted that Crescendo does not operate in a silo, emphasizing that it serves as the operating platform for fully automated QA and insights across all interactions, regardless of source.
What Makes CX AI Projects Succeed or Fail
Applying AI to customer service is not a “set it and forget it” situation, Urlocker warned. It requires an ongoing effort to keep the knowledge base up to date and continually tune and refine how the AI handles operations as new products or policies are rolled out.
“This is one of the things that can go wrong in a do-it-yourself AI project. Everyone’s excited to launch the project, but who will continue updating things? You’ve got to plan for that,” he cautioned.
Urlocker agreed that his firm’s AI strategy is different from that of other companies. AI without human oversight and good guardrails is easily prone to running astray.
“We think it’s better to use AI to augment people rather than replace them. AI can be a very effective tool in providing highly empathetic customer service,” he advised.
He described AI as a very eager, fast-working junior intermediate employee. It works best with known issues.
“When it runs into a problem it’s not familiar with, meaning it hasn’t been trained, then you want it to escalate the problem to a more senior human. You don’t want it to just make up an answer,” he said.
That is the downfall of many companies that attempt to use AI on their own. If they try to apply AI to all problems, they may end up with AI creating new policies that alienate customers and cause problems, he suggested.
Loading ...
