We closed 2024 with mixed results for AI in customer experience (CX). While enthusiasm for the transformative potential of AI in customer service remained strong, fewer projects than anticipated moved beyond proof of concept (POC) to production, and even fewer scaled to meet sky-high expectations. A recent RingCentral-commissioned study, RingCentral Trends 2025: The state of AI in business communications, highlights the progress made, as well as the challenges organizations faced. 2024 provided tons of learning, and this article aims to unpack those lessons to help you refine your CX AI strategy and unlock its full potential.
AI projects demand a fundamentally different approach, beginning with a clear articulation of your strategic goals. These goals may range from focusing solely on cost reduction to leveraging AI to enhance the quality of the customer experience. Defining these objectives is important for driving change management and ensuring adoption—reassuring customers worried about AI hindering access to human support and addressing employee anxieties about job security.
AI requires a phased approach, tackling one use case at a time—beginning with a proof of concept before advancing to production. Adopting a product approach is equally essential for scaling results and driving continuous improvement. This phased, agile strategy must be paired with the establishment and reinforcement of a solid foundation to address governance issues and build a community of AI practitioners and experts that fosters experience sharing.
Proving ROI has been more challenging than expected, making it critical to start with a comprehensive view of potential impact—one that extends beyond cost reduction to include measures like quality, speed, reduced cognitive load on employees, and improved customer experience. Equally important is gaining control over the cost side of the equation, ensuring expenses are both trackable and predictable.
Central to success is the creation of a portfolio of use cases that can be sequenced into a roadmap, aligned with your key business goals, and structured to allow you to walk before you run. My research has identified over 30 potential use cases. This extensive list reflects AI’s broad impact across every element of your CX stack.
This diagram brings them all together. Start from the left, where the customer attempts self-service before potentially being transferred to a human agent. The self-service options and assisting agents present numerous AI use cases. Further to the right are supervisors and the back-end functions of customer service. The instrumentation layer below the middle line captures interaction and activity details. AI plays a key role in data collection, stitching elements together, and interpreting both structured and unstructured data. This layer also assembles the data that powers your AI. At the bottom, lay your enterprise applications—transactional systems and customer databases. Above the middle line is the decision layer, where interactions are routed, work is allocated, and recommendations are provided to agents. At the top, your knowledge bases can be better managed, federated, and made available to customers and employees using AI.
Reflecting on last year’s challenges, we’ve observed that a top-down mandate to explore how GenAI can transform customer service operations often leads to a dangerous, technology-driven approach. We strongly advocate for flipping the conversation and starting with your business priorities. It is important here to point out that there is no single one-size-fits-all approach to implementing AI. While I identified over 30 different use cases, you should focus on the use cases that align with your business’s unique pain points and priorities. The layout above aligns with the top jobs to be done (JTBDs), offering a framework to identify the right AI technologies you can apply to your most pressing issues.
After selecting your most relevant use cases, create a framework to rank them based on several key factors:
- Their ability to impact your business goals
- Data readiness
- Potential risks
- The impact on stakeholders—both customers and employees
- Your ability to measure value and track costs
- The ease of sourcing the technology and solution
It’s impossible to capture all the learnings from the past year in a short article. We encourage you to watch this webinar, where John Finch, VP of Product Marketing for Customer Experience at RingCentral, and I delve deeper into these insights.
Originally published Feb 13, 2025, updated Mar 04, 2025
