Editorial
Research Report
Building the Case for Modern Public Sector Messaging
The Gist
- Humans are the standard. AI that replaces a person should be just as fast and easy for customers.
- Headless CX helps. Separating the presentation layer from the data delivers a more customer-centric experience.
- Contextual data matters. Customers get faster, more relevant resolutions when the AI has more high-quality data.
All I wanted was a cheeseburger before my flight. But when I scanned the QR code on the table at the airport restaurant, it told me to switch my WiFi to their ordering network … then create an account … then check a box if I wanted tomatoes, onions, or extra cheese. I was almost 10 minutes in when the server saw my frustration— and took my order in less than a minute.
When I shared this anecdote on LinkedIn, it racked up hundreds of reactions and dozens of comments. Clearly, I’m not the only one frustrated by having to jump through (digital) hoops.
When tech saves your customers time and hassle, it’s a win-win. But when customers have to put in more work to get the same results (or worse), they’re not going to be happy — especially when they’re already hangry.
AI and the Context Switching Problem
While we’ve seen significant progress recently, many of today’s AI-powered CX solutions still require context switching, which adds friction and results in a poor customer experience. Ask yourself— when your customers use your chatbot, can they do everything they need in the same window, or do they have to check their texts for a link, navigate a separate payment system or open a new tab for more info? As Microsoft noted, “if interacting with AI feels unnatural, adoption suffers.”
- Context switching hurts customer retention. The online restaurant menu frees up employees (at least theoretically), but at what price? How many of your customers would have given up after five minutes of trying to order a burger and fries? Journey fragmentation will be “the greatest barrier to CX excellence,” according to a Capgemini retailer survey.
- Employees often pay the price. AI platforms that don’t deliver a seamless experience put more stress on your frontline employees when they have to do what AI should have done — all while dealing with frustrated customers.
Of course, that’s not even accounting for the risks of AI-washing, which can further erode customer trust.
Your goal for agentic AI should be to make flow portable, allowing work and context “to move together across systems without breaking the customer story.” When it’s done right, you get success stories like Lenovo improving customer service efficiency by 50%, and IBM supporting an AI-powered assistant with a 94% customer satisfaction rate.
Related Article: Where Agentic CX Pays Off First (And Why It's Not Customer-Facing)
Why You Need to Hold AI Agents to the Same Standards as Humans
Is a 2x time savings for the server worth making the customer’s experience 10x worse?
Even if it were the other way around — a 10x savings for the waitress for a 2x worse customer experience, many CX leaders would say it’s not a smart trade-off.
Humans are still more efficient at many CX tasks, at least from the customer’s standpoint. The server at the restaurant didn’t ask me to scroll through pages of appetizers and entrees — she just took my order. And even if she messed up, customers are willing to accept more mistakes from humans than from AI.
Overall, people still prefer human support vs. AI, according to Capgemini research. That said, the desire for human-driven CX is stronger in high-risk industries like banking and pharmaceuticals, while customers actually preferred conversational AI support in retail and travel — all the more reason to improve your AI platforms.
If you’re thinking about firing a human and replacing them with AI, the standard shouldn’t be “does the AI function from a technical standpoint?” Instead, look at it from the customer’s perspective, and ask if the AI is at least as good as the human. If not, you’re just disrespecting your customers.
Content Leaders Collective: Proving Content's Business Impact Starts With the Right CCMS
Register
Replacing Tasks, Not Roles: The Changing Nature of Contact Center Work
Register
How Fast-Moving Content Teams Keep Up as Sites Grow
Register
From Automation to Intelligence: How Leading Teams Are Rethinking Operations
Register
Ultimately, AI-to-AI commerce — when your customers’ agents talk directly to your company’s agents — may remove the friction. But until then, it’s your job to watch out for any solution that makes your customers do more work.
Reducing AI Friction in Customer Experience: Key Shifts and Actions
Editor's note: The following table highlights the most important lessons, actions and strategic considerations emerging from how to deliver a more frictionless AI-driven customer experience.
| Key Area | What Happened | Why It Matters | Recommended Action |
|---|---|---|---|
| Architecture | Headless CX decouples the front-end presentation layer from back-end data | Enables a consistent experience across any channel or device | Evaluate headless architecture to minimize context switching |
| Interface Design | Drivers completed tasks roughly twice as fast with physical knobs than touchscreens | Human-centered design can outperform flashier AI-driven interfaces | Prioritize speed and ease of use over interface novelty |
| AI Orchestration | Orchestration layers route requests to the agents and data that can act on them | Poor orchestration causes failures on out-of-the-box requests | Ask vendors to demonstrate exactly how orchestration connects LLMs, chatbots and data |
| Service-Level Objectives | SLOs define the outcomes that actually matter to the business, like a three-minute booking flow | AI adoption without SLOs risks optimizing for the wrong thing | Establish SLOs first, then evaluate whether AI helps meet them |
| Contextual Data | Real-time context (like weather) can shift what's promoted to customers | Contextual relevance drives more authentic, timely interactions that deepen customer loyalty, per IBM | Build contextual data into AI-driven recommendations |
| Outcome-Driven CX | Nearly 80% of executives are shifting CX from query-based interactions to outcome-driven experiences, per Capgemini | Customers increasingly expect nuanced, goal-based requests rather than simple queries | Design AI to handle outcome-based requests, not just direct questions |
| Device Uptime | Kiosk, POS and endpoint outages undermine AI regardless of its quality | AI is only as reliable as the hardware delivering it | Invest in endpoint monitoring to proactively catch and fix device issues |
Right now, you’re caught between two worlds — your CFO wants the cost savings of AI, but your customers want their burger and fries without tapping through five different screens. But your customers shouldn’t be paying the price for your shiny new AI. Because even if your customers don’t have a plane to catch, their patience — and their loyalty — will only go so far.
Learn how you can join our contributor community.