Your AI May Sound Smart. It Still Can't Do the Job.
DCX Links April 19, 2026
Welcome to the DCX weekly roundup of customer experience insights!
Most companies say they are getting smarter with AI. A lot of them are just getting better at hiding the same old problems.
This week’s stories all point to the same tension. Companies want faster service, cleaner operations, and more personalized experiences. Fair enough. But speed is useless if the agent cannot resolve. Personalization is weak if it just dresses up a clunky decision process. And better dashboards do not matter if leaders still cannot see where customers are getting stuck.
That is what makes these pieces worth reading together. One shows why most AI agents still fail the real frontline test. Another gets at a harder truth: the biggest drag on outcomes is often buried in handoffs, dependencies, and work nobody can fully see. Starbucks offers a more practical example of where this is going, using AI to help customers figure out what they want before they ever hit checkout. Then the stat brings it back to earth. Executives think the experience is strong. Customers are a lot less convinced.
The common thread is simple. The gap between what companies think is happening and what customers actually live is still wide. AI can help close that gap. It can also make it worse by adding polish where the real issue is still unresolved.
So here is the standard: is your team using AI to remove friction customers actually feel, or to make internal blind spots look more sophisticated?
Let’s dig in.
NEW: The DCX AI Today - AI Tool Directory - Every tool from the last 50 DCX AI Today editions, sorted by category, with links and descriptions. If you lead a CX team and want a curated shortlist of tools worth evaluating, this is your starting point.
This week’s must-read links:
Most AI agents still fail the job interview
AI is getting better at finding the stuff leaders miss
Starbucks is making drink discovery feel more like a chat
Your brain is not built to be locked in all day
DCX Stat of the Week: Leaders think CX is stronger than customers do
DCX Case Study of the Week: Cebu Pacific scaled support without losing the human touch
I’m obsessed with Wispr Flow Pro! Get a Free Month Now!
Most AI agents still fail the job interview
A lot of companies are buying “AI agents” the way people buy kitchen gadgets on late-night TV. The demo looks slick. The real thing disappoints. This Ada Guide on How to evaluate AI customer service agents is useful because it gives you a better test: stop admiring the pitch and start interviewing the agent like it’s applying for a frontline job.
What to pressure-test:
Ada says a real service agent should handle eight basics well, including multi-part questions, follow-ups, consistent answers, relevance, semantic understanding, and direct responses instead of article links.
That matters in the real world because customers do not speak in neat, one-intent tickets. They ask messy questions, stack requests, and expect the system to keep up.
Where the money gets wasted:
If the agent cannot hold context, answer the same issue consistently, or resolve without dumping links, you did not buy automation. You bought friction with a bigger invoice.
Ada puts it plainly: an agent without semantic understanding is basically an expensive search bar. Hard to argue with that.
The CX To-Do: Run your vendor bake-off on real customer contacts. Same issue. Different wording. Follow-up included. Does it resolve, or just perform?
🔗 Go Deeper: Ada
AI is getting better at finding the stuff leaders miss
This one lands because it names a problem a lot of companies live with every day. People are working hard. Meetings are happening. Dashboards are full. Results still feel stuck. The point here is that the problem often is not effort. It is the hidden patterns underneath the work.
What’s going on:
Most outcomes are shaped by a mess of decisions, handoffs, and dependencies spread across teams.
No single leader can really see all of that clearly while the work is happening.
What AI changes:
AI can look across time, teams, and data and spot patterns people miss.
That makes it useful for seeing why work that looks productive does not always produce a better outcome.
Why a CX leader should care:
This is not just an operations story. It is a customer story.
If AI helps you see which steps actually drive trust, speed, and resolution, you can fix the journey before customers pay the price.
That is a lot more valuable than using AI to just make the same broken process move faster.
🔗 Go Deeper: Microsoft
Starbucks is making drink discovery feel more like a chat
This is one of those moves that looks simple on the surface but says a lot about where CX is going. Starbucks is testing a ChatGPT experience that helps people find a drink by describing a mood, a craving, or even sharing a photo. That matters because most people do not walk around thinking in menu language. They think in moments. I want something cold. I want a little boost. I want something that fits today.
What stands out:
Starbucks is meeting customers where their thinking actually starts, which is usually fuzzy, not precise
The experience helps turn that loose intent into a drink suggestion without making the customer do the work
What it means:
This makes discovery feel easier and more personal
It also gives Starbucks a new way to shape demand before the customer ever gets to checkout
The CX angle:
This is a good example of AI being useful before the transaction, not just during service
The bigger idea is clear: brands can use AI to help customers decide, not just help them search
🔗 Go Deeper: Starbucks
Your brain is not built to be locked in all day
This one hit home for me because it pushes back on a bad modern assumption: that every quiet moment needs to be filled. Darius Foroux argues that mind-wandering is not the enemy of good work. In many cases, it is what helps good work happen. When your brain has a little room, it can connect ideas, process what you’ve taken in, and surface thoughts you were too busy to notice.
What makes it useful:
Darius separates focused attention from reflective attention and makes the case that you need both
He also makes a smart distinction between anxious mental spinning and healthy wandering driven by curiosity
Why it matters for CX people:
If your job depends on judgment, pattern recognition, and insight, constant input can actually get in the way
Some of your best ideas about customers, friction, and what to fix next probably will not show up while you are cranking through Slack and meetings
The practical takeaway:
Protect a few small pockets of real off-time
No podcast, no scrolling, no filling the silence
Sometimes the most useful thing you can do for better thinking is back off a little
🔗 Go Deeper: Darius Foroux
DCX Stat of the Week: Leaders think CX is stronger than customers do
84% of executives think customers would recommend their company. Only 45% of customers actually would.
Takeaway:
That gap is expensive. It means a lot of companies are managing CX from the conference room, not from reality. If your team thinks the experience is working but customers are far less likely to recommend you, you have a perception problem before you have a loyalty problem.
Source:
Capgemini Research Institute, Reimagining customer experience: Human-led, AI-powered (2026). The report draws on surveys of 9,500 consumers across 16 countries and 1,200 senior leaders and frontline staff across 13 markets.
🔗 MORE STATS: Daily Stats on Substack Notes
DCX Case Study of the Week: Cebu Pacific scaled support without losing the human touch
CX Challenge: As the Philippines’ largest airline by passenger volume, Cebu Pacific had to support large volumes of travelers quickly and consistently. The risk: long waits, uneven answers, and frustrated passengers during stressful travel moments.
Action Taken: Cebu Pacific partnered with Ada to deploy AI customer service across 11 languages, automating routine questions so support could be available at scale while freeing human teams for more complex cases.
Result: The airline now resolves 75% of customer interactions autonomously and increased customer satisfaction scores by 50%.
Lesson for CX Pros: The best automation does not feel like avoidance. It removes friction from repetitive journeys, gives customers faster answers, and creates more space for agents to show empathy where it matters most.
Quote: “Partnering with Ada allows us to set a new benchmark in customer service, leveraging AI to better serve our passengers.” — Candice Iyog, Chief Marketing and Customer Officer, Cebu Pacific
Further Reading: Ada
Have a case study to share? Reply and let me know!
See you next week.
If this edition sparked ideas, share it with a colleague or team member. Let’s grow the DCX community together!
✉️ Join 1,500+ CX leaders who stay ahead of the next customer curve.
Human-centered insights. Plug-and-play frameworks. Smart tools that actually work.
All designed for CX pros who want to build with purpose—and deliver with impact.
👉 Subscribe today and get the tools to elevate your strategy (and your sanity).