The Real CX Question: What Should Humans Do More Of Now?
DCX Links May 10, 2026
Welcome to the DCX weekly roundup of customer experience insights!
AI is getting better at handling the routine. That part is no longer the headline.
What matters now is what happens next. Where do you redeploy human judgment once the easy tasks are off the queue? How do you keep accountability clear when AI starts acting like a teammate? And how do you make sure the customer signals buried in calls, meetings, and escalations actually turn into better decisions?
This week’s stories point in the same direction: the next CX advantage will not come from automation alone. It will come from how well brands use AI to reduce effort, preserve trust, capture insight, and free their people to do the work customers still value most.
Let’s dig in.
This week’s must-read links:
IKEA’s Better AI Lesson: Redeploy the Capacity
Return on Imagination Is Becoming a Business Metric
Memoket Wants to Save the Stuff We Keep Losing
AI Agents Need Owners, Not Job Titles
DCX Stat of the Week: Customers want AI explanations—most brands still don’t provide them
DCX Case Study of the Week: Wyndham Automates the Stuff Guests Shouldn’t Wait For
I’m obsessed with Wispr Flow Pro! Get a Free Month Now!
IKEA’s Better AI Lesson: Redeploy the Capacity
IKEA gives you the more useful AI question: once automation removes routine work, where should human judgment go next?
The case is simple. Ingka Group says its AI chatbot Billie resolved about 47% of customer inquiries from 2021 to 2023, covering 3.2 million interactions and about €13 million in savings. At the same time, 8,500 call center co-workers were reskilled into areas like remote interior design, digital retail sales, relationship building, and complex problem solving. Remote customer meeting points reached €1.3 billion in FY22 sales.
The better lesson:
Do not just celebrate deflection. Study what the bot cannot handle.
Those escalations often show where customers still want confidence, advice, taste, or reassurance.
That is where your people can create value the bot cannot.
The CX move:
IKEA did not turn service capacity into a spreadsheet savings story.
It turned frontline knowledge into a paid advisory channel.
That should make every CX leader look at unresolved contact reasons differently.
The catch:
Reskilling only works when leaders design the next role clearly.
“Use AI and do more” is not a plan.
Name the customer need, the human skill, and the revenue path.
The CX To-Do: Pull 90 days of bot escalations. The growth idea may already be sitting in the “too hard” pile.
🔗 Go Deeper: CXToday
Return on Imagination Is Becoming a Business Metric
Nish Patel’s says that AI changes what companies should value next. Execution gets cheaper. Building gets faster. The scarce thing becomes imagination: the ability of people across the business to spot what could exist, shape it, test it, and turn it into value.
The point:
Return on Imagination asks a practical question: how much of your team’s creative and customer-facing intelligence actually becomes something useful?
Most companies waste that capacity. Ideas get trapped in meetings, backlogs, feedback notes, and “we’ll revisit this.”
AI lowers the cost of turning a rough thought into a draft, prototype, workflow, offer, or experiment.
The CX read:
Your frontline teams are sitting on more than service issues. They hear unmet needs, confusing moments, workaround behavior, and buying hesitation every day.
The opportunity is not just faster execution. It is building a system that captures those signals and moves the best ones into action.
The catch:
More imagination also creates more judgment work.
Leaders still need filters, ownership, funding rules, and the discipline to kill weak ideas quickly.
The CX To-Do: Ask where your company’s best customer-informed ideas go after someone notices them.
🔗 Go Deeper: Tomorrow with Nish
Memoket Wants to Save the Stuff We Keep Losing
Memoket Gem is built around a familiar problem: the best parts of conversations usually disappear after the meeting ends.
You remember the concern, the idea, the next step, or the customer quote just well enough to know it mattered, but not well enough to do anything cleanly with it later.
What it does:
Memoket is a wearable AI device that captures meetings, conversations, or idea dumps with one press.
The app turns those recordings into summaries, tasks, and usable documents.
It also connects context across conversations, instead of treating every recording like a separate note.
The CX angle:
This could be useful anywhere customer insight gets lost between the conversation and the follow-up.
Think sales calls, research sessions, frontline feedback, project meetings, and customer escalations.
The value is not recording more. It is turning what was said into something the team can act on.
The catch:
Privacy needs a hard look before this touches sensitive customer work.
Memoket says user data is not used to train public AI models, but full storage and security details are coming before launch.
The CX To-Do: Find the conversations your team keeps forgetting, then decide whether they need better capture, better follow-through, or both.
AI Agents Need Owners, Not Job Titles
The “AI employee” idea sounds clever until something breaks and everyone quietly points at the bot.
That’s the useful warning in this HBR research. When companies frame AI agents as teammates or employees, people do not necessarily adopt them more. They often review the work less carefully, escalate more, and feel less clear about who owns the outcome.
The accountability problem:
Personal accountability dropped when AI was framed as an employee instead of a tool.
That is a serious CX risk. Customers do not care whether the mistake came from “ALEX-3,” a workflow, or a human.
They care that the wrong answer, bad decision, or broken handoff landed on them.
The hidden cost:
Managers caught fewer errors when reviewing AI “employee” work.
Escalations also increased, which means the work may move faster at first, then slow down through rework and second-guessing.
That is how AI creates new friction while everyone is celebrating productivity.
The better move:
Treat AI agents as powerful software with named human owners.
Define decision rights, review rules, escalation triggers, and consequences before the agent touches customer-facing work.
The org chart does not need a bot. The operating model needs clearer accountability.
The CX To-Do: Before naming your AI agent, name the human who owns its output.
🔗 Go Deeper: Harvard Business Review
DCX Stat of the Week: Customers want AI explanations—most brands still don’t provide them
95% of customers want to know why AI makes the decisions it does, but only 37% of CX leaders say they currently offer any reasoning behind AI’s decisions.
Takeaway: AI transparency is becoming part of the customer experience, not a compliance footnote. CX teams need to design explainability into customer-facing AI before trust becomes the bottleneck.
Source: Zendesk CX Trends 2026
🔗 MORE STATS: Daily Stats on Substack Notes
DCX Case Study of the Week:
Wyndham Automates the Stuff Guests Shouldn’t Wait For
Wyndham’s contact center problem was not really a technology problem. It was a guest-effort problem. Travelers calling from airports, hotels, or in transit were running into disconnected systems, slow responses, and agents forced to jump across CRM, loyalty, and booking tools while customers repeated details. That is the kind of internal mess guests feel immediately.
The problem:
Wyndham’s partially on-premise contact center platform was complex, unreliable, and hard to scale across locations.
Agents lacked clean access to guest context, which made urgent travel issues harder to resolve.
Slow onboarding also made it harder to support seasonal peaks and new markets.
The move:
Wyndham moved to Five9, with Salesforce integration and AI agents.
The goal was not just automation. It was giving agents better context while removing repetitive tasks from the queue.
The result:
Five9 AI Agents now handle about 40,000 password resets a month.
Wyndham automated 80% of booking cancellation calls.
The company reports a 62% automation rate, under 1% abandonment, and millions saved by moving to cloud.
The CX lesson: Automate the tasks that create customer waiting and agent drag, then protect human capacity for the moments where judgment, reassurance, or revenue matter.
Further Reading: Five 9
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!