The AI Is Not Just Finding the Trip. It Is Picking It.
Plus: Travel is giving us a useful preview of what happens when AI stops answering questions and starts helping customers make the call.
Your daily signal on AI and CX — minus the hype.
DCX Stat of the day: Locus says 45% of consumers now use AI as either a primary or secondary online shopping tool. Locus
In this issue:
→ AI moves closer to booking decisions
→ Reliability beats pure delivery speed
→ Marketing agents need visible guardrails
→ Risk agents follow the whole customer
→ Vertical bots meet contact center complexity
🔎 Deep dive
The AI travel agent now has a point of view
Priceline just moved Penny from helpful travel bot into something closer to a decision partner. The new version can compare live travel options across flights, hotels, cars, maps, pricing, traveler preferences, and booking context inside one conversation.
That matters because travel is not a simple search problem. Customers are not only asking, “What is available?” They are asking, “What makes sense for this trip, this budget, this risk, and this moment?” When an AI starts surfacing its own pick, explaining the tradeoff, and carrying preferences from one trip to the next, the journey shifts from browsing to guided commitment.
Here’s where it breaks. If the AI gets the preference layer wrong, the customer does not experience a bad recommendation as a model issue. They experience it as a brand that misunderstood the trip. So this is not just a chatbot question anymore. Once AI shapes choice before checkout, product, service, legal, data, and brand all have a stake in what the agent recommends.
Source: Priceline
📬 Copy-Paste Take
AI booking agents are not just a support deflection play. They are becoming part of the decision architecture. If we let them recommend, rank, and remember, we need to know what they are optimizing for. Accuracy is only half the job.
OPERATOR PLAYBOOK
Audit the moments where AI narrows choice
Before you celebrate an AI agent for reducing clicks, look at the choices it is quietly removing. Convenience can absolutely help the customer. It can also hide the wrong tradeoff.
Audit every AI-guided purchase or booking flow for four things:
What the agent is allowed to optimize for.
What customer preference data it remembers.
How it explains why one option is recommended.
Where a customer can challenge, reset, or escalate.
Then test whether a reasonable customer would understand the recommendation well enough to trust it without feeling pushed.
Ask your team: If this AI pick goes wrong, who owns the customer explanation?
Signal: Fewer clicks only count as better CX when the customer still feels in control.
📈 Market Reality Check
AI discovery raises the standard for fulfillment
This week’s Locus survey gives this travel story a wider commerce reality check. AI is already changing how people shop, with 45% of consumers using AI as a primary or secondary shopping tool. Among those shoppers, 39% say they are more likely to try new brands or products, and 37% are more likely to buy more items when using it.
That does not prove AI makes customers loyal. It proves AI can move demand into places your operation may not be ready to support. The important detail comes after discovery: 68% of shoppers say a fast refund makes them more likely to shop with a retailer again. AI may create the intent, but delivery, returns, and service still decide whether the relationship survives.
AI can open the door. Operations still has to keep the promise.
🧰 Tool Worth Knowing
MoEngage Merlin AI Custom Agents
What it does: MoEngage launched Merlin AI Custom Agents for lifecycle and CRM teams. Marketers can build agents that run on MoEngage data, define audience, channel, content, and budget guardrails, and see an activity log of what each agent did.
CX use case: A retention team could ask an agent to review campaign performance, draft a journey, QA an in-app message, or point out which segment is not responding before another generic campaign goes out.
Worth watching because: The useful part is visibility. MoEngage is not asking teams to trust a mystery box in the middle. It is giving them a way to see the data pulled, decisions made, channels touched, and content sent. No public performance metrics were shared in the launch.
Bottom line: Marketing AI gets more useful when the team can inspect it, limit it, and stop it.
The DCX AI Today - AI Tool Directory - If you lead a CX team and want a curated shortlist of tools worth evaluating, this is your starting point.
⚡ 90-Second CX Radar
Oscilar launches Agent Hub for fraud, compliance, credit, and onboarding
Risk is part of CX whether we like it or not. Oscilar is trying to coordinate more than 30 agents across the customer risk lifecycle, which matters because fraud checks, onboarding, disputes, and credit decisions are often where customers feel the cost of internal silos.
Vonage adds vertical AI agents to its contact center
Vonage is embedding healthcare, financial services, and retail-trained agents into Vonage Contact Center through Avaamo and Syndeo. The useful question is not whether a vertical bot sounds smarter. It is whether it can handle compliance, handoff, and context without fragmenting the customer.
🧭 Your Move
The pattern today is simple: AI is moving upstream. It is not waiting for a customer to ask for help after the journey breaks. It is helping the customer choose, buy, book, return, verify, and decide.
That is useful. It is also where sloppy ownership gets expensive. When AI starts shaping the customer’s options, the work can no longer sit with one platform owner or one innovation team. Someone has to own the promise the agent is making on behalf of the business.
If AI helps the customer decide, CX has to help the business decide what the AI is allowed to value.
Until tomorrow,
👥 Share This Issue
Think of one person who’s wrestling with AI in CX right now
and forward this to them.
I’m obsessed with Wispr Flow Pro! Get a Free Month on me.
If someone forwarded this to you, they thought you needed to see it before your next AI planning meeting. Get your own copy.