AI That Forgets the Journey Will Cost You the Customer
Plus: smarter AI is nice, but smoother customer moments are what people actually remember
To opt-out of receiving DCX AI Today go here and select Decoding Customer Experience in your subscriptions list.
📅 March 10, 2026 | ⏱️ 5 min
Good Morning!
We keep talking about AI like the big win is better answers. It is not. The big win is fewer annoying moments. Today’s edition is all about consumer journey moments: the heads-up that comes early enough to help, the handoff that does not make people repeat themselves, the follow-up that feels useful, and the recovery step that does not sound like a robot reading a policy. So let’s get into where the journey is still breaking, and what smart CX teams should fix first.
The Executive Hook:
Here is the shift. AI is moving from “answer the question” to “help before the customer has to ask.” That sounds great. It can also go sideways fast. A proactive message can feel helpful or creepy. A smart handoff can feel seamless or sloppy. A remembered preference can feel thoughtful or weird. The edge for CX leaders is not more automation. It is knowing when to step in, what context to carry forward, and when a human should take the wheel.
🧠 THE DEEP DIVE: Proactive AI Is Starting To Show Up Before The Customer Complains
This is the big idea from Decagon’s latest move, and it is worth paying attention to. The company is pushing “proactive agents,” which is another way of saying AI that does not just sit there waiting for a ticket. It watches for signals, remembers context, and reaches out earlier in the journey.
That matters because most customer pain does not start when someone opens chat or calls support. It starts before that. The package is late. The payment fails. The appointment changes. The product setup gets confusing. By the time the contact happens, the customer is already annoyed.
A few things stand out here.
First, memory matters more than speed now.
We all love fast service. But you know what customers love even more? Not having to explain the same thing twice. If someone already told your bot, your app, or your agent what went wrong, the next step should pick up from there. That is the real promise here. Not faster replies. Better continuity.
Second, the best service moment may be the one that never becomes a support case.
This is the part CX teams should care about. If AI can flag a likely issue and step in early with something useful, that is not just efficiency. That is effort reduction. And effort reduction is where loyalty gets protected. A clean heads-up before a customer has to chase you is worth a lot more than a polite apology after they waited on hold.
Third, proactive outreach is easy to mess up.
Let’s be honest. Brands are not always great at timing. A well-timed message feels like service. A badly timed message feels like surveillance with a smile on it. That is why the trigger logic matters so much. Why are we reaching out? What exactly are we helping with? Would a customer immediately understand why they got this message? If the answer is no, the whole thing gets shaky.
Fourth, this raises the bar on every handoff.
The moment a brand says it can remember and guide, customers expect the whole journey to work that way. They expect chat to know what email said. They expect the live agent to see the issue without starting over. They expect the follow-up to match the problem they actually had. Once you promise continuity, you have to deliver continuity.
Fifth, the vision is strong, but the proof is not there yet.
This is still a vendor announcement. No hard metrics were shared on conversion, containment, CSAT, or FCR. So let’s not act like this is proven magic. Still, the direction is clear, and it is the right direction. The industry is moving upstream, into the journey before the complaint.
My take:
This is where AI gets useful in a very unsexy way. Not by sounding smarter. By cutting one dumb step out of the customer experience. That is the game now. Remove friction. Preserve context. Show up at the right moment. Then get out of the way.
What CX leaders should do this week:
Pick one journey where repetition or late response is making customers work too hard. Returns are a good one. Delivery issues are another. Same with onboarding, payment recovery, or appointment changes. Then pressure-test the moment. What should trigger outreach? What context should carry over? What should the customer never have to repeat? And where does the human step in before trust drops?
Source: Decagon
📊 CX BY THE NUMBERS: Digital Friction Is Still Wrecking Good Intentions
Data Source: Contentsquare Digital Experience Benchmark 2026
Digital Experience Benchmark 2026
Traffic fell 3.8% year over year. That means fewer chances to get the journey right.
Engagement dropped 10%. Customers are giving brands less time before they tap out.
Conversion rate fell 5.1%. The journey is not just getting judged faster. It is getting punished faster.
Retention sat at 13%. Getting people in the door is still easier than getting them to come back.
The Insight:
This is the part too many teams miss. Customers are not sitting around waiting for a brand to sort itself out. They are bouncing faster, converting less, and coming back less often. So when people talk about AI improving CX, this is the test: did it make the journey clearer, quicker, and easier, or did it just add more stuff?
🧰 THE AI TOOLBOX: Trusted Outbound Voice
The Tool: AVOXI Trusted Outbound Voice helps brands improve outbound answer rates by using AI to manage caller identity and protect number reputation.
Problem: Customers do not answer calls from numbers they do not trust, even when the reason for the call actually matters.
Solution: Picture a delivery exception, fraud alert, payment problem, or last-minute appointment change. The brand needs to reach the customer now, not six hours from now after three missed calls. This tool helps choose the right number identity for the call, which can improve answer rates and make the outreach feel more legitimate from the first ring.
Benefits:
Time: Cuts down on retries and wasted outbound attempts.
Quality: Gives urgent service calls a better shot at being answered.
Experience: Makes the outreach feel more credible when timing matters most.
Where it sits: Front stage.
Best Fit:
Works best when outbound voice is part of service recovery, reminders, collections, or fraud prevention.
Not a great fit when the real issue is poor targeting, weak message design, or no clear reason for contact.
Key Takeaway: Use it to help customers trust the call, not to rescue a bad outreach strategy.
Source: AVOXI
⚡ SPEED ROUND: Quick Hits
OpenAI to acquire Promptfoo — This is a backstage story with front-stage consequences because any AI touching refunds, account changes, or service workflows needs stronger testing before it reaches customers.
Copilot Cowork: A new way of getting work done — The CX angle here is post-contact work, where promises often fall apart because escalations, follow-up, and exception handling still live in messy internal workflows.
CallMiner Delivers Breakthrough AI Advancements to Accelerate CX Automation — More flexible classifiers and summaries could help teams spot journey failures faster, especially when they need to understand why customers are contacting them in the first place.
📡 THE SIGNAL: Customers Do Not Care About Your Stack
Here is the truth underneath all of this. Customers do not care how advanced your AI is. They care whether the next step is easier than the last one. That is it. The practical move this week is to stop talking about AI in the abstract and pick one ugly journey moment to fix all the way through. Then ask your team a simple question: where are we still making customers repeat themselves, wait too long, or wonder what happens next?
See you tomorrow.
👥 Share This Issue
If this issue sharpened your thinking about AI in CX, share it with a colleague in customer service, digital operations, or transformation. Alignment builds advantage.
📬 Feedback & Ideas
What’s the biggest AI friction point inside your CX organization right now? Reply in one sentence — I’ll pull real-world examples into future issues.