The Customer Doesn’t See Dirty Data. They See a Bad Experience.
How leaders can build resilient customer experiences in a messy, real-world data environment
I had a simple issue: my TSA PreCheck didn’t show up on my boarding pass.
It’s a recurring problem tied to how my name was originally entered in the system - middle name formatting, exact match needed, etc. Annoying, but fixable.
Or so I thought.
I contacted support.
Three different agents.
One virtual assistant.
Each time, I had to re-explain the situation. Re-share my confirmation number. Re-live the same broken loop.
The virtual assistant didn’t carry context to the agent.
The agent didn’t have access to past chats.
And none of them could solve the issue without making me repeat every detail.
Let me be clear. This wasn’t a human failure.
It was a system failure.
A data failure.
TL;DR:
Everyone is racing to automate CX, but if your data is messy, your automation will fail.
Customers don’t just notice poor experiences, they lose trust because of them.
The companies that stand out do so because their product flows are powered by clean, connected data.
CS leaders can’t wait for perfect data. They need to lead through the mess now.
Use the “Automation-Ready” CX checklist (at the bottom) to start fixing your foundation today.
The real problem isn’t automation. It’s what you’re automating.
Everyone wants to automate customer interactions right now. Virtual agents. Chatbots. Triggered workflows. AI-powered suggestions.
But if your backend is built on fragmented, inconsistent, or outdated customer data?You’re not creating a better experience. You’re just speeding up the broken one.
And customers notice.
70% of consumers say they see a clear gap forming between companies that use AI well and those that don’t. - Zendesk CX Trends 2025
Here’s another one:
Only 42% trust businesses to use AI ethically, a sharp drop from 58% just last year. - Salesforce. State of the Connected Customer, 7th Edition. Page 10.
They’re losing patience and trust, fast.
Here’s what bad data creates:
• Customers repeating themselves to every agent and bot
• Bots that can’t remember what just happened in the same conversation
• Playbooks that never trigger because the fields don’t match
• Internal teams guessing instead of knowing
Automation doesn’t fix chaos. It just amplifies it.
Companies doing it better aren’t perfect. But they’re ready.
Let’s talk about a few.
UberEats makes smart assumptions and even smarter product decisions.
If your order is wrong or late, they don’t drop you into a chatbot immediately.
They guide you through a fast, intuitive in-app flow that handles the issue without ever requiring a human. It’s not just automation. It’s intelligent friction reduction.
The system has your order details, delivery time, item list, and driver history. You’re not re-explaining anything. You tap a few buttons and you’re done.
That’s what good customer experience looks like. Again, they aren’t perfect.
But they assume the obvious and they solve it quickly. As a customer, don’t make me work.
Tesla is another standout. Their app already knows your vehicle, your location, your charge level, your service history. When you need help, you’re starting with full context. Support is just an extension of a product experience that’s already deeply personalized.
Amazon, Ramp, Apple. Same story.
The automation works not because it’s extraordinary, but because it’s built on solid, well-connected infrastructure. And that becomes obvious when you realize how hard it is for most companies to replicate the same experience.
Consistent, centralized, clean data. And most importantly, transparent.
73% of customers say it’s important to know when they’re communicating with an AI agent, and to have a clear path to a human.³
Salesforce. State of the Connected Customer, 7th Edition. Page 19.
When that clarity is missing, trust erodes fast.
Here’s the part CS leaders need to hear and do:
You can’t sit back and wait for perfect data. You’ll be waiting forever.
Great CS teams don’t wait. They build forward.
In the meantime, you can:
Advocate for cleaner fields and consistent naming
One CS leader I know worked with RevOps to eliminate over 200 variations of the same customer account name. “Acme Corp,” “Acme Inc,” “ACME Corporation.” All unified. That one project made their entire QBR workflow 10x more accurate and allowed lifecycle triggers to fire properly.
Patch gaps with lightweight tooling, tags, or workflows
At a startup I worked with, the CS team couldn’t get engineering to prioritize building full product usage visibility. So they hacked together a simple Google Sheet and Zapier setup to tag customers based on login activity. Was it perfect? No. But it let them proactively reach out to at-risk accounts and reduce churn while they waited for a full solution.
Partner with Ops to flag duplicate or missing data
You don’t need to own the CRM to influence its integrity. Work with RevOps or SalesOps to identify where duplicates are being created (e.g., SDR-created accounts, mismatched billing records). Even a monthly “data sanity sweep” across a few key fields - account name, owner, renewal date - can drastically improve how well playbooks and reporting actually function. Make this a standing part of your CS/Ops relationship.
Document customer context in shared places
Context shouldn't live in your CSM’s head - or worse, in a private Google Doc. Use shared fields or notes in your CRM or CS platform to summarize customer goals, known risks, and success milestones. A simple 3-line narrative ("customer just went through reorg, VP of Product is new, expansion convo paused until Q3") can save hours, reduce internal miscommunication, and make transitions between CSMs seamless.
Push for the right data to be captured, not more of it
A bloated CRM full of irrelevant fields is worse than a simple one that captures the right signals. Be ruthless about identifying which data points actually drive proactive action. For example: usage frequency may matter more than number of logins. Strategic fit might trump ARR. Work cross-functionally to eliminate vanity fields and make space for the data that truly supports retention, expansion, and customer health visibility.
You don’t need a pristine data lake to build trust.
Clean, connected experiences start with consistency. Standardize the fields that matter most - like account name, owner, and plan type - so your systems speak the same language. Centralize key customer context in one place, so reps and bots aren’t digging through five tools to piece together the story. And streamline your handoffs. If context gets dropped between automation and human, your experience resets - and your customer feels it.
You need just enough structure to carry context and anticipate needs.
Start there.
And no matter how good your chatbot is, it can’t fix what your systems forgot.
Let me bring it back to that travel issue.
In the end, I spoke to four different agents, plus the chatbot.
Each time, I had to log in again using a link the agent sent me, just to re-verify the same basic details: my name, date of birth, confirmation number.
I didn’t have to remind them of my Medallion status. But that didn’t matter. The system didn’t remember me anyway.
And just to top it off, the seat number they gave me didn’t even match the aircraft type. A completely different issue, but same root cause. Systems that don’t talk to each other.
That’s not a support problem. That’s a systems problem.
And it’s exactly what happens when automation moves faster than your data strategy.
Don’t automate the chaos.
Fix the foundation first.