A recent trip (actually, several trips) to Costco made this painfully clear.

Three Systems, One Customer… and a Lot of Friction

My Costco membership is tied to a credit card that lives across three systems:

• The Citi credit card company
• The entry scanners at the front of the warehouse
• The checkout terminals inside the store and gas station

On paper, this is straightforward. One membership. One card. Three systems talking to each other so I can buy toilet paper in bulk and move on with my day.

In practice, it’s anything but straightforward.

Some days I can walk into the warehouse without a hitch, only to find that the checkout system refuses my membership-linked card. Other days I can buy gas without issue, but the entry system insists I’m not an active member and refuses to let me in.

Same membership. Same card. Same customer. Completely different outcome depending on which system happens to be in the loop that day.

To Costco’s credit, their customer service has been consistently kind and patient. I’ve had:

• 5 phone calls
• 3 in-person membership service visits

Everyone I’ve dealt with has tried to help. But the underlying issue persists, because it’s not a “frontline staff” problem. It’s a systems and governance problem.

And that’s exactly why this matters for AI.

This Isn’t About AI. That’s the Problem.

It’s tempting to look at a situation like this and think, “Once we have better AI, we’ll fix it.” But AI won’t magically reconcile mismatched systems, unclear ownership, or inconsistent data. In fact, it usually amplifies the mess that’s already there.

Underneath my Costco saga is a familiar set of gaps:

• No single, reliable view of the customer (me) across systems
• Unclear data ownership and accountability
• Flaky integration between systems that should be tightly connected
• Limited visibility into where the breakdown actually occurs

That’s not an AI readiness issue. That’s basic digital and data maturity.

If we can’t reliably connect a membership and a credit card across three systems, what happens when we:

• Add AI assistants into customer support flows?
• Use predictive models to determine “high-value” customers?
• Rely on automated systems to approve, deny, or flag suspicious transactions?

Without solid foundations, AI will happily make confident decisions on top of fractured data and inconsistent logic. And your customers will feel that—not just as inconvenience, but as a lack of reliability and trust.

The Foundations Everyone Wants to Skip

When I work with organizations on digital, data, and AI governance, the “boring” work is usually the work that actually makes everything else go smoothly.

That includes things like:

1. Clear user personas and journeys
Who are your core users, really? How do they move across your channels and systems? In the Costco example, “member who uses a membership-linked credit card across store, gas, and online” is a simple but critical journey that should be rock solid.

2. Well-defined use cases
What exactly should happen when a member taps their card at the entry, the pump, or the checkout? What are the rules, dependencies, and exceptions? These should be designed intentionally—not discovered by frustrated customers.

3. Mapped data flows
Where does the membership data come from? Where does it go? How is it updated, and how quickly? Which system is the source of truth? This isn’t glamorous work, but without it, “single customer view” is just a slide, not a reality.

4. Effective digital governance
Who owns which systems and policies? Who is accountable for fixing cross-system issues? What is the escalation path when something goes wrong? Governance isn’t just committees and documents! It’s clarity about who does what, and how decisions get made.

These are the same foundations that support data provenance and AI governance: knowing where data comes from, how it’s used, and who is responsible for it at each step.

Why This Matters Long Before AI Enters the Room

Customer trust rarely evaporates because of a single catastrophic event. It usually erodes slowly, through repeated small failures:

• A system that doesn’t recognize an active membership
• A payment method that works sometimes, but not always
• A support experience where kind people can’t fix systemic issues

By the time an organization is ready to roll out AI chatbots, decisioning engines, personalized offers, the relationship may already be fragile. Customers won’t care how advanced your models are if the basics don’t work.

And from a risk perspective, weak foundations are even more dangerous in an AI-powered world:

• Incomplete or inconsistent data feeds models that confidently draw the wrong conclusions.
• Poorly governed systems make it hard to explain or audit AI-driven decisions.
• Lack of clarity around data ownership and lineage makes regulatory compliance much harder.

AI doesn’t fix these problems. It multiplies them.

Getting Ready for AI Means Getting the Everyday Right

So, what does real AI readiness look like?

It’s not just buying tools or standing up a lab. It’s quietly, methodically making sure that:

• Core customer journeys are reliable and repeatable
• Systems that “should” talk to each other actually do, and do so consistently
• Data flows are understood, documented, and managed
• Governance is in place so that issues can be traced, owned, and resolved

Only then does it make sense to layer on more sophisticated capabilities: AI-driven personalization, automated decision-making, advanced analytics, or whatever comes next, be it quantum, XR, agents, or something we have not thought of yet.

If the membership-card-plus-three-systems problem is still unsolved, that’s a signal to slow down and invest in the basics before racing ahead.

The Quiet Test of Trust

Customer loyalty is built (or broken) in the small touchpoints: the door scan that works every time, the payment that never fails, the support interaction that doesn’t just empathize, but actually solves the problem.

Strong governance is what makes those touchpoints work. It’s also what determines whether an organization is genuinely ready for AI or just hoping that new technology will cover up old cracks.

If your organization is thinking about AI, and most are, start with a simple question: Are we consistently getting the everyday experiences right?

If the honest answer is “not really,” that’s not a reason to give up on AI. It’s an invitation to do the foundational work first, so that when you do adopt AI (and other emerging technologies), you’re building on something solid.

Because if three systems can’t reliably agree that I’m a member, it’s not just a Costco problem. It’s a preview of what happens when we skip the basics and expect AI to save us.