A few years ago, QRIS digital payments were a novelty. A shop that had the sticker by the register was the exception, a slightly progressive outlier next to the cash-only norm. Today the reverse is true: a cash-only stall selling to a customer under 40 is the exception, and it's the one losing sales, not the one being cautious. That flip happened faster than most owners expected, and it happened without most of them making a deliberate decision. It happened because customers changed first and businesses caught up under pressure.

I bring this up because the QRIS curve is a clean, recent, local case study in a pattern that repeats with every wave of business technology: the cost of adopting early is small and forgettable, while the cost of being last is customer loss you can't always win back. Owners who lived through the QRIS shift without noticing the lesson are about to relive it with AI tools, and I'd rather they see the pattern this time.

How QRIS Actually Moved

QRIS didn't spread because of a single marketing campaign or because merchants suddenly loved the idea of interoperable QR payments. It spread because Bank Indonesia mandated interoperability, e-wallet adoption among consumers had already crossed a critical mass, and merchants who resisted found themselves losing sales to the stall next door that had the sticker. The sequence matters:

  1. Infrastructure got standardized. One QR code that worked across GoPay, OVO, Dana, and bank apps removed the integration excuse merchants used to have.
  2. Consumer habit shifted quietly. People started leaving cash at home before merchants noticed, because carrying a phone was already the default.
  3. Peer pressure did the rest. Once enough stalls in a market or enough stores in a mall had QRIS, a cash-only holdout looked outdated rather than careful.

By the time most reluctant merchants adopted QRIS, they weren't making a strategic bet, they were catching up to a customer expectation that had already formed. That's the expensive way to adopt technology: after the market has already decided, when you have zero leverage and full urgency.

The Cost Asymmetry Nobody Calculates

Adopting QRIS early cost almost nothing: a registration process, a printed sticker, a small transaction fee. Adopting it late cost something much less visible but larger, lost sales from customers who simply walked past a cash-only stall to one that took digital payment, without ever complaining or explaining why. Lost sales from friction are silent. A customer who can't pay the way they want to doesn't file a complaint, they just don't come back, and you never see the transaction that didn't happen.

This asymmetry is the actual lesson: early technology adoption costs are visible and small, late adoption costs are invisible and compounding. Owners consistently overweight the visible cost of adopting early (learning a new system, a small fee, staff training) and underweight the invisible cost of adopting late (customers quietly choosing competitors). I've seen this exact miscalculation with a retail chain in Tangerang that delayed digitizing its loyalty program for two years because the upfront cost looked unnecessary, only to realize competitors had already captured the mobile-first customers who wanted point tracking without a physical card.

Applying the Pattern to AI

The same curve is playing out now with AI-assisted customer service, AI-generated content, and AI-driven back-office work. Right now, in early 2025, a business using an AI-assisted WhatsApp responder or automated invoice reconciliation is doing something a little ahead of the curve, and it costs relatively little in money and disruption to test. In two or three years, businesses without any AI-assisted operations will look the way cash-only stalls look today, not wrong exactly, just quietly behind in a way customers notice and respond to by leaving.

The specific technology will keep changing. The pattern won't. Ask yourself the same question about AI today that a cautious owner should have asked about QRIS in its early years:

  • Are my customers' expectations already shifting, even if I haven't felt direct pressure yet?
  • Is the early cost of testing this small enough that being wrong barely hurts?
  • Am I waiting for proof, or am I waiting for permission because change feels uncomfortable?

What This Means for Timing Your Bets

The framework that comes out of the QRIS pattern is simple: adopt when the downside of being early is small and reversible, not when you're forced to by customer attrition you can already see happening to competitors. That doesn't mean chasing every new tool. It means distinguishing between fads that fade (most single-purpose apps) and infrastructure shifts that don't (interoperable payments, and now AI-assisted operations). QRIS was infrastructure. Most SMEs treated it like a fad until the data proved otherwise, and paid for that delay in invisible lost sales.

For owners building out a broader plan rather than reacting feature by feature, this is exactly the kind of decision that belongs in a real technology strategy instead of being made ad hoc every time a new payment method or tool shows up. It's also worth pairing with how you handle the data these systems generate, which I cover in owning your customer data, since QRIS transaction data is itself an asset most merchants never bother to use.

Takeaway

The businesses that adopted QRIS early didn't win because they predicted the future perfectly. They won because they treated a small, reversible bet as cheap insurance against a customer shift they could already see starting. Look at what's shifting quietly among your customers right now, most likely around AI-assisted service and self-service tools, and ask whether you're testing early while it's cheap, or waiting until being last actually costs you the sale.