There’s a question every payments executive eventually has to answer: What costs more, backing the wrong innovation or moving too slowly?
Phil Bruno, Chief Strategy and Growth Officer at ACI Worldwide, doesn’t hesitate. “I would say it’s moving slowly.”
That answer only holds up if you take the rest of it seriously, though. Backing the wrong innovation isn’t a disaster, Bruno argues, as long as you’re learning from it, and reading the signals early enough to adjust or walk away. A failed pilot that tells you something true about your customers, your rails, or your risk exposure is a cheap lesson. A two-year deliberation that ends with a competitor already at scale is not.
Your battle plan never survives first contact with the enemy, he says. Plans change. That’s not a failure of planning, it’s what planning is for. The job isn’t writing a plan that turns out to be right. It’s building an organization that notices when the plan is wrong fast enough to change it.
“What costs more? Backing the wrong innovation, or moving slowly? I would say it’s moving slowly.”
— Phil Bruno, Chief Strategy and Growth Officer, ACI Worldwide
Learning fast is a governance problem, not a culture problem
Every company claims to tolerate failure. Very few have actually built the machinery to recognize it on schedule.
“There have to be clear milestones that you’re trying to achieve,” Bruno says, and these should be set before the pilot starts, tied to what it’s actually meant to prove. Milestones are what turn an experiment into a decision: whether it’s working, whether it’s time to scale, or whether it’s time to pull the plug and move on.
Without that structure, innovation portfolios don’t fail. They linger. They quietly eat budget and engineering time for quarters, sometimes years, without ever producing the one thing they were funded to produce: a clear answer.
The discipline Bruno is describing is easy to state and hard to enforce. You need to treat scale, test, or walk away as three equally legitimate outcomes of every bet, and know in advance what evidence would trigger each one.
Stablecoins aren’t waiting for proof
“Stablecoin for commerce is actually ready to scale,” Bruno says. This isn’t a thesis waiting on proof anymore. Somewhere between $300 and $400 billion in stablecoin transactions are already moving through commerce, a large share of it business-to-business, on track to hit $1 trillion this year.1
Most of that volume is business-to-business with companies moving money to solve real treasury and cross-border problems, not experimenting with a novelty. For any bank or merchant still treating stablecoins as a “watch and see” item: the risk has quietly flipped. The exposure isn’t in moving too early anymore. It’s in still waiting.
Agentic commerce: close, but not there yet
Agentic commerce is drawing real money and real usage. AI is already showing up inside commerce flows. But it needs more time before it scales, and the holdup is trust rather than technology.
The consumer picture here is more specific than either the hype or the skepticism suggests. “They certainly want the input… but they’re not ready to turn over the keys to their wallet,” Bruno says. People want an agent to help them find, compare, and recommend. They don’t yet want an agent to actually spend their money.
That line should draw the roadmap. Build the discovery and decision layer now, as that’s where the demand already is. And build the consent, spending-limit, dispute, and accountability architecture for the payments layer at the same time, because the moment trust catches up, the companies without that infrastructure will be starting from zero.
The next customer might not be human
Bruno expects a growing share of transactions to come not from a person directly, but from an agent acting on behalf of one. Almost every layer of commerce today assumes a human is on the other end: interfaces built to hold human attention, fraud models trained on human behavior, checkout flows engineered to reduce human hesitation. None of that holds when the buyer is software. Merchandising becomes machine-readable data. Conversion becomes API reliability. A brand’s carefully built loyalty has to survive an agent optimizing on criteria the brand never got to set.
Treat agentic commerce as a front-end feature, and you’ll end up rebuilding the whole stack underneath it later, under worse conditions.
Fraud doesn’t get solved. It gets managed, forever.
Fraud isn’t a project with an end date. Adversaries are a permanent, adaptive, well-funded part of the environment, and their tools improve at the same pace everyone else’s do. The only real posture is staying ahead of them, continuously.
“We have such great locks because we have such great lock pickers in the world,” Bruno says.
That has a direct consequence for both rails above. Stablecoin settlement and agent-initiated purchases each open new attack surfaces that today’s fraud controls were never built for. New rails need new defenses funded at launch, not bolted on after the first loss.
The trend that should worry every bank CEO
Bruno points to a widening gap between what banks are spending on technology and what they’re earning back, and the trajectory concerns him.
The good news first: Banks have got the message. They’re investing in speed, in modernization, in the things that actually drive revenue and engagement, for consumers and for business customers alike.
But a cost line that compounds faster than a revenue line only ends one way. The banks best positioned to survive that gap are the ones already growing at roughly twice the industry average—they have the capital to keep spending through it. Everyone else eventually has to choose between the investment and the margin.
Bruno is specific about where the underinvestment actually sits: product, services, and customer engagement—the exact areas that would generate the revenue growth needed to close the gap in the first place. Which is the real case against waiting. When technology spend grows faster than revenue, delay doesn’t preserve your options. It spends them. A bank that skips this year’s product investment doesn’t have more capital next year. It has less.
“The plan will change. It always does, on first contact. The advantage goes to whoever notices first and moves.”
— Phil Bruno, Chief Strategy and Growth Officer, ACI Worldwide
Payments are changing fast. Are you keeping up?
Get straight answers from ACI leaders on AI, stablecoins, payments routing, and fraud in The Future of Payments video series. Practical takes, no jargon, and each video takes only minutes.
Article Sources
- ACI’s estimate, based on current growth trajectory. McKinsey’s Global Payments Map (2026) puts current annualized stablecoin payment volume at roughly $400 billion, with select B2B and card-linked use cases growing 600–700%+ year-over-year. Extrapolating that growth rate, ACI expects stablecoin payment volume to approach $1 trillion this year. ↩︎
