Will Messina had the kind of idea that sounds sharpest while you are broke in college. He used a debit card. The person behind him could put the same purchase on an American Express card and collect points. Why not let shoppers pay online straight from a bank account, bypass some card economics, and share the savings? GrailPay began as that consumer pay-by-bank proposition. It was neat. It was cheaper. And, in the form Messina was selling, the market mostly shrugged.
E-commerce merchants did not feel an urgent need to save one or two percentage points if a new checkout method did not also improve conversion. Consumers had not been waiting for another way to type bank details into a purchase flow. Customer-acquisition costs climbed, monetization resisted, and every sales conversation felt like a shove. GrailPay had built a fair amount of infrastructure just to learn that its thesis was wrong.
“We were selling a solution to a problem people hadn’t decided they had yet.”Will Messina, co-founder and CEO
That sentence is the useful beginning of the company, not an embarrassing preface to edit out. In 2024, GrailPay stopped trying to make consumer checkout happen and turned toward B2B payment risk. The pivot did not come from a brainstorming retreat or a tour of fashionable fintech conferences. Messina went where ordinary businesses complained about money: SmileCon for dentists, veterinary gatherings and physical-security trade shows.
The bug report hiding at a dental conference
The operators he met were not asking for a lower-cost checkout button. They were asking why ACH payments returned without warning, why fraud appeared too late to prevent, and why collections and operations teams had to mop up the consequences. GrailPay changed its question from “How do we make bank payments more attractive to consumers?” to “How do we know whether this business payment will work before we send it?”
That sounds like a modest reframe. Technically, it changes the product. Commercially, it changes the buyer. The old GrailPay needed consumers to adopt a new habit and merchants to care about a small fee advantage. The new GrailPay sells to payment processors, financial institutions, fintechs and software platforms whose existing bank-payment volume already creates failed-payment losses, manual reviews and support tickets. The budget and the pain live in the same place.
When the pitch keeps needing explanation, visit customers outside your industry’s conference circuit. Ask what failure they reconcile every week, what spreadsheet exists because software does not, and which loss already has a budget attached.
The processor was the accidental moat
Many bank-account tools begin with outside data. GrailPay took the expensive route: it built an ACH processing operation. That taught the team how money actually moved from payer to payee, through onboarding, debits, payouts, settlement, returns, refunds and clawbacks. More importantly, processing produced first-party outcomes. A payment cleared, returned or raised a flag. Each result could sharpen a model.
This is what survived the consumer failure. The infrastructure was no longer just plumbing for an unwanted checkout experience. It became a machine for collecting ground truth about bank payments. GrailPay combines those outcomes with external data and dedicated machine-learning models. Its pitch is not merely “this account exists.” It is closer to “given the account, identity, history, amount and context, how likely is this payment to settle?”
A decision before movement
Two questions, not one green check
GrailPay now packages the answer in layers. Account Intelligence validates an account and routing number without requiring the user to log in to a bank. It looks at structure, ownership and historical signals, then returns a confidence score. The company says version three covers about 99 percent of US bank accounts and can deliver very high confidence on its strongest scores. Open Banking adds live balance, ownership and transaction visibility when a connected-account experience makes sense.
Transaction Intelligence moves from the account to the event. It evaluates identity, fraud, credit, industry, payment amount and situational context. A platform can use its score to approve, reject or route a payment. For qualifying transactions, Guaranteed ACH adds dynamic indemnification - GrailPay can stand behind settlement. Money movement sits alongside the risk layer: standard and faster ACH, FedNow-enabled instant payouts, merchant onboarding, configurable payout timing, webhooks and refunds are exposed through APIs.
The commercial model follows the embedding. GrailPay sells infrastructure to businesses rather than a consumer app, and each client’s payment program is configured around its risk tolerance, origination limits and settlement rules. Partners can set transaction fees and offer guaranteed payments as a product of their own. That makes GrailPay both a vendor and an ingredient: its customer owns the interface and relationship, while GrailPay earns from the risk and movement layer underneath. The arrangement has a practical appeal for vertical-software companies. They can add bank payments, earn payments revenue and avoid hiring separate teams to stitch together verification, compliance, underwriting and operations. It also explains the company’s emphasis on close implementation support. A generic score is less useful than one calibrated to the customer’s payment mix and appetite for returns.
The old answer
Micro-deposits, prenotes and static database checks can confirm format or past status. A return code explains the failure after it happens.
GrailPay’s wager
Score the account and the specific transaction in real time, then choose the rail, timing and guarantee before the money becomes difficult to recover.
This combined stack is the distinction. Plaid, MX, Finicity and Early Warning supply account and open-banking signals. Dwolla, Stripe, Modern Treasury, Trustly and GoCardless help move bank payments. Alloy, Sardine and SentiLink help orchestrate or detect risk. GrailPay’s proposition is that processing, first-party outcomes, credential-free validation, transaction decisioning and the payment itself belong in one feedback loop. It also integrates into Alloy, proof that a competitor in one layer can be a distribution partner in another.
The numbers after the turn
By June 2026, GrailPay said it was approaching $5 billion in annualized payment volume, up roughly twentyfold in a year. Revenue had grown more than 2.5 times since January. The company counted about 30 mid-market and enterprise customers and reported no churn. Those figures came from the company, but they are concrete enough to show what the consumer pitch never found: repeated use by buyers with a live operational problem.
The integrations make the product legible. Infinicept put GrailPay inside Launchpay, its payment-facilitator platform. Merchants can be underwritten once, see card and ACH information in one place and receive one consolidated daily payout. iStream paired the system with account verification, cash-flow inquiry and financing. Alloy made Account Intelligence available inside its risk-orchestration platform. The customer is rarely the business owner staring at GrailPay branding. It is the platform that wants the risk machinery to disappear behind its own interface.
“Their platform covers risk and money movement all in one place.”Scott Agatep, CEO of Infinicept
Faster rails send the safety net away
ACH’s slowness is annoying, but the delay has done hidden work. A return can surface. An operations analyst can review a strange pattern. Someone can stop the next payment. FedNow and RTP settle in seconds. Stablecoins and agent-run purchasing can remove intermediaries or the human pause altogether. Once money is final, detecting a bad payment tomorrow is bookkeeping, not prevention.
This is where GrailPay fits in the market: above the rail, before settlement, making an authorization decision that bank transfers historically lacked. Its $10.5 million Series A in June 2026, led by MissionOG with EJF Ventures, Counterpart Ventures, Construct Capital, Commerce Ventures and SSC Venture Partners participating, brought disclosed funding to $17.2 million. The capital is meant to expand the Payments Identity Network and move farther into transaction-level decisioning.
The model is not universal. A card-only merchant has little reason to buy it. A company with tiny ACH volume and negligible losses may not recover the cost of an enterprise integration. Credential-free account validation trades richer, consented account context for speed, while open banking adds friction and depends on connection coverage. Tailored risk thresholds also demand good outcomes data and a customer willing to tune them. GrailPay works best where volume is meaningful, failure is expensive, and seconds matter.
What a founder can actually steal
The obvious lesson is “listen to customers,” which is so broad it is nearly useless. The GrailPay version is narrower. First, get close enough to the transaction to observe outcomes, not opinions. Second, notice the work customers already perform after failure. Third, preserve the hard-won asset when the original product dies. GrailPay did not throw away its processor; it reinterpreted the processor as a data engine.
There is also a warning in the success story. Building infrastructure to test a consumer thesis is an expensive way to learn about demand. The pivot worked because the leftover system produced proprietary information and because B2B customers had losses large enough to pay for a fix. Without those conditions, sunk cost is merely sunk cost. The move to copy is not “pivot to enterprise.” It is to ask whether the failed product left behind a capability that becomes more valuable in a market with stronger urgency.
GrailPay’s team describes itself as low-ego, truth-seeking and curious, with “methodical velocity” rather than motion for its own sake. That language fits the history better than the usual startup wallpaper. The company changed its mind in public, followed a less glamorous complaint, and rebuilt around the evidence. Its future rests on a simple, testable claim: as bank payments accelerate, the valuable product is no longer the rail. It is the confidence to use it.