Agentic AI Automation
Shopping with Instant Bank Payments at the Point of Sale and Real-Time Dispute Resolutions
Table of Contents
Instant Bank Pay vs Pay-by-Bank 3
Market Precedent: Pix in Brazil 4
Blockchain for Verified Credentials 5
Vertex AI-based Fraud Detection Service 7
Agentic AI Automated Dispute Resolution 8
Introduction
The RTP® network, owned and operated since 2017 by bank-owned The Clearing House (TCH), and the more recently launched FedNow® network, owned and operated by the Federal Reserve Board (FRB), promise to revolutionize the U.S. banking and payments industry. TCH represents the sole private-sector operator originating core payments infrastructure system and services for ACH, wire transfer, and check volume in the U.S., processing $2 trillion USD daily.
The organization also serves as the forward-thinking adviser for the entire U.S. banking ecosystem. Rostering over 900 participating financial institutions (FIs) and processing an average 1.17 million transactions valued at $5.03 billion daily, TCH owned RTP network has enabled instant payments for 71%+ U.S. demand deposit accounts (DDA).
FRB established FedNow’s footprint in July 2023 as the second instant payments network supporting direct cash deposit transfers between bank accounts. Housing more than 1,300 FI members that reflects 30%+ U.S. deposit accounts, FedNow has processed over 2.8 million transactions worth more than $955 million. Cumulatively, both instant payment networks have augmented and accelerated funds transfer speed among use cases including business-to-business (B2B), business-to-consumer (B2C), consumer-to-business (C2B), peer-to-peer (P2P), account-to-account (A2A), government-to-consumers (G2C), and more: brokerage account funding, digital wallet defunding to DDAs, earned wage access and emergency payroll, instant wage access, merchant disbursements, real estate closing payments/title insurance payments, consumer loan funding, and bill pay.
To date, the predominant use case for real-time payments has been B2B bill pay, followed by gaming and gig worker payouts. The use case with the largest Total Addressable Market (TAM) is real-time C2B payments at the point of sale.
Currently, neither bank network is equipped to support C2B point of sale transactions. Lacking is an alias directory similar to Zelle where DDA information is masked, transactional fraud detection and mitigation, and dispute resolution comparable to the chargeback process used by credit card networks. These deficiencies contribute to the lack of adoption by FIs for Send and Request for Payment (RfP) personas. All FI participants in either bank network must minimally support Receive.
Instant Bank Pay vs Pay-by-Bank
In today’s digital age and rapidly evolving payment landscape, the term “Pay-by-Bank” is being frequently used by many FIs and fintech companies around the globe. And while the terms “Pay-by-Bank” and “Instant Bank Pay” may often be thought of and used interchangeably, the underlying technologies and customer experiences can differ significantly. It’s crucial to understand the difference, as they’re not the same.
Traditional Pay-by-Bank offerings frequently rely on ACH (Automated Clearing House) rails — a decades-old system designed for batch processing. While some providers offer “same-day ACH,” these transactions still settle only during regular business hours and take 1 to 3 days to fully process and post to the recipient’s account (NACHA). In contrast, Instant Bank Pay utilizes modern real-time payment rails, such as RTP Network or FedNow, to settle transactions within seconds, 24/7/365, providing instant execution and irrevocable settlement. Understanding this distinction is crucial in today’s unyielding economy, where consumers and businesses expect and require seamless and instantaneous transactions.
The distinctions don’t end there. The difference also extends to payment risk and user experience. ACH-based Pay-by-Bank systems lack the built-in fraud protections, transparency, and finality that real-time systems may offer. Because ACH transactions can be reversed or delayed, they carry a more significant risk of fraud exposure and uncertainty, especially for merchant businesses. On the other hand, however, Instant Bank Pay includes irrevocable settlement and is augmented by real-time fraud detection, account verification, and immediate confirmation. Sionic takes this a step further by embedding its AI-powered Fraud Detection System, modeled with Slalom and built on Google Cloud’s Vertex AI, along with agentic AI automated dispute resolution tools to ensure transactions are not only fast but also more secure, trustworthy, and fully supported from initiation to settlement.
Sionic’s Instant Bank Pay is more than just another Pay-by-Bank solution; it’s a reimagined infrastructure for digital payments. By leveraging real-time rails and solving critical limitations in alias identification, fraud detection, and post-payment resolution, Sionic can deliver a future-proof, friction-reduced, and merchant-friendly experience. As real-time payments grow globally — led by systems like Pix in Brazil and UPI in India — Sionic’s Instant Bank Pay positions U.S. FIs and merchants to participate in a similar leap forward, without being held back by legacy systems disguised as innovation.
Market Precedent: Pix in Brazil
Pix, the instant payment system created and operated by the Central Bank of Brazil (BCB), has become overwhelmingly embraced by natives since being established in 2020. Having mandated participation of BCB authorized issuing institutions that have greater than or equal to 500 thousand transaction accounts, or approximately 1,000 institutions, Pix was officially named the alternative payment method most used in Brazil, officially out-reigning previous leaderboard competing debit cards (69.1%) and cash (68.9%) at 76.4%, as of March 2025. According to ACI Worldwide, since its launch in November 2020 by the Central Bank of Brazil, Pix has experienced extensive growth, processing over 37 billion transactions in 2023, making Brazil the world’s second-largest real-time payments market after India.
Similar to Know Your Customer (KYC) identifiers, “Pix Keys” employ user attributes, such as phone number, email, and government-issued IDs, to enroll users and initiate payments. Presentment of BR Codes, Brazilian branded dynamic and static QR codes, are 32% and 8%, respectively, of all transactions in May 2025. As of June 2025, the Brazilian payments infrastructure has expanded real-time transfer capabilities with Pix Automático, or Automatic Pix — the feature that enables processing for recurring payments such as utility bills, subscriptions and streaming — with market-designated “Pioneers”: Amazon, Globoplay, Grupo OLX, Cemig, Disney+, Netflix, Spotify, and Visão Prev. Further augmenting the platform with such a practical feature has resulted in an overall increase in adoption and transaction volume. Pix now heavyweights 182+ million Pix users, reflecting nearly 88% of the adult population, provisioning 817 million Pix keys linked to more than 547 million accounts, and 19 million business participants. Merchant payments now represent over 40% of nearly 6 billion transactions per month:
It’s worth noting that the evolving nature of ubiquitous instant payments has a different orientation in Brazil compared to the U.S. For starters, US-based FIs are not mandated to join either instant payment network (the RTP network or FedNow), ultimately leaving a portion of end users underserved and underconditioned for evolving payment innovations and resources. Additionally, Pix is a centralized, real-time payment system that supports instant transfers across the entire Brazilian national landscape. Creating a more decentralized infrastructure for instant payments support, the RTP network, a private-sector operator, and FedNow, a public-sector operator, are two distinct entities facilitating faster payments. Moreover, because the systems are not interoperable with one another, FIs must subscribe to each network to gain access and transfer funds to their respective in-network participants. Unlike Pix, the two US-based real-time payment networks also do not currently support C2B payment capabilities, leaving consumers to resort to conventional payment methods for everyday spending and adjacent commerce. Among other differences, the three aforementioned highlight tremendous gaps impacting real-time payment acceleration in the U.S.
Blockchain for Verified Credentials
Instant payments require near-instant identity verification. Blockchain-based Verified Credentials (VCs) offer a tamper-proof, decentralized method for proving identity. Verified Credentials embed trusted digital identities into real-time payment systems. As highlighted in an article by Michelangelo Frigo for Zyphe, “How Blockchain Technology is Revolutionizing Identity Verification,” issuing digitally signed credentials stored in user wallets enables banks and payment providers to instantly authenticate both payers and payees through cryptographic proofs, thereby eliminating the need for manual verification of identity documents. This streamlines onboarding, reduces friction, and cuts fraud risk, while enabling seamless reuse of credentials across institutions.
Such approaches have been explored by international trade financiers, such as HSBC, which has piloted blockchain trade finance platforms like Voltron and Contour, reducing letter of credit processing from 5–10 days to under 24 hours and enhancing cross-border transparency. Additionally, HSBC is prototyping decentralized identity (DID) solutions using Polygon ID, which enables the issuance of cryptographically verifiable credentials during the digital onboarding process. This approach aims to replace manual identity checks with blockchain-based proofs that enhance privacy and reduce fraud risk.
At a global interoperability level, blockchain VCs and DIDs bridge jurisdictional identity gaps in instant payments. Systems such as India’s Aadhaar and the EU’s EUDI Wallet rely on these standards to ensure credentials are recognized across borders. Additionally, solutions like alongID provide orchestration among international KYC providers, allowing users to store reusable credentials in digital wallets for instant cross-border onboarding. This empowers both businesses and individuals to transact globally with minimal delay.
Take Singapore’s Project Orchid, by the Asian taxi and delivery platform Grab, for example. It exemplifies how blockchain VCs can integrate with programmable payment infrastructures. Under MAS guidance, Orchid pilots incorporate VCs to gate programmable, “purpose-bound” money — vouchers, grant tokens, and even merchant payments by verifying credential conditions before execution. This adaptive compliance model embeds AML/KYC rules directly into payment flows, validating required credentials instantly via blockchain and smart contracts, enabling compliant, automated instant payments.
Blockchain for VCs is not yet ubiquitous, but it’s quickly becoming a critical enabler for trust, compliance, and interoperability in global instant payment networks. It reduces friction, prevents fraud, and lays the groundwork for next-gen digital payment ecosystems. Looking forward, these systems promise robust fraud prevention, real-time compliance, and frictionless cross-border transfers. However, widespread deployment hinges on the evolution of standards for VCs and DIDs, coordination with regulators, and privacy-preserving mechanisms for revocation and consent. As various markets mature and evolve — Brazil’s Pix, India’s UPI, EU’s EUDI, and Singapore’s Orchid — all driving toward VC integration, we’re likely to see blockchain-anchored credentials become foundational to next-generation instant global payments.
Vertex AI-based Fraud Detection Service
Today’s instant payment systems are groundbreaking and show promise, but they’re not complete. They lack alias directories, forcing consumers to remember complex bank details rather than using simple, verified identities. They’re also missing real-time fraud detection for point of sale transactions, leaving both payers and payees exposed. To top it off, there’s often no structured way to resolve disputes, which can leave minor problems growing into major frustrations for both payers, payees, and their financial institutions.
Sionic fills those gaps by embedding industry-proven fraud detection protocols, such as anomaly scoring and behavioral analytics, into its Instant Bank Pay flow. We’ve introduced VCs and DIDs, eliminating the need to share or remember bank routing numbers and other information. We’ve partnered with Slalom and Google Cloud Professional Services Organization, leveraging Slalom’s synthetic fraud data expertise and Vertex AI from Google Cloud — to build a robust AI-powered Fraud Detection System that flags suspicious activity in real time. With the help of Large Language Model (LLM) integration through Google and Slalom, these models, enhanced by AI, are trained on both synthetic data and real-world fraud scenarios, enabling Sionic’s Instant Bank Pay system to respond in real-time to evolving threats and analyze and interpret language-based data. This, combined with a dedicated alias directory and real-time dispute resolution mechanisms, enables Sionic to deliver a fully integrated, end-to-end service.
When issues do arise, payers can immediately identify and resolve disputes directly with the merchant through the Instant Bank Pay microsite, enabling instant and transparent settlements between payers and payees via AI agents. The result is a more secure, real-time bank payment experience that goes beyond speed. Across the industry, most real-time payment systems lack these types of embedded fraud tools, creating major risk exposure.
AI Personal Shopping Agents
Emerging as another value-added feature to enhance US-based payments has been the integration of agentic AI, or designated agents performing consumer shopping and purchasing activities and business back office operations for payment transaction processing. Unlike traditional chatbots, web forms and other tools, AI agents are autonomous, being able to initiate transactions, make decisions and interact with other systems independently and without human intervention.
Visa, a global card payments processor, recently launched their Intelligent Commerce platform, a groundbreaking initiative that opens Visa’s payments network to developers and engineers building the first generation of true AI commerce. This platform allows consumers to authorize AI agents for routine shopping experiences that include browsing, selecting and purchasing items based on owner defined preferences, budgets and restrictions.
Sionic is offering payer-controlled AI shopping agents to deliver “best-value” shopping engagements. Consumers may add an AI Shopping Agent to their VC in their Bank Perks pass inside Apple® or Google® wallets by specifying explicit permissions to perform specific actions, operating within a defined scope, context, and timeframe. Upon setup completion, the AI agent is cryptographically bound to the VC.
AI Shopping Agents may then be used to autonomously place orders for goods or services, whether online or in-store and submit Instant Bank Pay transactions for payer approval.
Agentic AI Automated Dispute Resolution
In the realm of dispute resolution, AI agents can proactively manage the end-to-end dispute life cycle, automating case initiation, documentation, workflow orchestration, and even decision-making recommendations. The shift from static, rules-based automation to adaptive, learning-driven agents enables organizations to resolve disputes more efficiently and accurately. Leading innovators like HighRadius, Host Merchant Services, and Celent are already demonstrating how agentic AI is transforming dispute resolution workflows across finance, payments, and insurance sectors.
HighRadius, a leader in finance automation software, supports that agentic AI can significantly transform deductions and dispute management. It enables two- or three-way matching between invoices and purchase orders to detect potential disputes before they escalate, while also automating tasks like claim backup aggregation, case creation, and routing. According to Host Merchant Services, agentic AI delivers substantial benefits by autonomously analyzing chargeback disputes. These systems can gather relevant evidence, assess transaction anomalies, and generate resolution recommendations for human reviewers while reducing costs, accelerating resolution, minimizing human errors, and maximizing security and compliance. Finally, Celent highlights how agentic AI is revolutionizing insurance claims by automating data collection, validation, and settlement proposal processes, thereby reducing resolution times from weeks to hours. This technology can also enhance compliance, reduce fraud, and improve customer experience, showcasing AI’s transformative potential across industries.
Sionic is redefining how agentic AI can automate dispute resolutions between payers and payees, empowering intelligent agents to autonomously mediate and resolve issues between payers and payees in real time. Our automated payment resolution service targets fraud claims and errors, leaving product or service quality and non-payment disputes to be managed through traditional returns and refunds processes and procedures. FIs participating in our Instant Bank Pay service enjoy the benefit of AI- generated dual governance (Reg E and UCC 4A) compliance reporting with human handoff at completion, minimizing FTE overhead.
Our agentic AI platform enables autonomous dispute resolution at the point of sale, structured around four operational phases: Perceive, Reason, Act, and Learn. This approach ensures transparency, compliance, and alignment with the policies of both merchants and FIs.
Here’s how we’re defining and creating guardrails for autonomous dispute resolution processes:
- Perceive: AI agents collect data from Sionic systems — transaction details, payer history, merchant policies, and prior resolutions.
- Reason: Using LLMs and retrieval-augmented generation (RAG), agents analyze disputes and determine appropriate resolution paths based on contextual data and business rules.
- Act: Through ULink® APIs, agents carry out actions such as issuing refunds or escalating cases, aligned with merchant-defined logic.
- Learn: Resolution outcomes are captured and fed back into the system, refining future decisions and providing visibility to merchants, FIs, and customers.
Dispute outcomes follow the payee’s refund policy, which can include rules-based parameters for instant refunds, such as:
- Date of purchase (i.e., within 30 days from payment)
- Sale item (final)
- Payment amount
- Reason
- Payer history/loyalty
- TrustMe® score
FIs maintain real-time visibility into disputed transactions and can intervene based on defined escalation protocols. This AI-driven framework ensures fast, fair, and policy-aligned resolutions at scale.
Market Profile and Summary
According to ACI Worldwide, real-time payments are rapidly reshaping global finance, with transactions increasing by 42.2% year-over-year and totaling over $266B, back in 2023. Meanwhile, across the globe, real-time payment volume is projected to grow at a 36.4% CAGR, reaching nearly $169.32 billion by 2029, as stated in a report by The Business Research Company. The surge in real-time payments is not just about speed, instant clearing and settlement free up liquidity, empowering both consumers and businesses with immediate transaction clarity, and dramatically reduces processing costs, according to a study by Mastercard. Additionally, real-time payment systems like FedNow are a game-changer for small businesses, delivering instant access to cash, lower fees, and greater control over cash flow. By eliminating costly delays and card network fees, these systems enable the kind of agility small businesses need to grow quickly and compete more effectively.
The U.S. real-time payments market is on the rise, with RTP and FedNow introducing critical infrastructure. Still, current systems lack support for consumer-to-business (C2B) point-of-sale transactions, alias-based identity, and embedded fraud and dispute resolution. This leaves FIs and merchants without the tools to fully capitalize on the speed and transparency these networks promise. By contrast, global systems like Brazil’s Pix and India’s UPI have shown how real-time payments can rapidly scale when combined with the right ecosystem enablers, unlocking massive economic value.
Sionic is building that missing layer. Our Instant Bank Pay platform enhances real-time rails with AI-powered fraud detection, blockchain-based verified credentials, and autonomous dispute resolution — all designed for frictionless, more secure, alias-based payments at scale. Built on Google Cloud’s Vertex AI and developed in partnership with Slalom, Sionic’s infrastructure is API-first, highly adaptable, and purpose-built for modern use cases, from retail to gig economy payouts. As market adoption accelerates, we’re positioned to lead with a platform that closes the gaps others can’t, unlocking new revenue models and setting a new standard for intelligent, real-time payments.
NOTE: All references cited in this document are publicly available and the property of the respective authors or publishing entities. All registered trademarks and service marks are the property of their respective owners.
