Comerica Bank will become an early adopter of The Clearing House’s revised rules for domestic On-Behalf-Of (OBO) payments on the RTP® network. Comerica Bank, and its client Monex USA, a leading provider of international payments, corporate FX and currency risk hedging services, took part in one of the first OBO payments under the revised rules, reinforcing Comerica’s dedication to delivering instant payment solutions designed for speed, efficiency and flexibility for both direct and indirect customers. OBO payments are RTP transactions originated by a sender to make a payment for another person (i.e., on behalf of that other person). The sender is the titled owner of the account from which the RTP payment is sent. The new rules replace existing requirements for Payment Service Providers and apply more broadly to intermediated RTP activity, introducing a consistent framework focused on payment transparency, due diligence, risk management and fraud reporting obligations. “Introducing this new capability is a testament to Comerica’s commitment to providing our customers seamless, on demand access to funds, both for themselves and their own customers,” said Allysun Fleming, Comerica Bank Executive Director of Payments. “RTP OBO payments unlock real-time disbursement use cases at scale, such as payroll and benefits, marketplace payouts, embedded payments, and more. We are excited to participate in an ecosystem that enforces transparency of funds.” A comprehensive risk management framework for OBO payments on the RTP network broadens the benefits of RTP payments through enhanced oversight of intermediary payments providers. Through the new requirements, the RTP network strengthens participants’ ability to manage risk for growing payment use cases and will increase momentum for the 24/7 real-time payment network that already processes more than one million transactions per day for more than 950 banks and credit unions. RTP OBO payments for domestic transactions will enable Monex USA to facilitate instant payments at scale, greatly simplifying their operational process and providing enhanced transparency end to end.
Citizens Bank stays with strategy of approaching open banking as an “enterprise” or business project, irrespective of the fate of the 1033 regulation
The Trump administration is doing away with a regulation designed to promote data sharing between banks and third parties. But for Citizens Financial, the move isn’t hindering its bullishness on the broader concept. “This doesn’t change our strategy,” Eric McCabe, Citizens’ head of embedded finance, told. “1033 was focused more on consumers,” McCabe said, adding Citizens has approached open banking as an “enterprise” or business project, combining people from a wide range of departments. That doesn’t directly relate to the 1033 regulation but more about the general market demand for improved data sharing, and how internal collaboration can streamline project management. “Most banks have their own staff for corporate and consumers,” he said. Citizens built its open banking application programming interface to enable clients to have secure and seamless access to their Citizens data from their external platforms of choice, McCabe said. “We then pursued our development in a manner that would comply with 1033, but it wasn’t the driving factor,” McCabe said. The bank has spent more than two years working on its open banking project and has suggested banks continue their own work on open banking and other forms of data sharing. Citizens’ open banking service enables businesses to link their Citizens’ accounts to a third-party platform through an API. Businesses use the API to support permissioned data sharing without using an older practice called screen scraping that has long raised security and privacy concerns among banks. Citizens says its API has reduced screen-scraping usage by more than 90%. “The impetus was to get away from screen scraping. Banks don’t like that. So the APIs give an alternative,” McCabe said. While no one knows exactly what will come next, open banking is here to stay, Kiernan Hines, principal banking analyst at Celent, said in a research note about the uncertainty of the 1033 rule. More than 25% of banks in the U.S. say open banking-led product innovation is one of their three leading technology investment priorities for 2025, according to Celent, though Hines noted that survey was taken before the CFPB’s recent suit. While the CFPB’s recent move to eliminate the 1033 rule will cause the development of open banking to “take a hit,” open banking will continue to grow and has already had a huge impact on financial services, Hines said. “Many financial institutions now view open banking as a genuine opportunity to sit on the other side of the value chain and enhance their own workflows and customer-facing services,” Hines said. “While the rules of the road will continue to evolve, if not formulate, the open-banking toothpaste is out of the tube; customers expect to be able to leverage account data without friction”.
Adflex’s straight-through processing tech to enable Coupa to fully automate the B2B spend management process by enabling virtual card numbers to be read automatically from within emails
Coupa has integrated Adflex’s straight-through processing service, Adflex STP, into its total spend management platform, saying this will fully automate the B2B payment process for Coupa customers and their suppliers. The Adflex STP service and virtual card reader technology enable virtual card numbers to be read automatically from within emails, allowing transactions to be processed without manual entry. This collaboration will simplify supplier acceptance and process integration, enabling suppliers to scale virtual card programs effectively, Bill Wardwell, senior vice president and general manager, Coupa Pay and Treasury, said. This new service is exclusively available through Coupa and Adflex partner Barclaycard Payments in the Europe, Middle East and Africa (EMEA) region. Adflex also offers its automated virtual card payments service to all card issuers and their corporate clients. “Adflex STP levels the playing field of digital B2B payments by delivering the benefits of prompt, secure and pain-free transactions to both buyer and supplier,” Andy Downman, commercial director at Adflex, said.
Startup AuthZed ‘s open-source permissions system can scale to trillions of access control lists and millions of authorization checks per second to support RAG and agentic AI systems with real-time permissioning
Permissions management startup AuthZed announced new support for retrieval-augmented generation and agentic artificial intelligence systems, expanding its authorization infrastructure to address security challenges in enterprise AI. The expanded support is designed to give engineering teams the tools they need to ensure that AI systems respect permissions, prevent data leaks and maintain complete audit trails. AuthZed uses its open-source permissions system, SpiceDB, to support RAG and agentic AI. SpiceDB, based on Google’s internal permission system, Zanzibar, was built for scale and complexity and can scale to trillions of access control lists and millions of authorization checks per second. AuthZed says that supporting AI is a natural evolution for the system. AuthZed ensures that only authorized data is retrieved, embedded and displayed to users throughout the RAG process. Using AuthZed, teams can enforce access control by filtering documents before embedding them, post-filtering vector search results to exclude restricted content, and synchronizing permissions in real time with platforms such as Google Workspace and SharePoint. The controls allow organizations to build secure, high-performance RAG systems that minimize the risk of data leaks. On the agentic AI front, AuthZed’s Agentic AI Authorization Model is designed to manage what agents can do by aligning their capabilities with the permissions of the users they act on behalf of.
Klarna allows callers to provide feedback and suggestions about its products to “AI-Sebastian,” who has been trained using the real CEO’s voice, personal insights and experiences
Klarna has introduced an AI-powered hotline, enabling customers to engage directly with a digital version of CEO Sebastian Siemiatkowski for product feedback. This initiative places technology at the centre of customer communication, enabling a more streamlined and efficient feedback process.This innovative approach allows callers to provide feedback and suggestions about Klarna’s products to “AI-Sebastian,” who has been trained using the real CEO’s voice, personal insights and experiences. The deployment of this technology aims to transform the way customer input is gathered and integrated into product development. The new service is initially rolling out in Sweden and the US. By leveraging advanced AI, Klarna’s customers can now submit ideas, ask questions and provide feedback simply through a regular phone call. This eliminates the need for traditional, often cumbersome methods of data collection. Each conversation is analysed and transcribed in real time using AI. This ensures that all customer interactions are captured and processed efficiently, providing immediate insights into customer sentiments and requirements. Upon completion of a call, the technology generates a summary which is immediately forwarded to an internal information flow, where it is assigned to Klarna’s product and development teams. This rapid dissemination of information ensures that relevant departments can quickly address customer feedback. Customer feedback, depending on the nature of the issue or suggestion, can lead to tangible product improvements within 24 hours. This rapid response capability highlights the agility and responsiveness enabled by integrating AI into the customer feedback loop. Klarna seeks to challenge conventional standards by offering customers a faster and more interactive way to contribute to the improvement of its products and services. The AI-driven hotline is designed to overcome the limitations of traditional feedback methods, ensuring a more engaging and insightful customer experience. The introduction of the direct line to “AI-Sebastian” builds on Klarna’s ongoing efforts to leverage AI for increased efficiency and improved customer experience. Klarna’s AI chatbot now handles two-thirds of all chats, representing over 1.3 million customer cases each month, and the work of 800 full-time employees. Simultaneously, the average handling time has decreased from 12 minutes to under two, while maintaining customer satisfaction. The company has recently replaced over 1,200 external SaaS solutions with a proprietary technical infrastructure. By implementing AI throughout the organisation, revenue per employee has increased by 152%, approaching $1m per employee per year.
Autobook integrates Fundbox’s embedded capital infrastructure tech into its platform to enable small business owners to apply for and access funds without ever leaving their banking app
Autobook has launched Autobooks Capital, powered by Fundbox, the embedded capital infrastructure for small businesses. This new offering adds fast, flexible funding directly within the Autobooks platform—no redirects, no extra accounts—just seamless access to capital where businesses already manage their finances. By layering in embedded capital infrastructure from Fundbox, financial institutions can offer small businesses a seamless way to access working capital, right when and where it’s needed. Because Autobooks Capital is embedded directly within digital banking, small business owners can apply for and access funds without ever leaving their banking app. This direct-to-account experience simplifies cash flow management and reinforces the financial institution’s position as the primary operating hub for small businesses. Fully integrated into the Autobooks platform, Autobooks Capital offers fast underwriting, competitive rates, and flexible repayment options—allowing businesses to apply for and receive funding without ever leaving the platform. Whether it’s restocking inventory, expanding operations, or navigating cash flow challenges, capital is now just a few clicks away. Autobooks Capital compliments traditional lending programs by giving business owners convenient access to short-term working capital for everyday needs. This enables financial institutions to better retain primacy of the customer relationship and compete more effectively with online lending providers — delivering modern capital access without pushing business customers to third-party platforms.
Applicant trust platform Snappt offers automated, reliable verification of an applicant’s rent payment history delivering 25x more coverage than credit bureaus and over 80% verification success
Snappt, a leading platform for applicant trust in multifamily housing, has announced the addition of Verification of Rent (VOR) powered by Trigo. The platform also includes enhanced Verification of Assets (VOA) and bank account linking, supported by a new partnership with Mastercard’s open banking platform, Finicity. VOR enables property managers to automatically and accurately verify applicant rental payment history, eliminating the need for manual outreach to landlords. With less than 5% of rental history available through traditional credit reporting, VOR delivers 25x more coverage than credit bureaus and achieves over 80% verification success. Key features of Snappt’s Applicant Trust Platform: Verification of Rent (VOR): Automated, reliable verification of an applicant’s rent payment history. Bank Account Linking: Reduce friction for applicants by offering the option of instant verification and speed up application processing. Verification of Assets (VOA): Approve creditworthy applicants with significant assets who may not meet traditional income requirements, directly improving occupancy rates and NOI. Industry-Leading Fraud Detection: Snappt’s proven AI-driven detection technology and fraud forensics team have analyzed over 13 million documents, ensuring accurate results. Connected Payroll: Direct integration with 90% of US payroll providers instantly verifies income and employment status in real-time. ID Verification: Best-in-class biometric technology, complete with 30+ checks on an ID and the ability to scan 4,600+ global ID types.
Postman’s platform improves API discoverability by enabling any public API on its network of over 100,000 public APIs to be turned into an MCP server, with a verified domain, auth controls and good documentation
API discoverability has always been important, but it’s becoming increasingly more important as AI agents become more prevalent, says Abhinav Asthana, CEO and co-founder of Postman. Sterling Chin, senior developer advocate at Postman, said that the industry needs to get to a point where an API is so easy to digest that it’s just like building with LEGO. Postman launched a network for verified MCP servers. “We basically took all the remote MCP servers available, verified them, and put them on the public network because everybody’s gonna need a verified place soon.” Postman also released an update to its platform that enables any public API on its network of over 100,000 public APIs to be turned into an MCP server, making it more important than ever that API developers ensure their APIs are discoverable by the people that will want to use them. Chin said that what is typically seen of APIs is only the tip of the iceberg. “We only see the top maybe 10 percent. Those are the external APIs that get all the hype. The majority of services are internal to us, and those are the ones that when MCP starts to really take off, those are the APIs that are going to blow everyone away.” Allen Helton, ecosystem engineer at Momento, maker of reliability solutions and a customer of Postman, told that the most important benefit they get out of Postman is that it allows their APIs to be easily discovered by developers. Another recommendation is to make sure your public profile is filled out. The public profile includes everything an API publisher owns, including workspaces, collections, and API specs. He advises everyone to have a profile picture and links to their social media and website on that page. Getting verified by Postman will also help, as verified publishers get a badge that essentially proves that you’re the domain owner, increasing confidence among API consumers. Postman’s requirements for getting verified include things like having a verified domain, setting up authentication for public APIs, and having good documentation.
Frontier models are multimodal, capable of zero-shot learning, display agent-like behavior, offer real-time inference and are characterized by massive data sets, compute resources, and sophisticated architectures
You can intuitively apply the word “frontier” to know that these are the biggest and best new systems that companies are pushing. Another way to describe frontier models is as “cutting-edge” AI systems that are broad in purpose, and overall frameworks for improving AI capabilities. When asked, ChatGPT gives us three criteria – massive data sets, compute resources, and sophisticated architectures. Here are some key characteristics of frontier models to help you flush out your vision of how these models work: First, there is multimodality, where frontier models are likely to support non-text inputs and outputs – things like image, video or audio. Otherwise, they can see and hear – not just read and write. Another major characteristic is zero-shot learning, where the system is more capable with less prompting. And then there’s that agent-like behavior that has people talking about the era of “agentic AI.” If you want to play “name that model” and get specific about what companies are moving this research forward, you could say that GPT 4o from OpenAI represents one such frontier model, with multi-modality and real-time inference. Or you could tout the capabilities of Gemini 1.5, which is also multimodal, with decent context. A team of experts analyzed what it takes to work in this part of the AI space and create these frontier models. The panel moderator, Peter Grabowski, introduced two related concepts for frontier models – quality versus sufficiency, and multimodality. Douwe Kiela, CEO of Contextual AI, pointed out that frontier models need a lot of resources, noting that “AI is a very resource-intensive endeavor.” “I see the cost versus quality as the frontier, and the models that actually just need to be trained on specific data, but actually the robustness of the model is there,” said Lisa Dolan, managing director of Link Ventures.
New data observability solutions are addressing the full lifecycle of AI/ML inputs as 42% of enterprises still don’t trust AI model outputs
Ataccama’s new report in partnership with BARC finds that while 58% of organizations have implemented or optimized data observability programs – systems that monitor detect, and resolve data quality and pipeline issues in real-time – 42% still say they do not trust the outputs of their AI/ML models. The findings reflect a critical shift. Adoption is no longer a barrier. Most organizations have tools in place to monitor pipelines and enforce data policies. But trust in AI remains elusive. While 85% of organizations trust their BI dashboards, only 58% say the same for their AI/ML model outputs. The gap is widening as models rely increasingly on unstructured data and inputs that traditional observability tools were never designed to monitor or validate. 51% of respondents cite skills gaps as a primary barrier to observability maturity, followed by budget constraints and lack of cross-functional alignment. But leading teams are pushing it further, embedding observability into designing, delivering, and maintaining data across domains. When observability is deeply connected to automated data quality, teams gain more than visibility: they gain confidence that the data powering their models can be trusted. The report also underscores how unstructured data is reshaping observability strategies. Kevin Petrie, Vice President at BARC said “We’re seeing a shift: leading enterprises aren’t just monitoring data; they’re addressing the full lifecycle of AI/ML inputs. That means automating quality checks, embedding governance controls into data pipelines, and adapting their processes to observe dynamic unstructured objects. This report shows that observability is evolving from a niche practice into a mainstream requirement for Responsible AI.”