Tandem is taking a major step forward with the launch of Connect by Tandem – a next-generation lender platform designed to transform the broker experience through automation and smart technology. The introduction of Connect is a pivotal moment in this journey, redefining how brokers interact with the bank by streamlining case management and saving valuable time. Connect is a pioneering loan processing platform that allows brokers to transact seamlessly with Tandem. Tandem has always been known for speed and efficiency and Connect takes this to the next level. An extensive test phase has already demonstrated its impact, shaving an impressive 4.5 working days off processing times. Key features include: 1) Advanced API integration – Enables brokers to submit applications electronically, significantly reducing manual data entry. 2) New broker portal – A single, central hub for automated underwriting policy requirements and case tracking, document uploads and case updates. 3) Automation – Streamlines processes such as EPC discounts, AVM’s and affordability checks.4) Enhanced document management – Simplifies uploads, tracking and case notifications for brokers.
MoneyGram launches API for embedding crypto on/off-ramp functionality enabling fast, compliant cash-to-crypto and crypto-to-cash integration through a single implementation
MoneyGram has launched MoneyGram Ramps, a developer-centric API that enables fast, compliant cash-to-crypto and crypto-to-cash integration through a single implementation. With just a few lines of code – wallets, exchanges and fintech apps can instantly access the MoneyGram global cash network, making it easier than ever to embed crypto on/off-ramp functionality at scale. Built for speed, MoneyGram Ramps equips developers with tools to get started in minutes: No banking integrations required; Instant API credentials and sandbox access; Comprehensive documentation and SDKs; Live onboarding. Powered by the Stellar blockchain, Circle’s USDC and MoneyGram’s global network, MoneyGram Ramps bridges physical and digital currencies, enabling movement between cash and crypto and expanding access to the digital economy. With MoneyGram Ramps, developers can easily offer users – even those without a bank account – the ability to deposit or withdraw cash at thousands of participating MoneyGram locations worldwide. “With this launch, MoneyGram is quickly becoming the connective tissue between traditional finance and the digital economy,” said Anthony Soohoo, MoneyGram Chief Executive Officer. “We’ve taken the complexity out of integration, opening the door to seamless connection with the world’s largest cash on/off-ramp for digital wallets1.”
LLMs can still be prohibitively expensive for some, and as with all ML models, LLMs are not always accurate. There will always be use cases where leveraging an ML implementation is not the right path forward. The key considerations for AI project managers to evaluate customers’ needs for AI implementation include: The inputs and outputs required to fulfill your customer’s needs: An input is provided by the customer to your product and the output is provided by your product. So, for a Spotify ML-generated playlist (an output), inputs could include customer preferences, and ‘liked’ songs, artists and music genre. Combinations of inputs and outputs: Customer needs can vary based on whether they want the same or different output for the same or different input. The more permutations and combinations we need to replicate for inputs and outputs, at scale, the more we need to turn to ML versus rule-based systems. Patterns in inputs and outputs: Patterns in the required combinations of inputs or outputs help you decide what type of ML model you need to use for implementation. If there are patterns to the combinations of inputs and outputs (like reviewing customer anecdotes to derive a sentiment score), consider supervised or semi-supervised ML models over LLMs because they might be more cost-effective. Cost and Precision: LLM calls are not always cheap at scale and the outputs are not always precise/exact, despite fine-tuning and prompt engineering. Sometimes, you are better off with supervised models for neural networks that can classify an input using a fixed set of labels, or even rules-based systems, instead of using an LLM.
LLMs can still be prohibitively expensive for some, and as with all ML models, LLMs are not always accurate. There will always be use cases where leveraging an ML implementation is not the right path forward. The key considerations for AI project managers to evaluate customers’ needs for AI implementation include: The inputs and outputs required to fulfill your customer’s needs: An input is provided by the customer to your product and the output is provided by your product. So, for a Spotify ML-generated playlist (an output), inputs could include customer preferences, and ‘liked’ songs, artists and music genre. Combinations of inputs and outputs: Customer needs can vary based on whether they want the same or different output for the same or different input. The more permutations and combinations we need to replicate for inputs and outputs, at scale, the more we need to turn to ML versus rule-based systems. Patterns in inputs and outputs: Patterns in the required combinations of inputs or outputs help you decide what type of ML model you need to use for implementation. If there are patterns to the combinations of inputs and outputs (like reviewing customer anecdotes to derive a sentiment score), consider supervised or semi-supervised ML models over LLMs because they might be more cost-effective. Cost and Precision: LLM calls are not always cheap at scale and the outputs are not always precise/exact, despite fine-tuning and prompt engineering. Sometimes, you are better off with supervised models for neural networks that can classify an input using a fixed set of labels, or even rules-based systems, instead of using an LLM.
AI radically transforms agile software development by reducing need for multiple teams and diminishing cross-team dependencies
Agile’s focus on delivering working software frequently has evolved into continuous integration/continuous delivery practices. AI is now pushing this boundary further toward what we might call “continuous creation.” When code generation approaches real-time, the limiting factor isn’t producing code but verifying it. AI offers solutions here as well—automated testing, security scanning, and quality analysis can be AI-enhanced. AI agents can write unit tests for new code and help create end-to-end tests, improving quality guarantees. The most successful teams will master this balance between acceleration and validation, exploring more ideas, failing faster, and converging on optimal solutions more quickly—all while maintaining high quality. These transformations create opportunities to streamline traditional Scrum processes. Teams can allocate a higher percentage of their sprint to spontaneous improvements as implementing features and bug fixes with AI may be faster than the overhead of including them in sprint planning. For architecture reviews, AI can serve as your first wave of feedback—a mental sparring partner to develop ideas before presenting to a committee. The AI-written summary can be shared asynchronously, often eliminating the need for formal meetings altogether. Retrospectives should now include discussions about AI usage. The improved individual productivity allows organizations to streamline overhead processes, leading to further increases in velocity. Teams can tackle larger, more complex problem spaces, and projects that previously required multiple teams can often be handled by a single team. Cross-team dependencies—a perennial challenge in scaled agile—diminish significantly. What’s most remarkable about AI’s impact is how it reinforces rather than replaces agile’s core values.
AI meeting summarization tool Jump AI frees up about 10 hours per week for each advisors
Artificial intelligence-powered offerings in wealth management regularly hammer home one benefit they provide in particular: saved time. “Advisors are saving 10-plus hours per week on average by leveraging AI to streamline their client meeting process,” said Startup Zeplyn CEO Era Jain. “That’s about 500-plus hours per year or 20 new clients they can service per year.” These time savings are primarily spent on business development and relationship building. Solo advisor Kelly Klingaman, founder of Kelly Klingaman Financial Planning, said she wanted to utilize an AI notetaker in her business so she could be more present during client meetings. Having tried out a few AI notetaking tools so far, Klingaman said Fathom is “affordable, easy to use and dynamic” — and it saves her between five and eight hours per week. For Gregory Furer, the founder and CEO of Beratung Advisors, one of the biggest game changers has been the integration of Holistiplan tax planning software. “With AI, we can now analyze a client’s tax return and generate insights in just three minutes — a process that used to take an hour and was prone to human error,” he said. From there, Furer said they create tax modeling for clients in 20 to 30 minutes, compared to the two to three hours it used to take. He said his firm is also leveraging AI within eMoney, its financial planning software, “to instantly calculate the amount of life insurance needed to maintain client-defined success rates and goals.” “This real-time decision support enhances the accuracy and speed of our recommendations,” he said. Like Klingaman, Furer has been utilizing AI for meeting notes; he uses Jump. “As the tool continues to learn our systems and language, it could eventually save five to 10 hours per week of high-value planner time, potentially becoming our most cost-effective AI tool.” Rob Schultz, senior partner and wealth manager at NWF Advisory said he also uses Jump for meeting summarization. “The quality of the notes was significantly better than I ever wrote down during a meeting and it allows me to focus solely on the client in front of me. It saves me time in the post-meeting review, probably 30 minutes per client interaction.” Samuel Flaten, co-founder of Narrow Road Financial Planning said he mainly uses ChatGPT, which he calls a “total game-changer.” In addition to the writing assistance, Flaten said he has also trained a custom GPT with “everything I know as a CFP” to workshop ideas, stress-test strategies and pull in creative alternatives he might not have considered. Across his average weekly schedule of 20 meetings, Schultz said his use of Jump AI frees up about 10 hours. Jain of Zeplyn recommends that firms optimize their scheduling by identifying advisors who successfully use AI to save time, establishing their best practices and training or coaching other advisors.
U.S. Bank opines RfP offers the promise to drive increased instant payment adoption in the US by enhancing payer security through bank-driven authentication, and reduces risk for RfP senders by eliminating returns for unauthorised or insufficient funds compared to traditional debits
What RTP and FedNow have achieved mark significant inflection points in US payments infrastructure. After the ISO 20022 deadline in November 2025, faster payments rails will be smoother with straight-through-processing (STP), streamlined reporting and compliance, and potential for enhanced interoperability. The future of US digital payments must include automated reconciliation, STP, and arguably most importantly, request for payment (RfP) if firms want to remain competitive. The RfP message type could turn real-time payments into bi-directional, eventdriven workflows that are useful for just-in-time billing for utilities for instance, or consumer initiated B2B payments through a QR code, for example. In conversation with Finextra, a representative from U.S. Bank indicated that RfP “has the promise to drive increased instant payment adoption in the US.” Returning to the point around data, they elucidated on how RfP allows businesses to send data-rich digital requests through the receiver’s bank. This initiates an immediate payment, at any time of day, any day. U.S. Bank’s view is that “RfP improves customer experiences by enhancing payer security through bank-driven authentication, and reduces risk for RfP senders by eliminating returns for unauthorised or insufficient funds compared to traditional debits.”
Intersection launches LinkNYC the largest free public Wi-Fi network to expand free and low-cost advertising opportunities – community organizations and local small businesses can now display ads on six LinkNYC screens at a time
Intersection, an experience-driven out-of-home media and technology company, and LinkNYC, the public-private partnership that created the largest free public Wi-Fi network in the world, announced the official launch of the LinkNYC Self-Service Ad Portal. Featuring expanded free offerings and a lower cost of entry for paid campaigns, advertising on LinkNYC is now easier than ever for New York City’s small business community. Intersection’s launch of the portal includes an enhancement of LinkNYC’s LinkLocal free advertising program. Eligible New York City-based community organizations and local small businesses can now display ads on six LinkNYC screens at a time—up from four—at no cost. The update also includes customizable advertising templates and colors, the ability to add a logo or QR code, and an interactive map to help advertisers select optimal LinkNYC locations. The LinkNYC Self-Service Portal is also the home of LinkDirect, Intersection’s new offering for purchasing larger, branded LinkNYC campaigns, which is designed to accommodate budgets of all sizes. LinkDirect allows advertisers to select specific screen locations and campaign dates. Users can upload custom creatives and include a QR code to boost engagement. “These upgrades to LinkLocal will provide hundreds more community-based businesses and nonprofits with opportunities to amplify their messaging through the most coveted out-of-home advertising space in New York City,” said Margaux Knee, Chief Administrative Officer of LinkNYC. “In 2024 alone, LinkNYC provided over $4 million in free advertising space.”
Capgemini unveils perpetual ‘Know-Your-Customer’ real-time continuous compliance sandbox automatically alerting firms to changes in a customer’s circumstances that could affect their risk profile, enabling them to re-assess their risk exposure to the customer
Capgemini has launched a technology sandbox to help financial institutions transition from static Know-Your-Customer (KYC) processes to perpetual KYC (pKYC) and event-based reviews. The sandbox, a first of its kind, provides a secure environment for firms to test and demonstrate the effectiveness of pKYC processes. It allows firms to automatically alert firms to changes in a customer’s circumstances that could affect their risk profile, enabling them to re-assess their risk exposure to the customer. Capgemini’s sandbox model is flexible and modular, allowing organizations to implement it across their cloud platforms and technologies. The sandbox is designed to meet regulatory requirements and demonstrate how financial institutions are mitigating inherent risk exposure more effectively. It also demonstrates the industry’s ability to demonstrate excellence in achieving real-time KYC requirements. Key benefits of Capgemini’s new pKYC sandbox include: A safe testing environment: a secure environment where new KYC processes, policies, or technologies can be tested without risking real customer data leakage or compliance failures. Best-of-breed solutions: integration of key components from best-of-breed RegTech solutions and accelerators. Real-time visualization: ability to visualize pKYC in action to gauge benefits and showcase the framework to regulators. Quantifiable business impact: rapid end-to-end testing of the tech stack and processes leading to much faster feasibility of the pKYC operating model and creation of the associated business case. Operational readiness: identifies operational bottlenecks and optimizes workflows to enable full-scale deployment with confidence.
Google Wallet deploys Zero-Knowledge Proof age verification technology uses blockchain to process the condition (age) in encrypted form, generating a proof that can be verified by an external service through public keys
Google has introduced Zero-Knowledge Proof (ZKP) technology to its Google Wallet service, allowing users to verify their age without sharing personal information. This cryptographic technology is a significant turning point for online privacy protection, as it eliminates the risk of privacy violations and identity theft. The system uses blockchain technology to process the condition (age) in encrypted form, generating a proof that can be verified by an external service through public keys. Unlike traditional methods, the ZKP system maintains total control over users’ information. Bumble, a popular dating app, will use digital IDs issued through Google Wallet to verify their age, while the confirmation will be managed through the ZKP system. This will improve the user experience and increase trust in the platform. The adoption of ZKP technology by Google could mark a decisive turning point, attracting attention from developers, companies, and investors in the decentralized privacy sector. The future of age verification and digital identity could be marked by a greater balance between security and privacy. If successful, the adoption of systems based on ZKP could lead to a safer and more respectful internet for individuals.