Top Innovations
FormFree, a mortgage fintech uses open banking to create RIKI (residual income knowledge index), ranging from 80 to 150—with scores above 100 indicating positive cash flow
A future glimpse into how…
Google DeepMind open sources its AI text watermarking tool
Google DeepMind is open-sourcing its…
Startup Novaprime debuts AI-powered loan platform to massively cut processing time- utilizing smart contracts that conform to MISMO standards and extract all the loan data
Mortgage technology startup Novaprime is…
Key News
Notary firm Proof is now offering an automated deepfake detection tool during a verification or notarization session
Remote online notarization and identity…
Socure to create a unified identity, fraud and risk decision engine and offers a “single view of identity”
Socure plans to acquire Effectiv…
Meta debuts slimmed-down Llama models for low-powered devices with a reduced memory footprint and support faster on-device inference, with greater accuracy.
Meta Platforms is striving to…
DeepMind’s Talker-Reasoner framework brings System 2 thinking to AI agents- enabling AI agents to find the right balance between different types of reasoning and provide a more fluid user experience.
Researchers at Google DeepMind introduce…
Criminals reverting to old-school tactics with new twists, Visa’s State of Scams report shows
The resurgence of physical theft:…
Meta Platforms plans to integrate a custom search engine into its Meta AI chatbot to provide users with information about current events. Currently, Meta AI relies on technology from Google LLC and Microsoft Corp. to retrieve information about current events. With an in-house search engine, there is no risk of the company’s chatbot losing access to key information because of a change in a third-party component. Google and Microsoft provide developers with access to their respective search engines through paid application programming interfaces. By switching to an in-house search engine, Meta can avoid the fees associated with those APIs. Additionally, the company can align the feature set of its search engine with its chatbot’s requirements more closely that would be possible with an off-the-shelf service. It’s unclear how Meta’s search engine will work. One possibility is that it will display a standard list of webpage links in response to queries. Alternatively, Meta may be planning to have the service output AI-generated responses that directly answer the user’s question.
Productivity-focused AI- startup Read AI has created a Gen AI-powered text, audio and video content summarization tool that enables enterprise workers to understand quickly what’s happening across all of the communication and collaboration channels they use. Its generative AI tools listen in on user’s video calls on Teams, Meet and Zoom, taking notes so they can quickly summarize what’s been said at the end of the call. Read AI also analyzes the user’s performance in meetings, then provides tips on what they can do to improve their speaking abilities, so they can be more clear and concise in future meetings. Users can connect their email and collaboration platform accounts to it, so it can read them all — just as its name suggests. Doing this, it effectively unifies all of the user’s communications, so it can provide them with a personalized, actionable briefing each day that lays out the tasks they need to prioritize. Read AI is launching a new tool called Read AI for Gmail, which transforms Gmail into a productivity hub with smart AI summaries, a feed that provides contextual insights from previous messages and meetings, and automatic responses to emails and messages for busy workers.
AI startup Sierra Technologies helps enterprises implement proactive virtual assistants for customer service roles. The startup’s technology is built on a conversational AI model that allows it to engage in natural language conversations with customers. It also has the ability to “see” and understand a company’s own business data, enabling its AI agents to be contextually aware of the ways they might be able to solve customer’s issues. By taking actions on behalf of customers, Sierra’s chatbots can reduce the instances in which calls have to be handed off to a real human agent to be solved. Besides its agentic AI capabilities, Sierra also differentiates itself through the way customers can personalize their customer service bots. Another thing that sets Sierra apart from its competitors is its reliance on a “constellation” of well-known LLMs. Though it primarily uses one model to do the heavy lifting, there’s an expectation that no single LLM is 100% reliable and accurate. So it uses secondary models as a kind of “backup,” in order to monitor the accuracy of the first and help when necessary.
Moondream has built a vision-language model that operates with just 1.6 billion parameters yet rivals the performance of models four times its size. The company’s open-source model has already captured significant attention, logging over 2 million downloads and 5,100 GitHub stars. “What makes it special is that it is one of the tiniest models that is peculiar in its high accuracy, and it works just really well,” said Jay Allen, Moondream’s CEO and former AWS tech director. “It can run everywhere really easily and quickly. It can even run on iOS, on mobile phones.” Moondream’s approach allows AI models to run locally on devices, from smartphones to industrial equipment. Retailers use it for automatic inventory management through mobile scanning. Transportation companies deploy it for vehicle inspections, while manufacturing facilities with air-gapped systems implement AI locally for quality control. Recent benchmarks show Moondream2 achieving 80.3% accuracy on VQAv2 and 64.3% on GQA — competitive with much larger models. The system’s energy efficiency impresses, with CTO Vik Korrapati noting “per token consumption is something like 0.6 joules per billion parameters.”
Armis Inc provides a device detection and protection platform for enterprises, Centrix, that automatically maps out the devices in a corporate network and scans them for cybersecurity issues. The platform can find employee endpoints, printers, servers and network equipment. After it spots a new device, Centrix searches for vulnerabilities such as unpatched firmware with a known security flaw. The platform then ranks those vulnerabilities by severity to help administrators prioritize their remediation efforts. Centrix also spots more subtle cybersecurity issues. The platform can identify situations where a device doesn’t have an antivirus installed, or does have one but can’t make full use of its features because of networking problems. Armis says that Centrix tackles the challenge with a mechanism called Smart Active Querying. According to the company, the feature collects data about devices directly without relying on network traffic analysis. It can glean what a device is used for, whether the firmware it’s running is up-to-date and other technical details.
Newly launched Trainspot, is the first transparent, open-to-all marketplace for business-ready AI training data, empowering content owners of all sizes to take control of their data in order to participate in and profit from the AI economy. Content creators and rights holders who have published content online are encouraged to register their content or data on Trainspot. Once verified, they can manage how their data is used for AI, whether to block, sell, or donate it. Buyers, from individual AI developers to large companies building AI-driven products, can browse the Trainspot directory to license content that fits their needs, ensuring their AI models are trained on legally compliant, high-quality data. The two-sided marketplace also gives AI developers, businesses, and large enterprises the ability to confidently purchase necessary training data, at scale, from sources large and small across the Web without fear of legal liability. On Trainspot, owners sell, donate, or block the use of their data. This empowers data and content owners with the control they need to protect their rights and profit from their intellectual property.
With the proliferation of generative AI expected to result in US$40 billion in fraud losses by 2027 — and that’s just in the United States — financial institutions need to step up their fraud-prevention tactics. Recent research from Jumio found that 78% of consumers in Singapore are concerned about whether their bank is doing enough to protect customers against deepfake-powered fraud, with the same number of consumers ready to switch banks over inadequate fraud protection. Frederic Ho, Jumio VP of APAC said his eKYC firm addresses the vulnerabilities of database and biometric identity verification systems to deepfake and identity theft, by implementing advanced liveness detection to counter sophisticated fraud tactics like face morphing, face swapping, and camera injection in eKYC processes. Key risk services that Jumio checks for include: global identity, Social Security number (SSN), phone number, government database, email, geo IP, address, device, and bank identification number (BIN). The Jumio Platform provides AI-driven eKYC, risk assessment and AML compliance services in a unified workflow to verify the identities of new and existing users, fight fraud and help meet compliance mandates throughout the customer journey.
Google is expected to launch a major update to Quick Settings in Android 16, and one of those new updates may include the ability to resize individual tiles. As in the current implementation, users are able to move tiles to fit the order they desire, but this update would also allow some tiles to be bigger than others. 1×1 and 1×2 sizes seem to be the only options, with the larger size adding a label alongside an icon. This update would allow users to set up a denser layout for Quick Settings, or retain current setup with its larger tiles, or find any mix of those two. Google’s exact plans for this new setup remain murky, but we’re getting a clearer picture of what a split notification and Quick Settings setup in Android looks like.
One of the controversial aspects of Samsung’s upcoming One UI 7 update, based on Android 15, is the decision to split notifications and Quick Settings into two different panes. But, as it turns out, that won’t be forced. In its new behavior, you can swipe down from the left side for notifications while swiping down from the right side for the Quick Settings menu. There are also new editing options, with individual blocks of the revamped Quick Settings menu able to be moved around. Much more important in that editor interface, though, is a “Settings” button in the top corner. From there, users are able to decide how the Quick Settings panel in One UI 7 works. Samsung seems to default to “Separate,” but a “Together” option seems to merge the two, leaving the functionality as it exists today where one swipe shows some Quick Settings tiles and your notifications, and a second swipe then reveals the full Quick Settings panel.
Google Messages has been working on dual SIM RCS since the start of this year, and support is now more widely rolling out in beta. To date, you could only have RCS active on one phone number at a time in Google Messages. Dual RCS support lets both be active simultaneously. Going to Settings > RCS chats notes the status of both as “Connected” with checkboxes to disable/enable each SIM. After testing in recent months, a lot more users are seeing this with the latest Google Messages beta (version 20241018_01_RC00+) in recent days, with another rollout today. There are more than a few reports in the US on Google Fi, T-Mobile, Verizon, etc. across various phone models, including Samsung. You can have both a physical SIM and an eSIM instead of needing variant hardware. However, not all Google Messages users report having dual SIM RCS support yet.
Marks & Spencer is rolling out a new option to streamline the purchasing process in its clothing stores, by adding add self-checkouts to fitting rooms in its 180 clothing stores. The move is designed to avoid to prevent shoppers from having to get in line twice during the shopping experience. The self-checkouts are expected to be rolled out to about 100 stores by early 2028. Initial plan is to add one of the checkouts per changing room area, with the number to increase depending on customer demand. “We’d like customers to be able to walk straight into the fitting room with no queue, try on what they’ve chosen, then pay there and just walk out,” Sacha Berendji, M&S’s operations director, told. To protect against shoplifting, the stores would have staff “hosting” the changing room areas to make sure customers did not leave without paying.
Marks & Spencer is rolling out a new option to streamline the purchasing process in its clothing stores, by adding add self-checkouts to fitting rooms in its 180 clothing stores. The move is designed to avoid to prevent shoppers from having to get in line twice during the shopping experience. The self-checkouts are expected to be rolled out to about 100 stores by early 2028. Initial plan is to add one of the checkouts per changing room area, with the number to increase depending on customer demand. “We’d like customers to be able to walk straight into the fitting room with no queue, try on what they’ve chosen, then pay there and just walk out,” Sacha Berendji, M&S’s operations director, told. To protect against shoplifting, the stores would have staff “hosting” the changing room areas to make sure customers did not leave without paying.
According to a new survey from e-commerce SaaS company AfterShip and polling partner Ipsos, 76% of respondents browse social commerce platforms for products, but buy through a retailer’s website, primarily due to trust and security concerns (41%). 52% don’t trust social media platforms but would use social commerce if the platforms felt more trustworthy (44%) and secure (28%). 79% of Americans use social media weekly, with 63% finding social commerce platforms “excellent” for inspiration and ideas. 52% have made a purchase on social media in the past. Despite some trust issues overall, younger shoppers are embracing social commerce, with 34% of 18- to 34-year-olds shopping on social media weekly. 65% of the age cohort are likely to continue shopping on social media platforms in the future, while only 15% said they are not at all likely to. This likelihood is 49% for 35- to 54-year-olds, and only 14% for those 55 and older. When it comes to online shopping overall, the survey found that consumers’ purchase decisions are primarily motivated by sales and promotions (87%), followed by reviews from other customers (78%). 38% say that a brand’s social content motivates them, while 34% cite live shopping videos as a factor and 31% of consumers say they are impacted by influencers.
Marqeta has unveiled a new solution that delivers BNPL payment options inside payment apps and digital wallets. The new Marqeta Flex is being developed with payment providers Klarna and Affirm and payments platform Branch. Marqeta Flex is designed to expand the distribution of BNPL and provide these options to consumers when they need them. For consumers, the solution will present personalized BNPL options inside the payment apps they use. It will also allow payment providers that offer BNPL options to access more consumers, and card issuers and digital wallets to access a variety of global BNPL providers through a single integration with Marqeta Flex. Branch plans to integrate Marqeta Flex into its payments app for W-2 and 1099 workers.
Fintel Connect, the provider of growth marketing solutions for the financial industry, announces the official launch of Fintel Check for BaaS, an automated, cost-accessible fintech partner monitoring tool designed for BaaS sponsor banks and middleware providers. With Fintel Check for BaaS, BaaS banks will have access to an automated, scalable solution to achieve comprehensive monitoring of fintech partners’ marketing activities. Fintel Check directly addresses the challenges BaaS banks face when monitoring and policing fintech partner marketing activity, including ensuring compliance, maintaining audit trails, scaling monitoring efforts, and managing costs effectively. With a robust rules engine, the platform automates the scanning, monitoring, evaluation, and documentation of fintech partner marketing activity, saving compliance teams hours of time-intensive manual monitoring that often leaves room for error.
Adyen, the financial technology platform of choice for leading businesses, has today launched a new report with Boston Consulting Group (BCG) which reveals a game-changing $185 billion market opportunity for SaaS platforms, a substantial 25% increase since 2022. Market growth is driven by a favorable interest rate environment — due to higher bank revenues from accounts — and an overall expansion of banking revenue pools. In parallel, the demand for embedded finance from small-and-medium-sized businesses (SMBs) — a sector typically underserved by traditional financial institutions — has broadened well beyond payments.1 This, coupled with the market’s expansion, opens a significant opportunity for SaaS platforms to embed into their offerings more advanced financial products like business accounts, loans, and issued cards. Given typical margins on these products, most platforms have an opportunity to multiply their current revenues by up to three or four times. The market is primed with strong demand from SMBs with 50% of SMBs expressing a high likelihood of utilizing a full suite of embedded finance products in the near future, reinforcing its growth potential. Top platforms are now generating more than 50% of their revenues from embedded payments and finance, and shines a light on embedded finance as the next critical growth lever for SaaS. To differentiate in an increasingly competitive market, platforms must consider integrating financial products in their offering.
Google and Ford have collaborated on big updates for Android Auto on the automaker’s electric vehicles for routing and finding charging stations. Specifically, these updates are designed with the Mustang Mach-E and Ford F-150 Lightning in mind. The first and biggest update is in regards to showing the estimated battery level of the vehicle within Google Maps on Android Auto for when you arrive at a destination. So if you pick a location on Maps, you’ll be able to see how much battery will be remaining when you arrive at that location. Google Maps already does this on vehicles with Android Automotive, but further integration is required to make this work through Android Auto. This works on all stops. Beyond that, Google Maps on Android Auto can also now show charging stations equipped with NACS chargers, primarily including Tesla Supercharger stations in Ford EVs equipped with the port. On the car’s side, these updates also allow the vehicle to start preconditioning the battery for charging while en route to the charging station. Other upgrades you’ll find include showing a list of “nearby amenities” at charging stops and a “minimum recommended charge” for a destination.
Apple Intelligence launched with iOS 18.1, and Apple has released two new ads to showcase some features with a touch of humor. Apple’s new iPhone 16 ads highlight its Apple Intelligence capabilities, including memory movie creation and writing tools. The ads showcase practical applications of these features and continue Apple’s tradition of using humor to make the generative technology relatable. With this new round of videos, the tone remains consistent, focusing on how AI can help reduce common frustrations or simplify tasks while evoking a smile. By using humor, Apple avoids overcomplicating the concept of AI for the average user. Instead of diving into technical intricacies, the ads illustrate relatable scenarios where the technology’s benefits are immediately apparent. This approach educates and entertains, positioning Apple Intelligence as a helpful tool rather than a futuristic abstraction.
More Innovations & Trends
Innovations & Trends
We’re on the look out for the latest trends, strategies and innovations. Do you have something to share? We’d love to hear from you.
Share an Innovation with Us.
Digital Brief delivers the latest disruptive financial innovations.