Mistral released an open-sourced voice model that could rival paid voice AI, such as those from ElevenLabs and Hume AI, which the company said bridges the gap between proprietary speech recognition models and the more open, yet error-prone versions. The company said Voxtral “offers state-of-the-art accuracy and native semantic understanding in the open, at less than half the price of comparable APIs.” Voxtral, at a 32K token context, can listen to and transcribe up to 30 minutes of audio or 40 minutes of audio understanding. It offers summarization, meaning the model can answer questions based on the audio content and generate summaries without switching to a separate mode. Users can trigger functions and API calls based on spoken instructions. The model is based on Mistral’s Mistral Small 3.1 and supports multiple languages and can automatically detect languages. Mistral added enterprise features to Voxtral, including private deployment, so that organizations can integrate the model into their own ecosystems. These features also include domain-specific fine-tuning and advanced context and priority access to engineering resources for customers who need help integrating Voxtral into their workflows. Mistral stated that Voxtral outperformed existing voice models, including OpenAI’s Whisper, Gemini 2.5 Flash and Scribe from ElevenLabs. Voxtral presented fewer word errors compared to Whisper, which is currently considered the best automatic speech recognition model available.
Vidnoz AI unveils product avatar and interactive avatar, redefining marketing video- Product Avatar- features realistic hand positioning, shadowing, lighting, and advanced lip-syncing; . Interactive Avatar enhances real-time engagement for customer-facing industries with natural speech interaction, multilingual communication
Vidnoz AI launched two groundbreaking features set to fundamentally transform marketing video creation, enabling businesses to produce engaging, personalized, and cost-effective content at scale. With these tools, users can easily empower their brand with UGC, product, and advertising videos, eliminating upfront costs or complex technical hurdles. Product Avatar is a breakthrough solution: upload any product image, select an avatar template, and instantly generate a lifelike avatar holding the item. This virtual human features realistic hand positioning, shadowing, lighting, and advanced lip-syncing. Such realism enables the creation of UGC and advertising videos that deeply integrate products into everyday scenes, fostering stronger resonance and empathy with audience for optimal brand promotion. Beyond the diverse pre-set library, users can also upload their unique brand spokesperson to create a custom Product Avatar. Interactive Avatar enhances real-time engagement for customer-facing industries. This intelligent virtual assistant supports natural speech interaction, multilingual communication, dynamic facial expressions, and contextual responses, making it ideal for sectors like banking, insurance, healthcare, training, and public services. Embedded into websites, apps, or smart terminals, Interactive Avatars provide 24/7 automated front-desk support, assisting with inquiries and information delivery, reducing human workload while maintaining consistent service quality.
OpenAI is planning to integrate a payment checkout system into ChatGPT in a bid to boost revenues.
OpenAI is planning to integrate a payment checkout system into ChatGPT in a bid to boost revenues. OpenAI is reportedly planning to take a share of eCommerce sales made via ChatGPT. That’s according to a report Wednesday (July 16) by the Financial Times (FT), which says this move is part of a plan by the artificial intelligence (AI) startup to derive new revenues from online shopping features. The company already displays products on its platform with the option of clicking through to online retailers, and inked a partnership with eCommerce platform Shopify in April. Now, sources told the FT, OpenAI hopes to integrate a checkout system into ChatGPT, to make sure users complete transactions within the platform. Sellers who make sales this way would pay a commission to OpenAI. PYMNTS has contacted OpenAI for comment but has not yet gotten a reply. The FT notes that this marks a shift in strategy for OpenAI, whose revenue has chiefly come from subscriptions to its premium services. Getting a slice of eCommerce sales would let OpenAI earn money from people using its free service. The report adds this move is another threat to Google’s business model, with consumers increasingly turning to AI chatbots for online search and shopping. OpenAI is already building its own browser in a separate challenge to Google, per a report last week by Reuters. While the feature is still in development, the sources told the FT that OpenAI and partners like Shopify have been presenting early versions to brands and discussing financial terms.
VantageScore 4.0 is coming but much work needs to be done and analysts suggest it would arrive sometime around the summer of 2026
In the days following the announcement that VantageScore 4.0 was suddenly in play for conventional mortgages, FinLocker‘s Brian Vieaux had to burst some bubbles. Most of the 15 to 20 loan officers he spoke to were ecstatic about being able to use the new credit scoring model. Millions of new prospective borrowers could be scored with VantageScore, they said. That’s opportunity in a down market. “I was like, ‘OK, guys, totally agree. If you can wave a magic wand and tomorrow — in your point of sale and your LOS, and the underwriting engines and all the technologies that that talk downstream — this was there and ready, yeah, I think it would open up some interesting opportunities with borrowers. And it will,” Vieaux said. “The other side of it is talking to the owner-operators of these companies, and they’re like, ‘This doesn’t become reality till 2026.’” For about a week, the mortgage industry operated in confusion about what comes next. It wasn’t even clear whether the Federal Housing Finance Agency (FHFA) wanted both FICO and VantageScore scores, or if they wanted lenders to choose between the two. FHFA Director Bill Pulte on Tuesday afternoon clarified several key points about Fannie Mae‘s and Freddie Mac‘s embrace of VantageScore, which is collectively owned by the three major credit reporting bureaus — Experian, Equifax and TransUnion. The tri-merge credit report will also remain. Pulte said lenders will be able to choose between FICO Classic, introduced in the late 1980s, and VantageScore 4.0. The GSEs will be working to update their Selling Guide policies, but at the moment they are not currently able to purchase mortgages with VantageScore 4.0 credit scores. The agencies will also need to create a new loan level pricing adjustment (LLPA) matrix before it can purchase loans with VantageScore 4.0 credit scores.
New Generation’s tech transforms static product catalogs into structured, AI-readable data through AI-native storefronts hosted on custom subdomains with embedded payments to power agentic commerce across chat, voice, and emerging agent-driven channels
New Generation (New Gen), a technology company building infrastructure for the AI internet, now transforms static product catalogs into structured, AI-readable data through intelligent storefronts hosted on custom subdomains like ai.brand.com. These storefronts support humans, who see a personalized, conversational interface, as well as AI agents, which access the same product information programmatically. Both paths provide frictionless, secure transactions using embedded checkouts without redirects or scraping. The platform now allows AI agents to quickly and securely check out from merchant sites across chat, voice, and soon through emerging agent-driven channels. New Gen New Gen is leveraging trusted payments infrastructure from Visa and is among the first collaborators in the Visa Intelligent Commerce sandbox. Together with Visa, NewGen is exploring enabling: Intelligent, embedded payments – Facilitate secure, seamless transactions initiated by AI agents on behalf of human users across chat, voice, or embedded experiences. Trust and visibility in AI-driven transactions – Integrate fraud, identity, and authorization protocols upstream in the agentic shopping flow, not only at checkout to give consumers control and confidence in each transaction. Simple and scalable integration – Easily plug into and scale with commerce platforms, interfaces, or ecosystems with flexible APIs and pre-built modules, without overhauling existing infrastructure.
Infobip taps NumHub’s API-driven cloud-based SaaS tech to offer Branded Calling ID (BCID) solution that integrates Rich Call Data (RCD) and FCC-mandated auth by streamlining all aspects of BCID registration, onboarding, vetting, billing, and reporting
Infobip has expanded its partnership with NumHub, to offer Branded Calling ID™ (BCID) as part of their Voice product solution. Infobip has adopted NumHub’s API-driven cloud-based SaaS solution streamlining all aspects of BCID registration, onboarding, vetting, billing, and reporting. As a valued partner, Infobip will engage enterprise businesses, empowering them to own their brand identity through the platform’s simplified BCID adoption process. This development comes at a crucial time, aligning with the FCC’s Eighth Report and Order, which encourages the industry to proactively implement Rich Call Data display with branded calling solutions that are supported in the mandated STIR/SHAKEN framework. “Branded calling helps businesses to improve answer rates and ROI, while keeping customer satisfaction at the forefront of each interaction,” said Ivan Ostojić, Chief Business Officer at Infobip. Branded Calling ID (BCID) is an industry-adopted, CTIA-governed framework designed to restore trust in phone calls by enabling verified businesses to display their brand name, logo, call reason, and number directly on consumer devices. BCID integrates Rich Call Data (RCD) and Internet Engineering Task Force (IETF) FCC-mandated STIR/SHAKEN authentication, ensuring that only legitimate, authenticated and vetted businesses can display branded caller ID information. By leveraging cryptographic signatures and industry-standard governance, BCID protects businesses from spoofing, enhances call answer rates, and ensures consumers can confidently identify and engage with trusted brands. Unlike traditional branded calling models that operate under vendor-controlled ecosystems, BCID is an open, standardized solution built to serve the entire telecom industry including enterprises, CSPs, OSPs, UCaaS/CCaaS providers, and consumers.
Amplitude acquires Kraftful to enhance ‘Voice of Customer’ insights – personalizes questions in real time, adapting based on previous responses to uncover deeper motivations and unmet needs
Behavior-tracking software firm Amplitude Inc. has acquired Kraftful Inc., a “Voice of Customer” startup. Kraftful offers a product research and insights platform that uses artificial intelligence to help product teams deeply understand and act on user feedback at scale. The company offers a centralized system that ingests user input from dozens of channels, into a unified, searchable space, rather than relying on manual analysis of disparate data sources. The platform allows teams to spend less time sorting through feedback and more time building what users actually need. The company also offers built-in hallucination detection to improve trust in AI-generated outputs, offering product managers a clear and reliable understanding of how users are experiencing their products. Along with passive analysis, Kraftful also allows teams to gather insights with AI-generated surveys and live conversational interviews. The tools personalize questions in real time, adapting based on previous responses to uncover deeper motivations and unmet needs. Surveys can be launched based on topics drawn from existing feedback or on entirely new initiatives, making it easy to test concepts, prioritize features and validate ideas at scale. Amplitude plans to integrate Kraftful’s capabilities as a native part of its platform, giving teams a faster way to turn customer input into better products. With these new capabilities, Amplitude customers will be able to see the full picture of what customers are doing and saying.
Everyday essentials and groceries score with Prime Day shoppers- household essentials segment was tied for the top category with apparel and shoes, with 30% of Prime users buying those kinds of items; health and wellness and beauty and cosmetics were other top categories for Prime Day 2025
Numerator published early results from the first two days of this year’s extended Prime Day promotion, indicating that shoppers are interested in scoring savings on items for their home and pantry. Four of the top five items sold on July 8 and 9 were household or grocery products. The household essentials segment was tied for the top category with apparel and shoes, with 30% of Prime users buying those kinds of items. Health and wellness and beauty and cosmetics were other top categories for Prime Day 2025. Groceries also fared well, at least during the upfront part of the promotion. 15% of Prime customers ordered groceries during the first days of the event, more than toys and video games (12%), office/school supplies (11%), small appliances (10%) and books, videos and media (10%). Also reflecting shoppers’ value-centric mindsets is Numerator’s finding that the average Prime Day spend per order is $57.12. 35% of orders were placed for $20 or less. 57% of Prime Day shoppers compared Amazon’s prices to other retailers before buying something. 69% checked prices at Walmart, 46% compared prices to Target, 29% compared prices to club retailers like Sam’s Club or Costco and19% compared prices to traditional grocery stores.
Mirriad to combine its visual quality, access to premium CTV content, and its control tools for content-owners with Rembrand’s AI-powered virtual placement software to enable brands to enhance attention and resonance to break through clutter
Rembrand and Mirriad, a virtual placement company, have formed a strategic joint venture for virtual product placement in the U.S. market. Brands will now have unparalleled opportunities to engage audiences and capture cultural moments as a tool to break through clutter and gain the attention of their audience. The joint venture will leverage Mirriad’s established expertise in virtual placement including: its visual quality, its incredible access to premium CTV content, and its sophisticated control tools for content-owners. This will be combined with Rembrand’s pioneering AI-powered virtual placement software, which enables significantly faster turnaround times while maintaining premium placement quality. The result is a cutting-edge solution designed to enhance brand attention and resonance in an increasingly competitive media landscape. The combination enables brands to connect with audiences inside, around, and beyond content like never before, whether via long-form storytelling or high-performance creator campaigns. The new joint venture will manage all U.S. publisher and brand relationships for virtual product placement, ensuring uninterrupted service and expanded opportunities for existing and new clients. The existing supply and demand team for Mirriad US will continue in their roles within the new venture, enabling all U.S. supply relationships to continue uninterrupted. Existing contracts, work, and payments will continue to operate through Mirriad US, per previous agreements.
Image and video API platform ImageKit’s new AI generative fill intelligently extends any image to a new width or height by painting missing pixels contextually, with the added flexibility to control what gets filled
ImageKit, a unified image and video API platform with integrated digital asset management (DAM), has expanded its AI-powered image editing capabilities, equipping businesses to streamline repetitive design tasks, reduce production costs, and speed up product time-to-market. Powered by simple URL-based transformation parameters, these enhancements bring scalable automation, intuitive workflows for developers and designers, and real-time delivery of brand-consistent visuals. A standout feature in this release is AI generative fill, which intelligently extends any image to a new width or height by painting missing pixels contextually, with the added flexibility to control what gets filled. Whether transforming a vertical product shot into a panoramic hero banner or adapting visuals for multiple screen sizes, generative fill replaces traditional reshoots and redesigns with intelligent automation, cutting both time and costs. AI background change makes it possible to instantly replace an image’s background. ImageKit’s AI composes different scenes behind the subject with photo-realistic precision. ImageKit also introduces AI background removal with realistic shadow generation, using advanced AI models to isolate subjects and apply natural drop shadows.