Customer experience is undergoing a seismic shift, and Gladly is leading the charge with The Gladly Brief.

It’s a monthly breakdown of market insights, brand data, and investor-level analysis on how AI and CX are converging.

Learn why short-term cost plays are eroding lifetime value, and how Gladly’s approach is creating compounding returns for brands and investors alike.

Join the readership of founders, analysts, and operators tracking the next phase of CX innovation.

Subscribe here to stay informed and ahead of the curve.

Subscribe now

Hey there! Welcome to the weekly dose of A.I stories you likely missed. You know how your uncle at Christmas tries to explain "the blockchain" and you just nod politely? This won't be like that. We're breaking down the biggest AI stories from this week in plain English – no tech jargon, no boring stuff. Just the good bits. Let's jump in.

OpenAI dropped a major update to ChatGPT on December 16th, and this one's pretty cool. The new image generation tool creates pictures four times faster than before, which means you're not sitting there watching a loading bar like it's 2005. But speed isn't even the best part. The new model (called "GPT Image 1.5") can now edit photos with scary-good accuracy. Want to change just the color of a car in a picture but keep everything else exactly the same? Done. Need to swap out a background while preserving all the lighting and shadows? No problem. OpenAI is basically calling ChatGPT a "creative studio in your pocket" now, and honestly, they're not exaggerating. The tool rolled out to everyone immediately, with business users getting access shortly after. For designers, marketers, or anyone who needs quick visual content, this is huge. It's also another sign that AI companies are racing to become one-stop shops for both text and images, turning what used to be separate tools into unified creative platforms.

Here's one that might make you go "hmm." Starting December 16th, Meta began using your conversations with its AI assistant to personalize your Facebook and Instagram feeds – including the ads you see. So if you chat with Meta AI about, say, hiking trails, don't be surprised when hiking boot ads start popping up in your feed. Meta says this feature now reaches over 1 billion monthly users across its apps, which is a staggering number. The company promises it won't use sensitive topics like religion, health, or sexual orientation for ad targeting, but here's the kicker: there's no way to opt out of this personalization. You either use Meta AI and accept that your chats will influence your feed, or you don't use it at all. This rollout is happening globally except in places with stricter privacy rules like the European Union, UK, and South Korea. Meta's CEO Mark Zuckerberg has been pushing hard to make "Meta AI the leading personal AI," and this move shows they're willing to turn those casual AI conversations into serious money. It's one of the first times a major platform has openly monetized AI chat data this way, and you can bet other companies are watching closely to see how users react.

Chip giant Nvidia surprised everyone on December 15th by launching its own family of cutting-edge AI models called Nemotron 3 – and they're completely open-source. These aren't small models either. They come in three sizes: Nano (about 30 billion parameters), Super (around 100 billion), and Ultra (a whopping 500 billion). To put that in perspective, bigger usually means smarter. But here's what makes Nemotron 3 special: these models can handle absolutely massive amounts of text at once – up to 1 million "tokens," which is like reading several books in one sitting. The smallest version processes information four times faster than its predecessor, which means near real-time responses even for complex questions. Companies like Accenture, Deloitte, Oracle, and ServiceNow are already testing these models for everything from coding assistants to multi-agent systems. Nvidia also claims these models cut costs by 60% for long reasoning tasks and showed a 15% accuracy boost in tests. Why is a chip company making AI models? Because Nvidia wants to show off what its hardware can do while giving businesses an alternative to closed systems like OpenAI. By making these models open and customizable, Nvidia is betting that companies will want to run powerful AI on their own infrastructure rather than renting it from someone else.

Amazon CEO Andy Jassy announced on December 17th that the company is creating an entirely new division dedicated to AI – and putting one of its most trusted executives in charge. Peter DeSantis, a 27-year Amazon veteran who helped build AWS from the ground up, will now lead a team focused on frontier AI models (like Amazon's "Nova" family), custom AI chips (the Trainium and Inferentia processors), and even quantum computing. This is Amazon basically saying "AI is now our top priority." The move comes after Amazon committed $50 billion to expand AI infrastructure for U.S. government agencies and invested billions in AI companies like Anthropic and potentially OpenAI. By putting all its AI eggs in one basket under DeSantis, Amazon wants to speed up innovation and better compete with Microsoft and Google. For AWS customers, this means more powerful AI tools that work seamlessly with Amazon's hardware. For Alexa users, it could mean a much smarter voice assistant. And for Amazon as a company, it's a clear signal that the next decade will be defined by how well they can combine huge AI models with huge computing power.

Google jumped into the news on December 17th with Gemini 3 Flash, a new AI model designed specifically for speed. While Google already has the powerful Gemini 3 Pro for heavy-duty thinking, Flash is all about giving you answers right now. Google describes it as "frontier intelligence built for speed at a fraction of the cost," which is marketing speak for "this thing is fast and cheap to run." The model improved accuracy by 15% compared to the previous generation while drastically cutting response times. Companies like Salesforce, Workday, and Figma are already using it to power real-time document processing, video analysis, and AI-driven workflows. The key here is that Gemini 3 Flash can handle text, images, and video all at once, making it perfect for businesses that need instant responses – think customer support chatbots or live data analysis. Google is clearly positioning this as their answer to enterprises who find most AI models either too slow or too expensive for everyday use. By offering a range of models (Pro for deep thinking, Flash for quick tasks), Google wants to make sure there's a Gemini model for every job, every budget, and every speed requirement.

Adobe is now facing a major class-action lawsuit filed on December 17th by author Elizabeth Lyon, who claims the company trained its AI models on copyrighted books without permission. Specifically, the lawsuit alleges that Adobe used pirated copies of books – including Lyon's own instructional works – to develop its "SlimLM" language models, which power document features in Adobe software. This is the first big copyright lawsuit to hit Adobe, though OpenAI, Google, and Anthropic have all faced similar challenges. In fact, Anthropic agreed to a $1.5 billion settlement earlier this year, the largest copyright settlement ever. The Adobe case could affect thousands of authors whose work may have been scraped without consent, and it raises serious questions about the legal use of copyrighted material in AI training. For Adobe, which has integrated AI across its creative tools (like the Firefly image generator), this strikes at a sensitive point: the company sells software to creators, and now creators are suing them for allegedly stealing their work. If the lawsuit succeeds, it could force Adobe to change how it trains AI models and potentially pay significant damages. More broadly, this case will help determine the rules for AI development – specifically, whether companies can freely use copyrighted content or need to license it first.

On December 18th, Microsoft and IT services giant Cognizant announced a major partnership to help big companies adopt AI faster. The deal focuses on embedding Microsoft's AI tools – including the powerful agentic AI features and Microsoft 365 Copilot – into real-world business workflows across industries like finance, healthcare, retail, and manufacturing. Cognizant will use its industry-specific platforms (like its healthcare system TriZetto) to integrate Azure AI services, creating pre-built solutions that companies can adopt without starting from scratch. The partnership also involves training thousands of Cognizant consultants on Microsoft's AI tech, creating an army of experts ready to help businesses deploy AI at scale. Cognizant's CEO said "AI underpins every transformation program we drive," while Microsoft's Chief Commercial Officer emphasized that combining Microsoft's AI with Cognizant's industry knowledge will "unlock transformative value." This partnership represents the next phase of enterprise AI – not just selling tools, but delivering complete solutions tailored to specific industries. For Microsoft, it expands Azure's reach into huge corporate projects. For Cognizant, it positions them as the go-to partner for AI transformation. And for businesses, it means faster access to AI solutions that actually work for their specific needs rather than generic tools they have to figure out themselves.

OpenAI announced on December 18th that it's teaming up with the U.S. Department of Energy to supercharge scientific research using AI. The partnership, formalized through a memorandum of understanding, will give DOE's 17 national laboratories access to OpenAI's most advanced AI models (like the GPT-5 series) to accelerate research in energy, climate, materials science, and national security. The idea is to pair OpenAI's cutting-edge AI with the DOE's world-class scientific facilities and massive datasets. Early projects include using AI to analyze complex experimental data, optimize fusion energy simulations, and help scientists generate and test hypotheses faster than ever before. OpenAI calls this initiative "OpenAI for Science" and believes AI can help researchers "explore more ideas, test hypotheses faster, and move from insight to validated results more quickly." The announcement came during a White House event where officials discussed making 2026 the "Year of Science" powered by AI. For OpenAI, this is a chance to showcase positive social impact while refining its models on unique scientific problems. For the DOE, it means potentially faster breakthroughs in critical areas like renewable energy and drug discovery. If successful, this partnership could lead to real-world advances – imagine AI helping discover new battery materials or accurately forecasting climate extremes – while setting an example for similar collaborations worldwide.

Here's a genuinely exciting breakthrough from December 18th: researchers at Italy's Institute of Technology used AI to design a molecule that makes pancreatic cancer cells more vulnerable to chemotherapy. The team developed an aptamer called "Apt1" that targets a protein (RAD51) cancer cells use to repair their DNA. When they added Apt1 to lab tests, pancreatic tumor cells became much easier to kill with chemo drugs, even at lower doses than normally required. The AI algorithm helped scientists rapidly screen potential candidates and identify Apt1 as the best option – a process that would have taken years through traditional trial-and-error. Pancreatic cancer is particularly deadly, with only about a 10% five-year survival rate in Italy, where around 14,000 people are diagnosed yearly. What makes this even more promising is that Apt1 appears to selectively target cancer pathways while leaving healthy cells largely unharmed in tests. The findings were published in Nature Communications as the team works toward clinical trials. This development shows AI isn't just analyzing data – it's actually inventing new medical solutions. If Apt1 or similar AI-designed molecules make it to human treatments, it could transform how we approach diseases that desperately need better therapies.

In a surprising partnership announced December 22nd, American ride-hailing rivals Uber and Lyft each teamed up with Chinese tech giant Baidu to launch driverless taxi trials in London starting in 2026. Baidu's electric Apollo Go robotaxis (the RT6 model) will appear in the Uber and Lyft apps, operating alongside regular cars with human drivers. This marks the first time American and Chinese autonomous vehicle leaders are competing directly on European streets, happening right after Google's Waymo began its own supervised tests in London. The UK has become a hotspot for self-driving car testing thanks to its new Automated Vehicles Act from 2024, which clarifies that the operating company (not passengers) is legally responsible for autonomous vehicles. London's complex roads will provide a tough challenge, but Baidu already runs one of the world's largest robotaxi services in China, so they're confident. For Lyft, which recently acquired European ride app FreeNow and is pushing a $200 million international expansion, this partnership is a major bet on the future. Both companies plan to run hybrid networks at first, mixing autonomous and human-driven cars to handle demand. Success in London could build public trust in robotaxis and open doors for rollout in other major cities worldwide.

Enterprise software company ServiceNow made headlines on December 23rd by announcing its largest acquisition ever: a $7.75 billion all-cash purchase of cybersecurity startup Armis. The deal reflects growing fears about AI-enhanced cyber threats. Armis specializes in protecting the Internet-of-Things and connected devices – essentially monitoring and securing all the smart gadgets, sensors, and equipment that businesses use. ServiceNow plans to integrate Armis's real-time device scanning and threat detection into its workflow automation platform, creating an AI-powered system that can spot suspicious device behavior and automatically trigger responses. The acquisition comes as cyberattacks boosted by AI are becoming more sophisticated, with ServiceNow's leadership arguing the deal will "triple the market opportunity" for their security business. Armis was valued at $6.1 billion just a month earlier and was preparing for an IPO before ServiceNow swooped in. At $7.75 billion, it's one of the biggest cybersecurity deals in recent years and represents ServiceNow's fourth major acquisition in the space (including a $2.85 billion purchase of AI startup Moveworks). The company's CFO said with Armis onboard, "we won't need to do any more M&A in the security space" for now. This deal underscores how AI and cybersecurity have become inseparable – companies need unified solutions that can handle AI-scale threats and protect the exploding number of smart devices that hackers might exploit.

The U.S. National Oceanic and Atmospheric Administration (NOAA) announced on December 17th that it's launching a new generation of weather prediction models powered by artificial intelligence. The AI Global Forecast System and hybrid AI-ensemble models use machine learning to improve forecast speed, efficiency, and accuracy. By combining AI with traditional physics-based simulations, NOAA's models can process atmospheric data faster than ever, giving forecasters quicker guidance with improved accuracy for weather patterns and hurricane tracks. The really impressive part? These AI-driven models match or exceed the accuracy of previous models while using only a fraction of the computing resources. NOAA's administrator called it a "significant leap forward" that provides faster warnings to the public at lower cost. The AI models enable more frequent forecast updates and better predictions of tropical cyclone intensity and heavy rainfall. Early testing showed notable improvements, and the efficiency gains mean taxpayers get better forecasts without constantly building expensive supercomputers. This is AI working quietly in the background to potentially save lives and property by improving something as fundamental as daily weather forecasts. Beyond national benefit, NOAA's adoption sets an example for weather agencies worldwide, many of which share data and could collaborate on AI techniques. It's a reminder that AI isn't just about flashy consumer apps – it can transform critical public services in ways that directly protect communities.

And that's a wrap! Twelve stories that show AI is everywhere right now – from making weather forecasts more accurate to designing cancer-fighting molecules, from turning your social media chats into ad gold to helping scientists unlock new discoveries. Whether you're excited, nervous, or just trying to keep up, one thing's clear: AI isn't slowing down. Same time next week for another round? Until then, try not to tell Meta AI about any expensive hobbies. Trust me on this one.