Google LLC is a global technology company and subsidiary of Alphabet Inc., best known for its search engine, digital advertising business, and a broad ecosystem of consumer and enterprise products. The company develops internet services, cloud infrastructure, mobile and desktop operating systems, and hardware devices used by billions of users worldwide.
Aviva built a shared factory for machine learning so data scientists can build, test, approve, deploy, and monitor models in a repeatable way instead of doing manual setup each time.
This is like an ultra-detailed 3D CAD tool for molecules, powered by AI. Instead of engineers designing car parts, RosettaFold3 designs and predicts how proteins, DNA, and smallâmolecule drugs fit and move together inside the body.
Give each AI agent a badge, a camera, and a fence: control what it can access, watch what it does, and stop forbidden actions in real time.
Think of NVIDIA BioNeMo as a set of very smart chemistry and biology "co-pilots" that can read and write molecules and proteins the way ChatGPT reads and writes text. Instead of scientists manually trying out millions of possibilities in the lab, BioNeMo helps them design and screen promising drug candidates on a computer first, massively narrowing the search space.
All factory data is stored in one organized place so analysts and data scientists can train models, test them, and build reports from the same data.
This is like giving every shopper their own smart personal assistant that knows the entire store, all the promotions, and the shopperâs preferences, and can guide them from âI have a needâ to âorder placedâ through natural conversation across web, app, or even voice.
Small businesses are using AI like a digital helper to speed up everyday work and get more done with the same team.
The AI writes a first draft of business documents so teams start from something useful instead of a blank page.
This is like putting a smart security camera on all your insurance transactions. It watches events in real time, spots suspicious patterns that look like fraud, and alerts your team before money goes out the door.
A pretrained language model is further trained on conversation examples so it responds more naturally in chat-style interactions.
AI assistants look up the right internal documents and data before answering, so responses are more accurate.
Instead of checking who changed what in each ML account one by one, a bank can collect all activity logs in one place to spot risky actions faster.
Think of this as building your own âNetflix-styleâ recommendation brain: it watches what each user does, learns their tastes, and then uses a mix of traditional recommendation models and modern generative AI to decide what to show or suggest next.
Imagine a huge classroom where different versions of Googleâs Gemini sit sideâbyâside answering the same homework and exam questions. A panel of judges then scores which Gemini answers are most helpful for students. This paper is about building that classroom arena and seeing how good Gemini really is as a learning assistant.
This is like giving an insurer a living, zoomable map of how cars and drivers behave in the real world, updated in near real time, and then using AI to spot risks, opportunities, and patterns that humans would never see by looking at tables and static reports.
This is like a superâsmart search and monitoring engine for banks and financial firms that can instantly scan all their data (transactions, logs, customer activity, documents) to spot risks, fraud, and opportunities, then plug into AI tools for answers and automation.
This is like having a smart assistant watch all your videos and automatically create a searchable index of whatâs said, who appears, where logos show up, and key momentsâso teams can quickly find and reuse the right clips without manually scrubbing through footage.
This is like upgrading an insurerâs old spreadsheet-based risk calculator to a smart assistant that not only predicts which policies are risky more accurately, but also clearly explains which customer or policy features drove each prediction.
This would be like a smart insurance analyst that reads articles and policy documents about social engineering fraud (phishing, fake invoices, business email compromise) and explainsâin plain Englishâwhat is and is not covered, where the gaps are, and what questions a broker or client should ask.
Imagine a hotel that remembers every guest like a great concierge: what room temperature they like, which pillow they prefer, when they usually arrive, and what they tend to order. AI in hospitality is the digital brain behind that experienceâquietly watching patterns in bookings, reviews, and operations so staff can serve guests faster, more personally, and with fewer mistakes.
Think of this as a citywide âcontrol towerâ that watches whatâs happeningâtraffic, utilities, emergency calls, citizen requestsâand then uses AI to suggest faster, cheaper, safer ways to run city services.
This is like having a smart, offline paralegal that can read through all your case files, contracts, and statutes stored on your own servers and then answer questions by mixing two skills: fast keyword search and âmeaning-basedâ AI search. It never has to send your documents to the cloud.
This is like giving your travel website a smart, 24/7 travel agent that chats with visitors, helps them find trips, and completes bookings automatically.
This is like giving every hotel guest their own smart local concierge who knows the city, the guestâs preferences, and the hotelâs offerings, and then auto-builds a detailed, bookable trip plan for their stay.
This is about using tools like ChatGPT as a very fast junior market researcher: you ask it questions about consumers, brands, or markets, and it drafts insights, survey ideas, and segment descriptions instead of a human doing everything from scratch.
Think of this as turning tools like ChatGPT into a smart study and research partner for a university: it helps students learn faster, teachers design better lessons, and researchers explore ideas more quickly, all while the university figures out how to use it safely and effectively.
This is like giving Netflix a smart brain that quietly watches what you watch, when you stop, what you search for, and then rearranges the entire app, recommendations, images, and streaming quality just for youâmillions of people at once, all differently.
This is like giving every teacher a super-fast, tireless teaching assistant that can read student work, score it, and draft feedback so the teacher can focus on teaching instead of paperwork.
This is like having an always-available teaching assistant that reads studentsâ short answers and reports, compares them to a grading guide, and suggests scores and feedback so instructors donât have to grade everything by hand.
Think of this as putting an AI âair traffic controllerâ on top of your customer support systems in the cloud. It quietly watches everythingâtraffic spikes, slow services, error logsâand automatically tunes the environment so support agents and customers get fast, reliable help 24/7.
Imagine a shopper can take a photo of a dress they see on the street, upload it to your online store, and instantly see similar dresses you sellâno need to guess keywords like âfloral midi dress with puff sleeves.â Thatâs visual search for ecommerce.
This is Netflixâs R&D lab for making sure every member quickly finds something theyâll love to watch. Think of it as a constantly learning concierge that rearranges the entire Netflix store for each viewer, in real time.
This is like having an AI-powered design buddy in a sneaker store: you tell it the vibe, colors, and style you want, and it helps you co-create a unique pair of shoes tailored to your taste.
This is like giving Mango its own smart âshop assistant in the cloudâ that can chat with customers and employees, answer questions, and help with tasks across web, app, and possibly in-store channels.
This is like giving LâOrĂ©alâs marketing team a tireless digital copywriter and designer that runs on Google Cloud. Marketers describe the campaign or product, and the AI helps generate onâbrand text, images, and variations for ads, social posts, and product pages in seconds instead of days.
This is like giving your marketing team a crystal ball that looks at all the clicks, calls, and purchases your customers made in the past and then guesses what theyâre likely to do next, so you can talk to the right people with the right offer at the right time.
Think of this as a smart store clerk who quietly watches what each shopper likes, remembers their habits, and then rearranges the shelves and offers just for that person in real timeâacross websites, apps, emails, and ads.
This is like giving every customer a tireless digital helper that can answer questions, solve common problems, and route issues to the right humanâ24/7âthrough chat on your website, app, or messaging channels.
Think of this as a map of all the ways online stores are using AI todayâlike a guidebook that explains how Amazonâstyle recommendations, smart pricing, chatbots, and fraud checks actually work and where theyâre going next.
This is like a standardized test for legal AI tools. Instead of trusting marketing claims, it builds exam-style questions and grading rubrics so you can see which AI systems actually understand law and which ones just sound confident.
Think of Azure AI Video Indexer as an AI librarian for all your videos. It automatically watches every video, recognizes people, objects, brands, spoken words, and emotions, and then turns that into searchable labels and timelines so your teams can instantly find the exact moments they need instead of scrubbing through hours of footage.
This is like giving every call center agent a super-smart sidekick that listens to customer interactions in real time, figures out what the customer is feeling and wants, and then quietly tells the agent the best next thing to say or do.
Think of this as a super-smart ad trader that watches billions of peopleâs clicks in real time and automatically decides which ad to show, to whom, at what price, and on which platform to get the best returnâfar faster and more accurately than any human team could.
Imagine every person watching TV or scrolling online sees an ad thatâs been instantly rewritten and re-edited just for themâdifferent script, images, and product angleâcreated automatically by AI instead of a big creative team doing one version for everyone.
This is like a game-making and story-writing assistant in one: you write or describe a story, and the AI helps turn it into an interactive, playable experience with scenes, characters, and branching choices.
Think of your marketing like a relay race where several runners (ads, emails, social posts, etc.) help score a sale. Dataâdriven attribution models use statistics and AI to figure out which runners actually mattered most, instead of just giving all the credit to whoever crossed the finish line last.
Think of this as a specialist AI toolkit for retailers and consumer packaged goods brands that helps them better understand shoppers, predict demand, and personalize experiences across stores and ecommerceâlike having a data-driven co-pilot for merchandising, marketing, and operations.
This is like giving your existing code to a very smart assistant and asking it to write the unit tests for you. The large language model reads the code, guesses what it should do, and then writes test cases to check that behavior automatically.
This is like giving your claims department a team of tireless digital assistants that can read documents, understand photos, and follow rules to move claims from âreportedâ to âpaidâ with minimal human involvement.
This is like having a smart digital tutor that learns how each student studies best, then automatically adjusts lessons, examples, and practice questions to fit that studentâwhile helping teachers design and manage this at scale.
Think of this as giving pharma companies a super-smart digital lab assistant and paperwork robot rolled into one. The assistant can sift through mountains of scientific data to suggest promising new drugs faster, and it can also take over a lot of the routine documentation and admin work that bogs down scientists and healthâcare workers.
Imagine every student and every teacher having a patient, always-available tutor in their laptop that knows the Khan Academy curriculum and can explain things step by step, ask questions back, and guide practice instead of just giving answers. Thatâs what Khanmigo is: an AI helper built into Khan Academy for learning and teaching.
This is like having an AI pairâprogrammer built into Visual Studio Code. As you type code or comments, it suggests whole lines or functions, helps you write boilerplate faster, and answers coding questions inside your editor.
Think of this as Netflix building its own very smart "taste brain" that understands movies, shows, images, and text, then wiring that brain into all the ways it personalizes what you see â rows, artwork, search, and more â instead of relying on a bunch of separate smaller brains.
This is like giving every college student a 24/7 smart study coach that can explain concepts in simple terms, quiz them, and help them plan their learning, rather than just giving them another digital textbook.
This is like giving your insurance claims department a tireless digital assistant that can read claim documents, check details, and help decide payouts much faster and more consistently than humans alone.
Imagine having a super-organized digital stylistâs assistant that can look at millions of fashion photos, read their descriptions, and automatically tag and sort everything by style, color, cut, and trend so your teams and algorithms can instantly find the right looks.
Think of Sarai as a smart, always-on hotel salesperson and receptionist that can talk with guests on your website or messaging channels, answer questions about your property, and complete reservations on its own â like your best front-desk agent working 24/7, but digital.
This is like having a super-fast digital media trader that watches your Facebook ads 24/7 and automatically shifts budget, bids, and creatives to whatever is working bestâwithout a human needing to click buttons all day.
Think of Copilot Arena as a public test track where many different AI coding copilots race on real developer tasks. Instead of trusting vendorsâ own benchmarks, this platform lets you see how each coding AI actually performs with real users and messy, real-world code problems.
Think of this as giving every journalist a smart digital assistant that can help research, draft, factâcheck, and personalize stories at scaleâwhile editors stay in control of what gets published.
This is about using smart software and robots as a âdigital brainâ for minesâhelping decide where to dig, how to run equipment, and how to keep workers safe, based on huge amounts of data from sensors, machines, and geological surveys.
This is like a super-accurate 3D blueprint generator for molecules inside the body. Instead of running long, expensive lab experiments to see how proteins and potential drugs fit together, AlphaFold 3 uses AI to predict those shapes on a computer in hours, so scientists can shortlist the best drug ideas much faster.
Imagine your entire IT environmentâservers, networks, apps, cloud servicesâconstantly watched by a smart assistant that never sleeps. It reads all the logs, alerts, tickets, and performance data, spots early warning signs, figures out whatâs really important, suggests fixes, and in many cases can trigger automated responses before users even notice a problem.
Imagine a blood pressure clinic that treats each patient the way a tailor makes a custom suit: it uses your genes, lifestyle, gut bacteria, and medical historyâanalyzed by AIâto pick the drug and dose that fit you best instead of guessing and adjusting over months.
Like having a research assistant that constantly reads and understands the FDAâs drug shortage pages and instantly answers your teamsâ questions about whatâs in shortage, why, and what the latest guidance is.
Think of this as a smart research analyst that constantly reads and updates all available reports, news, and data about military and commercial drones, then answers your questions in plain Englishâlike a âChatGPTâ specialized in the global drone market.
This is like a highâtech weather forecast, but for traffic jams. It looks at how traffic has behaved across a city over time and space (roads, intersections, hours of day) and then predicts where and when congestion will build up, so planners and operators can act before it happens.
Think of this as a smart listener that reads what your customers write (emails, chats, reviews, tickets) and instantly tells you if theyâre happy, confused, or angryâat huge scale and in many languagesâwithout needing a room full of people to read everything.
This is like giving your call center and support team a super-smart digital receptionist that can talk to customers, answer questions, and route issues 24/7 without getting tired.
Imagine a very smart digital artist and writer that has watched and read almost everything on the internet. When you ask it for a song, a video idea, a game character, or a script, it can instantly draft something new that looks like a human made it. Thatâs generative AI: a content factory that turns instructions into creative outputs (text, images, music, video, code).
This is like putting a smart âcheck engineâ light on every aircraft part and piece of ground equipment. Instead of waiting for something to break, Azureâs AI watches sensor data and tells you in advance when a component is likely to fail so you can fix it during planned downtime.
This is like giving a retail business a smart digital operations manager that can analyze sales and customer data, answer questions, and suggest actions to run stores and ecommerce more efficiently.
This is like giving every customer their own smart, always-on concierge that remembers who they are, what they like, and can talk to them naturally over chat, email, or other channelsâwithout needing a human to type every response.
Think of this as turning your marketing department into a super-targeted, always-on trading desk that continuously tests, learns, and optimizes where every dollar goesâusing AI as the brain that watches all the data and adjusts in real time.
This is like letting shoppers show your store a picture of what they want instead of typing words. The AI then finds the closest matching products across your catalog in seconds.
Imagine your streaming app as a smart host at a party who learns what each guest likes, suggests the right music and games at the right moment, and nudges people before they leave so they stay longer and have more fun. This system uses AI to do that automatically for every user in your mobile entertainment app.
This is like having a 24/7 digital concierge who looks and talks like a real person, remembers guest preferences, and can handle routine questions and requests for a hotel or luxury travel brand without needing more staff at the front desk.
This is like giving every marketer a smart digital assistant that can brainstorm campaigns, write and adapt content for lots of channels, and analyze whatâs workingâso a small team can operate like a much larger one.
Think of this as giving your marketing team a super-fast, super-smart analyst who studies every customer click, email, and ad impression, then quietly tells you: âshow this group offer A, show that group message B, and stop wasting money on these channels.â
Think of this as turning your marketing team into pilots of a self-driving ad machine: humans set goals and guardrails, while AI continuously tests, tweaks, and reallocates budget across channels to get you more customers for less money.
Imagine your whole supply chainâfactories, warehouses, trucks, and suppliersârunning like a smart GPS for your business. It constantly checks traffic (demand), fuel (inventory), and roadblocks (disruptions) and then suggests the best route and timing so you deliver on time with less waste and lower cost.
Think of this as a global field guide to âAI-as-a-junior-lawyerâ: it surveys how tools like ChatGPT-style assistants, contract analyzers, and legal research bots are being used in law firms and inâhouse teams around the world, and what that means for cost, risk, and competitiveness.
Imagine if every customer saw a version of your brand that felt like it was made just for themâa website, email, or ad that talks in their language, remembers their preferences, and adapts in real time as they interact. AI personalization is like giving every customer their own dedicated concierge who knows them well and continuously learns how to serve them better.
Think of Orbitae AI as a smart control tower for an automotive companyâs data. It connects to all your scattered data sources (production, sales, afterâsales, supply chain), lets managers ask questions in natural language, and then turns complex analytics into simple dashboards, forecasts, and recommendations to run the business better and faster.
This is like having a super-smart media planner that reads every page, video, or app screen in real time and decides whether your ad should appear there based on how likely someone is to act (click, visit, buy) â all without using cookies or following people around the web.
This is about using AI as a smart digital marketing assistant that creates, tests, and optimizes your online ads automatically so you sell more without manually tweaking every campaign.
Think of these AdTech AI agents as a team of tireless digital interns that understand ads, audiences, and campaign data. You tell them your goals (e.g., âget more app installs in Germany within this budgetâ), and they continuously research options, tweak settings, buy media, test creatives, and report backâwithout needing a human to click every button in every platform.
Think of AI in programmatic advertising as a super-fast trading bot for ad space: it constantly scans who is online, what theyâre doing, and in a split second decides which ad to show, at what price, and on which website to maximize your marketing results automatically.
This is like having an onâdemand creative team that instantly drafts lots of ad visuals and copy options for you, so your marketers just pick and refine instead of starting from a blank page.
Think of this as a tireless creative and analytics assistant that can draft campaigns, personalize messages for each customer, and learn from results to do better next timeâall in minutes instead of weeks.
This is like having an alwaysâon digital writersâ room that helps you brainstorm concepts, outline plots, write scenes, and refine scripts for films, TV, ads, or online videos in minutes instead of weeks.
This is like a very powerful âGoogle Maps brainâ that can look at extremely detailed satellite and aerial images, understand whatâs on the ground (roads, buildings, ships, fields, etc.), and connect that with other types of data, so many different applications can reuse the same core model instead of building their own from scratch.
Think of this as a tireless digital marketing assistant that can design ads, test many versions automatically, and keep tweaking them to get more clicks and conversionsâwithout a human having to watch it every minute.
This is like Netflix-style recommendations, but for news and media, where editors set the rules of the game and algorithms handle the heavy lifting of matching each reader with the most relevant stories and content.
This is like having a smart digital marketing assistant inside Facebook and Instagram that automatically builds and optimizes your ads so more of the right people see them, for less money, with less manual tweaking.
Think of this as a smart thermometer for customer feelings. It reads reviews, tweets, and comments at scale and tells you whether people are happy, angry, or worried about your products and brand.
This is like having a tireless junior creative team that studies which ads perform best, then automatically drafts new versions of those ads that are more likely to workâheadlines, copy, and visualsâover and over again.
Think of this as a smart ad-placing assistant that studies who actually clicks and buys from your ads on social platforms, then automatically shows future ads to more people who look and behave like those best customers.
Think of media buying as trading ads on a stock exchange. Programmatic buying is the robot trader that automatically bids on ad space in milliseconds. AI makes that robot trader much smarter, faster, and able to decide which impressions are worth paying for, at what price, and for which audience.