Brian Impact Foundation uses Anthropic's Claude in their Benevolence Enhancing Expert System (BEES) to review 100x more candidates for their social innovation fellowships. The AI processes large volumes of unstructured data, summarizing and organizing potential candidate information, allowing the small team to expand their reach, enhance equity in their application process, and allocate resources more efficiently.

FinQuery, a fintech platform that helps organizations manage financial accounting and spending, uses Gemini for Google Workspace to enhance productivity and collaboration across its remote workforce. Employees use generative AI features to brainstorm ideas, draft emails and documents faster, assist in coding tasks, and solve complex business problems, allowing teams to work more efficiently and focus on higher-value activities.

Clay has used both OpenAI and Anthropic’s models to power its AI sales tool, which automates outreach by identifying leads, enriching data, and generating personalized email messaging. This saves lots of hours of manual data work and improves the quality of outreach, increasing customer engagement and revenue while reducing costs.

BlueOcean, a brand intelligence platform, used Anthropic's Claude 3 models (Haiku and Opus) via Amazon Bedrock to consolidate four separate API calls into a single, more efficient one, resulting in over 50% cost reduction.

MavenAGI uses GPT-4 to power automated customer support agents. It integrates with CRMs like Salesforce and Zendesk to personalize responses, evaluate answers, and resolve 93% of customer queries autonomously. This results in a 60% reduction in resolution time and reduces the cost per ticket from $40 to $8, enhancing both customer and employee experience.

Canva uses OpenAI's GPT-4 to power its Magic Studio, integrating multiple AI tools like Magic Write, Magic Design, and Magic Switch. These tools allow users to generate content across text, images, and video, supporting multimodal workflows while maintaining simplicity. The suite enhances design processes for non-professionals by enabling content generation, format conversion, translation, and more, all through intuitive AI-driven prompts.

Lamini, an all-in-one LLM platform for enterprises to build open LLMs on proprietary data, uses Mistral-7B as one of the most popular open base models. Lamini enables Fortune 1000 customers across various industries to tune and deploy models in production efficiently, even on AMD GPUs with performance parity to NVIDIA GPUs. Mistral-7B's ease of use and high-quality results help customers move from proof-of-concept to production, deploying proprietary LLMs effectively.

HuggingFace introduced HuggingChat and HuggingChat Assistants, platforms that allow users to try out different open-source models and create customized assistants with unique personalities. By default, both platforms are powered by Mistral 8x7B, providing users with powerful, high-quality experiences and contextually relevant responses. Mistral 8x7B is currently the most popular model used on HuggingChat, enhancing user engagement.

Cloudflare, a leading content delivery network provider, serves Mistral 7B through their Workers AI platform, allowing users to run AI models on Cloudflare's global network. Mistral offers low latency, high throughput, and impressive performance, generating tokens up to 4x faster than Llama due to Grouped-Query attention. This enhances Cloudflare's AI offerings, enabling developers to build and deploy AI applications more efficiently.

Accenture uses Copilot Studio to create custom AI copilots that help employees automate routine tasks like checking project status. These AI-powered solutions make complex conversations easier by using relevant information, helping teams work more efficiently.

JourneyTEAM uses Copilot Studio's conversational and automation features to help clients replace repetitive tasks, allowing them to focus more on client engagement. This transformation enhances productivity and supports their mission of creating a positive, lasting impact.

Fusion5 uses Copilot Studio to offer consulting services that provide rapid Proof of Value for customers in generative AI. They tailor AI experiences for users within Microsoft 365, helping businesses establish a roadmap for further services and innovation.

Niantic, known for augmented reality games like Pokémon GO, uses Meta's Llama 2 to develop Peridot, a mobile AR game featuring virtual pets called "Dots." By integrating Llama 2, Dots exhibit dynamic and responsive behaviors, making interactions unique and enhancing user engagement. Additionally, Niantic employs computer vision with the Niantic Lightship ARDK to enable Dots to interact with the physical environment, creating a more immersive and lifelike gaming experience. This AI-driven approach accelerates development, reduces manual programming, and delivers personalized interactions, improving player satisfaction and engagement.

Amazon uses Anthropic's Claude with Amazon Bedrock to detect exaggerated claims in AWS Marketplace product listings, and is extending the AI to help partners improve listings through self-service. Additionally, auto-validations using Claude for product publishing have reduced publishing times from weeks to under an hour.

Klarna uses OpenAI's AI assistant to enhance the shopping and payments experience for 150 million consumers globally. The AI manages multilingual customer service, refunds, and returns, reducing query resolution time from 11 minutes to under 2 minutes. Additionally, Klarna integrates ChatGPT Enterprise across the organization, empowering 90% of employees to use generative AI for various tasks, significantly boosting efficiency.

Mendel has developed Hypercube, an AI platform that integrates Meta Llama 2-based LLM for natural language processing on clinical data. Hypercube uses a hypergraph semantic layer and the proprietary Eloquent language to enable querying and retrieving clinical information for applications like trial matching and patient cohorting. Llama 2 assists in converting natural language queries into Eloquent, bridging the gap for users unfamiliar with the query language.

Moderna uses ChatGPT Enterprise across the organization, empowering employees to create 750 custom GPTs in just two months. These GPTs, including Dose ID (for clinical data analysis), Contract Companion GPT (for contract summaries), and Policy Bot GPT (for internal policy queries), have significantly improved efficiency, accuracy, and decision-making without compromising privacy. The AI deployment is transforming operations in legal, research, clinical trials, and communications, enabling employees to focus on higher-level tasks.

OctoAI launched a production-ready Mixtral endpoint days after the model's initial announcement. Mixtral on OctoAI quickly became the most popular model on the platform, with about 9 out of every 10 new OctoAI sign-ups choosing Mixtral. Companies like Latitude Games and Otherside AI use Mixtral on OctoAI to deliver engaging LLM-powered experiences, generating billions of tokens daily.

Aive, an AI platform for video post-production, leverages Meta's Llama 3 to automate video summarization and smart clipping. By analyzing long-form content, Aive creates concise, engaging video formats optimized for platforms like Facebook and Instagram. This integration helps brands like LVMH, L'Oréal, and Stellantis streamline video workflows, reduce costs, and maximize content impact across channels within a secure and scalable AI environment.

Meditron, a suite of open-source large multimodal foundation models tailored for the medical field, leverages Meta's Llama 2 and 3 to assist with clinical decision-making and diagnosis in low-resource settings. Developed by researchers at EPFL and Yale School of Medicine in collaboration with humanitarian organizations like the International Committee of the Red Cross, Meditron provides evidence-based, contextually aware recommendations and diagnostic support. It automates tasks such as generating treatment regimens, determining radiation modalities, and assigning diagnostic codes, thereby enhancing clinical efficiency and improving patient outcomes. With over 30,000 downloads and top performance on medical benchmarks, Meditron democratizes access to advanced medical AI tools, enabling equitable healthcare innovation in underserved regions.

Amplitude's marketing and product teams use Perplexity to draft market landscape insights, accelerating research and improving the quality of strategic decisions.

Health editorial teams at Thrive Global are creating validated behavior change microsteps based on the latest peer-reviewed science with Perplexity.

Lambda uses Perplexity Enterprise Pro to streamline technical research and workflows across 15+ teams, including sales, content, and engineering. The AI provides reliable, cited information, helping users like technical writers, engineers, and sales reps accelerate documentation, sales outreach, and complex research tasks. This results in 475+ hours saved, faster task completion, and improved productivity.

Marketing and product teams at Amplitude use Perplexity to draft market landscape insights.

Product teams at Zoom use Perplexity’s Focus functionality for targeted search.

HP’s salesforce taps into Perplexity for rapid, in-depth prospect research, empowering them to craft compelling pitches and expedites the sales process.

Databricks uses Perplexity Enterprise Pro to accelerate R&D, enabling its engineering, marketing, and sales teams to execute faster and more efficiently. The company estimates saving 5,000 working hours monthly.

Fortune partnered with Accenture to develop Fortune Analytics, an AI-powered tool designed to analyze over 20 years of Fortune 500 and Global 500 financial data and related articles. This tool allows users to generate visualizations and receive ChatGPT-style responses, making data-driven insights more accessible. The platform is set to launch in beta on June 4, 2024, and aims to enhance enterprise decision-making by providing easy access to financial data and journalism archives.

Paradigm uses OpenAI’s GPT-4 API to evaluate clinical trial data sets, significantly improving the speed, accuracy, and cost of patient screening. By replacing traditional models with GPT-4, Paradigm achieved a 90% reduction in clinician time spent on model validation and a 10% increase in accuracy. This allows faster, more equitable access to clinical trials, especially for underserved patients with unstructured medical data.

Hugging Face is collaborating with Google across open science, open source, cloud, and hardware to enable companies to build their own AI with the latest open models from Hugging Face and Google Cloud hardware and software.