AI use cases
3058 ways people actually use AI, compiled from the videos in this
library. Each names the job and why it is worth doing, and links to the
moment in the source video.
-
-
Support vulnerability tracking and patch-management workflows
— Help security teams handle growing vulnerability volumes and prioritize remediation.
-
Discover vulnerabilities
— Increase the speed, scale, and depth of vulnerability research.
-
Assist with password hygiene and management
— Supplement human security work rather than replace people.
-
-
Running small-business operations
— Handle recurring administrative work and interface with systems of record, customers, insurers, and other agents.
-
Digitizing paper-based payments and records
— Convert checks, itemized invoices, and file-cabinet information into digital formats that agents can process.
-
Learning domain-specific workflows
— Use historical ERP data, staff feedback, office procedures, and documentation to improve agents' handling of edge cases.
-
Automating dental-practice administration
— Perform insurance-payment reconciliation, patient billing, claims work, and related operational tasks without a human in the loop.
-
Memory and contextual retrieval
— Use stored conversations, files, computer activity, and Chronicle-derived context to personalize responses and surface relevant information.
-
Personal planning and household management
— Manage calendars, workouts, meal planning, meal tracking, budgeting, and other personal productivity tasks over time.
-
Scheduled meme generation
— Review Slack, Docs, and other activity to create memes and surface unexpected connections.
-
Performance-review context gathering
— Search code, reviews, Slack, and other activity to identify what people have done and surface contributions for human review.
-
Research and self-play experimentation
— Create a playable board game, train AIs through self-play, explore parameters, and produce research artifacts and benchmarks.
-
Creating interactive Sites
— Turn reports, prototypes, presentations, spreadsheets, and research outputs into flexible interactive websites.
-
Creating spreadsheets and high-fidelity artifacts
— Generate and edit work products such as a retirement calculator while allowing users to iterate on a rendered result.
-
Locating a missed package
— Use an agent to inspect an image, search nearby listings, and identify the likely apartment complex where a package was left.
-
Automated testing
— Use AI to test software, an early application Akshay explored before large language models and vision language models.
-
Revenue generation and customer targeting
— Use business and neighborhood context to identify prospects, support inbound demand, and create new revenue opportunities.
-
Business workflow orchestration
— Coordinate operational rules, scheduling, service deployment, labor, and customer interactions across essential-service businesses.
-
Service triage and technician matching
— Determine whether a company can handle a request, who should handle it, when it should happen, and how to optimize the outcome.
-
Customer communication and intake
— Agents communicate with customers by voice, text, or website and gather information about their homes, records, and needs.
-
-
Train a neural approximation to an expensive quantum-chemistry simulator.
— Create a much faster validation model for screening scientific possibilities.
-
Use multiple agents or approaches with an evaluator to search possible solution paths.
— Improve reliability and performance in long-running agent workflows by retaining promising approaches and discarding failures.
-
Provide models with skills and guidelines for performance optimization.
— Enable an agent to measure benchmarks, modify code, evaluate performance, and iterate on improvements.
-
Automate machine-learning problem decomposition and experimentation.
— Run subproblems in tight loops, combine results, and produce improved machine-learning systems.
-
-
Investigate transaction context
— Improve categorization by considering the vendor, the team member who made the purchase, and notes or memos.
-
Categorize business transactions
— Automate monthly bookkeeping by proposing a category for each transaction and providing the rationale.
-
Prepare marketing assets
— Create advertising creatives and postcard designs for customer acquisition.
-
Analyze newsletter performance
— Review open rates, identify engaging subject lines, and inform future newsletter topics using actual performance data.
-
Generate personalized content
— Produce customized bedtime stories based on a child's name and preferences.
-
Create websites
— Generate websites for local businesses using publicly available business information.
-
Conduct business and market research
— Find companies, gather contact information, analyze marketing strategies, and compile findings into documents or spreadsheets.
-
Engineering assistance
— Help people understand systems, detect anomalies, triage failures, and produce code while humans evaluate quality and behavior.
-
Personalization and discovery
— Match the right title to the right person at the right moment across an expanding catalog of film, TV, games, live content, and podcasts.
-
Post-production modification
— Help filmmakers relight, reframe, reshoot, and change dialogue after filming while keeping creative direction with the filmmaker.
-
Content production and creative ideation
— Support pre-visualization, promotional assets, localization, subtitles, dubs, trailers, images, artwork, and other content workflows.
-
Data analysis, modeling, and information distillation
— Search Netflix's historical experiments, consumer research, metrics, and business knowledge to generate initial insights and hypotheses.
-
Prototyping and initial code generation
— Enable product, design, and data science teams to develop and test ideas faster before engineering productizes and scales them.
-
Create repeatable visual styles for brand assets.
— AI can generate cohesive photography, illustrations, backgrounds and supporting assets around an assigned aesthetic.
-
Generate prompts for other AI tools.
— ChatGPT produces more specific prompts for tools such as Runway, Midjourney and Nanobanana Pro.
-
Generate design sections from visual references.
— A screenshot of a preferred section can guide Claude in creating a similar website component.
-
Support coding and prototyping.
— Claude is used as the primary tool for coding and prototyping.
-
Generate images and video assets.
— Midjourney and Krea are used for image and video generation, while AI expands the types of assets a solo designer can produce.
-
Generate product requirements, features and user flows for complex software projects.
— Claude performs much of the UX thinking and reduces the need for extensive requirements-gathering conversations.
-
Route queries by difficulty
— Apply fast single-pass inference to easy requests and a full reasoning pipeline to harder requests.
-
Solve difficult reasoning problems
— Use additional inference-time computation, intermediate reasoning, search, or multiple reasoning paths to improve the probability of an accurate response.
-
Reviewing the Coldcard software repository for vulnerabilities
— The speaker says an AI system found the key-generation vulnerability quickly, reportedly after one prompt.
-
Execute work under human-set goals
— Enable people to focus on goal setting and supervision while ensuring AI produces the intended results.
-
Produce code and other work at much greater speed
— Allow people to spend more time on product purpose, customer demand, architecture and higher-level decisions.
-
Perform entry-level grunt work and other routine functions
— Reduce the need for people to perform work that AI can handle and free people for higher-value responsibilities.
-
-
Recognize contextual patterns in code
— Use large language models trained on programming languages, documentation, APIs, and developer discussions to provide more meaningful analysis.
-
Recommend fixes, improvements, explanations, and corrections
— Help developers address identified issues and understand why they occur.
-
Analyze software and scan code
— Detect security vulnerabilities, coding-standard violations, style inconsistencies, performance issues, and other problems.
-
Act on a user's behalf
— APIs can enable AI to perform actions such as booking a flight or sending an email.
-
Provide answers using a company's private information
— An API can feed private company information into an AI model so it can answer specific questions.
-
Receive instructions through text messages
— Allow the user to direct the AI agent via iMessage