Use cases

Teacher Casey Cuny uses ChatGPT to identify patterns in anonymized short answers and plan targeted follow-up teaching. A demonstrated workflow takes about 30 minutes. Replacing two evenings of analysis is his reported comparison, not a controlled study.

ENEOS Materials uses a custom GPT for training feedback. Reported task timings fell from one to two hours to 20 seconds; a separate aggregation tool reportedly reduced aggregation time by 90%.

An incident.io engineer asked Claude Code to diagnose slow API generation and implement improvements incrementally. The author reports an 18% runtime reduction, equivalent to 30 seconds, for approximately five minutes of personal work and $8 in Claude usage. Commands and workflow examples are included.

Jamie's Farm Granola uses Gemini to organize stock tracking. Google reports 20 hours less paperwork per week; the article provides no measurement method.

A KDDI Agile Development Center practitioner uses Claude Code with browser MCP tools to collect article links across four calendars and prepare Slack messages. The author reports a routine taking five minutes fell to 30 seconds and provides execution logs and setup commands.

Remberg CEO David Hahn uses a reusable ChatGPT prompt to research prospective customers, covering business models, finances, competitors and news. He reports saving at least ten minutes per research task. The article provides no independent timing study.

A New Mexico franchise operator uploads restaurant cash records and bank data to ChatGPT Work for monthly reconciliation. He reports reducing the task from three hours to five to ten minutes. These are operator-reported timings, without an independent audit.

Consultant Adrian Rohnfelder uses three Claude Skills representing his own position, a critic and a perspective focused on affected people. They can debate and produce a joint recommendation. He reports one to two hours of weekly maintenance; no measured decision-quality improvement is provided.

tl;dv uses Claude to summarize meetings, combine findings across conversations and assess sales calls against playbooks. The company reports 50% faster summary generation. Its reported revenue growth is a business outcome, not a measured estimate of users' time savings.

The Wild Hare and agency No Biscuits used Gemini for character preparation and Veo 3 for animated promotional clips. The founder reports over 20 assets produced in an afternoon at under 10% of a traditional agency video's cost. A finished campaign video is linked.

Bison Coolers uses Gemini in Meet to capture notes and email action-item summaries. Its owners also use Gemini in Gmail to catch up on conversations and draft replies while traveling. The story describes everyday use without quantified savings.

Block uses Claude through Databricks in its open-source goose agent to generate SQL, query datasets and connect internal tools. The company reports adoption across 15 job profiles. Reported engineering time savings concern goose's broader use and do not isolate natural-language data access.

Capitec used Microsoft 365 Copilot for Finance for a reconciliation previously handled through numerous lookup formulas. The bank reports that one employee's process fell from six hours of weekly work to one minute per session. This is a specific workflow result, not a bank-wide average.

Commerzbank uses Gemini to transcribe advisory calls, extract relevant facts and draft German documentation for human review. The customer-authored vendor post reports reducing work taking over 60 minutes to a few minutes. It includes architecture details and a sample implementation.

Microsoft's .NET team assigned simpler refactoring tasks to Copilot in dotnet/android. Over 28 days, 82.4% of 17 agent pull requests were merged. Harder dotnet/maui tasks had a much lower merge rate, showing the importance of task selection.

Fullstory uses Gemini to summarize meetings and turn takeaways into short video scripts. Staff then create narrated videos with Google Vids for customer communication, prospecting and event follow-up. The customer describes faster production without quantifying savings.

The Original Tamale Co. used ChatGPT to build and publish a farmers-market tracker in one afternoon. The operator had no coding experience. No comparative development cost is provided.

Pennsylvania piloted ChatGPT Enterprise for writing, research, summarization and IT support. Participants reported average savings of 95 minutes per day while using it. The result is self-reported, rather than an independently observed productivity measurement.

Rakuten engineers use Claude Code for tests, components, bug fixes and documentation. Rakuten reports average feature delivery falling from 24 working days to five. A separate seven-hour vLLM implementation reached 99.9% numerical agreement with its reference, rather than general code correctness.

REWE digital's Lumi answers employee questions from internal knowledge and creates support tickets when needed. t3n reports 8,000 to 9,000 monthly inquiries, with 58% resolved automatically. Adoption improved after the team introduced a recognizable name and avatar.

At t3n, Claude creates Asana orders and services when tagged in a Slack thread containing a sales confirmation. Rules live in Confluence. The publisher reports over 80% automated entry during the first two months and up to 75% less handling time, with human checks.

Wealth.com's Ester extracts beneficiaries, distributions and provisions from estate documents, linking findings to the original PDF for advisor review. Its customer-authored vendor post reports average analysis below two minutes. Building the system required annotated documents and legal expertise.

AppFolio's Realm-X Messages uses Claude to draft property-management replies using resident history, policies and maintenance context. Beta customers reported saving about 11 hours weekly. The source also reports 26 seconds saved per message; the weekly figure is customer-reported.

ASUENE compared Gemini and GPT-4o models on 200 sample utility receipts, extracting transaction months, consumption and costs. Its tests found Gemini more than four seconds faster on average. The article includes sample documents, outputs and benchmark charts; deployment of the tested models is not established.

Campfire uses Claude to parse bank statements, identify discrepancies and answer financial questions through Ember. The company reports 90% less reconciliation time and monthly closes shortened by three days on average. These are customer outcomes reported by the vendor, without a published study.

Canva product managers and designers use Claude to build interactive components and visualize proposed features. Projects preserve design principles and context. The case documents internal use but does not quantify time saved specifically by prototyping.

Cisco engineers use Codex to review both human-written and AI-generated code. OpenAI reports review-time reductions of up to 50%. The customer example does not publish the sample size or measurement procedure.

Cox Automotive uses Claude in VinSolutions CRM to draft personalized emails and texts based on shopping interests and verified vehicle data. Users review generated content. Cox reports more than twice as many lead responses and test-drive appointments, without publishing a controlled evaluation.

Credit Saison uses a voice-based custom GPT for new sales staff to rehearse prospect conversations and receive feedback. Reduced dependence on experienced colleagues is claimed; no roleplay-specific saving is quantified.

Doctolib runs Claude Code agents through GitHub Actions for implementation, review, documentation and production checks. Its updated case reports 25% shorter change lead time between January and September 2026. This is an internal operational measurement; causal attribution to Claude is the company's interpretation.

Drive Social Media built Project Phoenix, a 95-step Zapier workflow that collects company and marketing data for sales preparation. It combines AI steps, ChatGPT and external APIs. The story documents production use; it does not provide an isolated time-saving measurement for Phoenix.

Freshy uses Zapier and AI to enrich and prioritize requests for a support operation covering over 2,800 websites. Related workflows identify sales opportunities and draft articles from resolved developer tasks. The agency reports scaling support without additional headcount, without a controlled savings calculation.

A Funai Soken consultant used Gemini-generated code to build tablet applications recording production work times. These records support labor-cost analysis. The vendor story says 80% of that consultant's manufacturing clients use the application; it does not quantify financial savings.

FutureHouse uses Claude for agents that search literature, create cited reports and assess whether hypotheses have precedents. Its Aviary framework is open source. The story claims literature reviews can take days instead of months but does not publish a quantified time-saving study.

After Git training, Grab designers used Cursor to implement interface corrections directly in code. Design managers reviewed hundreds of merged fixes, often on the same day. Faster delivery is reported, but no controlled time-saving comparison is provided.

Harvey offers Claude within Assistant, Vault and Workflows for contract analysis, due diligence and drafting. Lawyers can revise intermediate outputs before approval. Harvey reports deployment in under a month and strong internal benchmark performance, but no quantified client time saving in this case.

IG Group analysts use Claude to generate SQL and automate routine data work. The company reports approximately 70 analyst hours saved weekly across AI-assisted workflows. This is a company-reported aggregate, not a documented per-user or controlled result.

Iteratec's MAIA combines booking data with visual interpretation of reference invoices to handle differing formats and anonymization requirements. t3n reports 20% greater processing capacity and involvement in 90% of invoices. Full automation of 90% to 95% was a target, not the reported achieved result.

Jiaming Shang used Copilot to generate a deployed typing practice website with English, Chinese Pinyin and 300 Tang poems. Source code and demonstrations are linked. The account documents repeated corrections and limitations rather than measured productivity gains.

The University of Johannesburg runs MoUJi, an Amazon Lex-based assistant available through web, Messenger and WhatsApp. Its guest post reports reducing peak-period temporary agents from 120 to 60 and saving over R500,000. Generative AI extensions are discussed separately from these established chatbot results.

JR Kyushu's sales team uses Gemini to draft landing pages, add booking-search functionality and improve SEO internally. The customer reports an SEO score rising from roughly 40 to 70. The article does not establish a resulting increase in bookings.

Lokalise integrates Claude into its translation orchestration and editor, where translators accept, revise or reject suggestions. An internal A/B test reported an 82.6% acceptance rate for Claude 3.5 Sonnet. Broader 80% cost-saving claims concern Lokalise's multi-model platform.

micro1 uses Claude to conduct adaptive technical interviews and produce assessment reports. The vendor story says Legal Soft maintained 10,000 monthly screenings while reducing its interviewing team from 33 to 12. Claims of unbiased assessment are not supported by an independent fairness study in the source.

A Microsoft team connected Copilot to Azure DevOps through MCP to retrieve manual test cases and generate Playwright scripts. Hundreds of domain tests were automated. Reduced manual effort is reported without a measured time comparison.

Money Forward QA teams feed Jira tickets and Notion documents to Cursor through MCP. One agent drafts test cases and another produces Playwright scripts. The vendor story reports 70% less test-generation time without publishing the measurement method.

MOTHERHOUSE's marketing team configured a Gemini Gem with its brand values to review social posts and newsletters. The customer story reports substantially less review work, without quantifying the marketing-specific savings.

Navan used self-hosted Cursor Cloud Agents to migrate eight applications and 15 libraries to Spring Boot 4 in about one week. A comparable earlier migration took over a month. These are company-reported project durations, not a controlled experiment.

NRI combines Claude, review checklists and previous examples to identify issues in specialized Japanese business documents. Its system filters insignificant suggested edits. The company reports a 50% reduction in review time; detailed evaluation data and code are not published in the story.

Paraná's education department integrates Gemini into Redação Paraná for linguistic and argumentative feedback, alongside teacher review. The department reports grading time falling from 10 to 15 minutes to under three minutes per essay. The article does not isolate Gemini's contribution from the wider platform.

Port engineers delegate small refactorings, bug fixes and TODO items to Copilot while developing other features. Their retrospective reports better self-review and fewer minor review comments, without quantified savings.

Quantium built an AI-assisted system using Claude to help prepare business proposals and bid responses. The company claims up to 90% less work time, with complex proposals moving from weeks to hours. The case study does not provide a controlled comparison or public implementation.

Quillit uses Claude to summarize interviews, query qualitative datasets and attach references to transcripts or recordings. The vendor claims up to 80% less report-writing time. It reports internal citation accuracy of 89% to 98% with Claude 3.5, so references still require checking.

A READYFOR engineering survey documents Claude Code use for coding, Git operations and pull requests. Among 12 respondents, 66% reported saving one to two hours daily, while 42% reported more fatigue. These are self-reported savings rather than controlled productivity measurements.

Teacher Sarah Dully uses ChatGPT to update examples, simplify directions, draft sentence starters and create alternative assignments. The account includes literature-linked mathematics activities. Reduced preparation burden is described without quantified savings.

A Schoo infrastructure engineer used Cursor Agent to investigate a manually maintained AWS account and import about 350 required resources into Terraform. The account contained more than 1,500 listed resources. The post provides technical steps but no measured time saving.

Law&Company uses Claude with retrieval of legal sources in SuperLawyer for research, precedent analysis, drafting and document questions. In a user survey, 92.5% reported saving time. These are self-reported results rather than independently measured lawyer productivity.

Thomson Reuters uses Claude through Amazon Bedrock within CoCounsel to synthesize tax and legal reference material using retrieval. Expert validation supports its quality process. The story includes customer expectations of time savings, rather than a measured tax-research productivity result.

Vanta uses Claude to turn failed compliance tests into tailored remediation instructions, Terraform templates or cloud commands. Its internal evaluation reported roughly 15% better performance than alternatives on Terraform remediation. Customer time savings are described qualitatively.

The Village Store's two-person team uses Gemini in Sheets for accounting. Google's customer roundup gives no quantified savings.

Zurich uses Copilot for Sales in Outlook and Teams to maintain CRM records and assist customer-facing correspondence, including non-native-language drafting. The company projects approximately 14,000 hours saved over the following year. That figure is a forecast, not a completed annual measurement.

Argentina-based AdGoat uses Gemini with Cloud Run for content analysis, bidding and audience targeting. Google describes the deployment but gives no quantified campaign improvement.

Advolve uses Claude to orchestrate advertising content, campaign deployment, validation and optimization. Named customers include Brazilian companies iFood and Cogna. The vendor story reports 90% less operational work and 15% higher return on ad spend without publishing an independent evaluation.

Aldo used Copilot's coding agent to add optional header images to an existing Markdown blog. The post links the agent session and describes compatibility and testing requirements. No task-specific time saving is measured.

ANDRITZ captures expert interviews, videos and documents, then organizes the material in SharePoint with AI agents. Employees retrieve structured guidance for onboarding and operations. Microsoft reports 6,000 active Copilot users and 504 active agents, but no measured onboarding-time reduction.

Australian Payments Plus uses ChatGPT Enterprise to locate payment specifications and prepare member communications. Staff review the results. Faster responses are claimed without a task-specific timing.

An Asahi Shimbun workshop demonstrated Cursor workflows for GitHub issue creation, project fields, deadline reminders and meeting documentation. Shell examples are provided. This was an internal hands-on exercise, with no measured productivity outcome.

BBVA's Credit Analysis Pro GPT extracts information from annual reports, ESG disclosures and news for risk analysts. Faster assessment is claimed; the story gives no workflow-specific savings measurement.

Benchling's Data Entry Assistant uses Claude to transform unstructured study reports and instrument files into searchable records. The vendor reports savings of up to two weeks on complex data transformation. The story does not disclose the sample or measurement procedure behind that maximum.

DBS deploys chains of 70 to 80 specialized agents to synthesize financial, regulatory and market information for credit memoranda. Google describes production use with human oversight, without quantified time savings.

Faire uses Cursor Cloud Agents and automations for migrations, CI failures, bug triage and code review routing. It reports more than 2,000 automated agent runs weekly and doubled PR throughput. The cited 18-month migration baseline was a prior estimate.

GitHub assigns Copilot issues to remove obsolete feature flags, conditional branches and associated tests across its platform codebase. Engineers review the proposed pull requests. The article describes reduced maintenance work but gives no savings measurement for this task.

Grab's Merchant AI Assistant uses Claude and merchant-specific data to answer platform questions and suggest menu, promotion and business improvements across Southeast Asia. Grab reports a 5.7 percentage-point rise in issue resolution and 25% less negative sentiment.

incident.io's marketing chief used Claude Code to build PagerTron, a game with stored high scores, for a campaign aimed at Anthropic. The post links the campaign and shows the game. It documents a shipped marketing asset but provides no measured conversion or time-saving result.

Jewelry designer Lina Prades built management applications with ChatGPT despite having no technical background. One supports daily operations. The source provides no measured financial or time outcome.

The Patrick J. McGovern Foundation uses Claude in Grant Guardian to extract financial indicators from nonprofit documents and prepare standardized profiles. The republished Anthropic case reports use by over 100 US philanthropies and qualitative time savings. Funding decisions remain with people.

During a healthcare software project, Microsoft ISE used Copilot to follow development guides, create required files and identify missing setup steps or tests. The team reports easier onboarding and time savings but provides no controlled measurement for this workflow.

Minnesota's Enterprise Translations Office uses ChatGPT Team for translation services serving multilingual communities. OpenAI reports lower costs and faster turnaround, but the cited account gives no numerical results or published quality evaluation.

Moderna uses ChatGPT Enterprise to extract facts and assumptions from evidence packages and draft target product profile sections. It reports that a core analytical step sometimes fell from weeks to hours. No detailed timing study is published in the case.

Newfront uses Claude in a benefits assistant that answers questions using each client's plan details. In an internal test, Claude answered 12 of 20 difficult questions correctly. Newfront claims HR teams recover one month annually; the story provides no measurement method for that saving.

Okta engineers used Copilot for routine coding and testing. A study involving 74 survey participants found improved motivation and perceived skills, but no significant change in pull requests, code volume or cycle time. Later results for Claude Code are separate.

Orange uses Claude in a manga-localization pipeline combining translation, image processing and professional editing. The company reports generation time dropping from more than ten hours to under two per volume. It says translators retain 50% to 75% of Claude's proposed translations.

A practitioner released a Windows application combining Whisper transcription with Gemini summaries. The post documents recording, importing audio and generating minutes. It provides source code and a downloadable application, but no measured time savings.

Rising Academies uses Claude in Rori to interpret students' answers and guide mathematics conversations over WhatsApp. The source reports reaching more than 150,000 students across African countries. It cites a 0.3 standard-deviation learning effect for Rori, without isolating Claude's contribution.

Google reports an eight-week randomized trial of Gemini Guided Learning across 48 classrooms in Sierra Leone. Students using the tool improved externally assessed mathematics scores by 0.26 standard deviations. This is a measured learning result, not evidence of financial or staff-time savings.

A Utah court filing contained a nonexistent case generated through ChatGPT. According to the court account reported by heise, the responsible lawyer had not checked the draft. The court ordered cost reimbursement and a $1,000 payment. This is a documented failure case.

Wayfair researchers used Cursor agents to implement and evaluate 110 variants of a product-attribute validation model. The company reports a measured 94% inference-cost reduction, with the selected architecture entering production. A later sprint reported a further 90% reduction.

Simon Willison used Claude to build a Python utility converting Claude Code sessions into navigable HTML. He publishes source code, example output and development transcripts. Codex handled a separate API reverse-engineering step. No quantified development-time saving is reported.

AES uses a Claude-based multi-agent system on Google Cloud to process safety documents, break down audit tasks and generate reports. The company reports reports taking about an hour instead of two weeks. Claimed cost and accuracy improvements lack a published evaluation method.

IMDEX used Cursor to rebuild fragmented geological data applications into an integrated platform delivered in eight months. Its comparison with two to three years and 20 additional engineers is a counterfactual estimate, not a measured alternative implementation.

Lotte Homeshopping's Moni assistant uses Claude through Sendbird to interpret quality documents and answer suppliers' routine questions around the clock. The company reports an average of 53 questions daily and 30% to 40% fewer inquiries reaching QA staff.

National Australia Bank's merchant services team used Cursor to ship a hardware-independent payment application in three weeks. The original scope was four months. The story reports this delivery against a planning estimate, without a parallel comparison project.

Novo Nordisk uses Claude in NovoScribe to draft clinical documentation from expert-approved text and study-specific variables. Claude Code helped build the platform. The company reports 90% less writing time for clinical study reports; human review and approval remain part of the process.

Project Themis uses Gemini prompts to identify inconsistencies, structure arguments and prepare document drafts. The practitioner publishes prompts and a repository, with human review in the workflow. The article provides no validated legal-accuracy or savings study.

Ramp connects Claude Code to tools such as Datadog and Sentry through MCP to gather logs, errors and metrics for on-call engineers. Early observations suggested initial triage could take up to 80% less time. The source presents this as developing tooling, not an established measured saving.

SIGNAL IDUNA describes Co SI, its Gemini-based internal AI platform, and a specialized assistant for health-insurance knowledge. Staff use these systems to access information needed for customer inquiries. The interview describes operational adoption without supplying a controlled savings measurement.

A practitioner at Ururu prototyped a workflow that retrieves Slack threads, summarizes them with Gemini and supplies the summaries to NotebookLM. The post includes Apps Script code and example questions with source links. Reduced inquiry volume remains an expectation rather than a measured result.

A developer built a browser application that sends receipt images to Gemini, returns editable structured records and exports CSV files compatible with Money Forward accounting. A hosted demo and repository are linked. The author reports API retries but provides no accuracy benchmark.

Amira uses Claude offline to prepare comprehension questions, hints and response paths. Content is reviewed before students interact with a curated response system. Claude does not perform live inference during student sessions. Reported reading gains concern Amira overall, not Claude alone.

Anthropic's legal team used Claude Code to prototype a phone-tree system directing colleagues to the appropriate lawyer. The example documents non-developers building an internal coordination tool. The article identifies it as a prototype and provides no quantified savings.

A University of Arizona astrophysicist uses Codex to propose numerical approximations and algorithms for plasma simulations. He reports generating candidates in minutes instead of days. Each approximation still requires scientific implementation and verification.

Australian developer Geoffrey Huntley ran Claude Code in repeated loops for three months to build Cursed, a programming language using Gen Z vocabulary. The compiler and demonstrations were published. This was a learning experiment with incomplete components, without measured labor savings.

Two Nokia engineers used Cursor to analyze more than 50 million lines of code in two weeks and prepare an architectural decomposition plan. The comparison with a dozen specialists working for months is an estimate, rather than an observed control project.

Deutsche Telekom and Gleiss Lutz helped develop Harvey Shared Spaces for joint regulatory and transactional work. The vendor describes shared matter context, document analysis and drafting. The feature was in early access at publication; customer-specific time savings were not quantified.