# Agents in the Wild

> A searchable directory of real-world AI agent use cases. Every entry links to its evidence, says who reported the result, and lists the caveats.

Each use case lists: organization, tools, reported outcome, evidence type (who made the claim), caveats, and source links.
Full data as JSON: https://agentsinthewild.co/api/use-cases.json

## Customer Support

- [Octopus Energy's AI email assistant Arlo scored higher satisfaction than human replies](https://agentsinthewild.co/use-cases/octopus-energy-arlo-emails): 76% customer satisfaction for AI replies vs 72% for comparable human replies, on ~8,000 emails a week (4% of UK customer emails). (Evidence: company-reported; tools: Arlo, Kraken)
- [Salesforce runs its own support on Agentforce and cut support headcount from ~9,000 to ~5,000](https://agentsinthewild.co/use-cases/salesforce-agentforce-support): About 1.5M customer conversations handled by agents, with CSAT about the same as humans; support headcount went from ~9,000 to ~5,000. (Evidence: company-reported; tools: Agentforce)
- [Lyft cut average support resolution time by 87% with a Claude-based assistant](https://agentsinthewild.co/use-cases/lyft-claude-customer-care): Average resolution time fell by more than 87%, with thousands of requests resolved per day (company claim). (Evidence: vendor-case-study; tools: Claude, Amazon Bedrock)
- [Intercom's Fin AI agent resolves about two-thirds of support conversations on average](https://agentsinthewild.co/use-cases/intercom-fin-resolution-rate): Average resolution rate of 66–67% across about 6,000 customers, up from about 51% at launch. (Evidence: vendor-case-study; tools: Fin, Claude)
- [Klarna's AI assistant took two-thirds of support chats in its first month](https://agentsinthewild.co/use-cases/klarna-ai-assistant-customer-service): 2.3M conversations in the first month, about two-thirds of all service chats, the work of roughly 700 full-time agents (company claim). (Evidence: company-reported; tools: OpenAI GPT-4)
- [DoorDash built a voice AI agent for Dasher support calls with answers in under 2.5 seconds](https://agentsinthewild.co/use-cases/doordash-dasher-voice-support): A voice agent ready for live testing in about two months, answering in 2.5 seconds or less, with 50x more testing capacity. (Evidence: vendor-case-study; tools: Claude 3 Haiku, Amazon Bedrock, Amazon Connect, Amazon Lex)

## Software Engineering

- [Stripe's Minions coding agents produce 1,000+ merged pull requests every week](https://agentsinthewild.co/use-cases/stripe-minions-coding-agents): More than a thousand pull requests merged each week are fully agent-produced and human-reviewed. (Evidence: company-reported; tools: Minions (internal), Goose)
- [Spotify's Honk background agent merges fleet-wide code migrations: 1,500+ PRs and counting](https://agentsinthewild.co/use-cases/spotify-honk-background-agent): 1,500+ merged AI-generated PRs; 1,000 merged PRs took three months at first and about ten days later on. (Evidence: company-reported; tools: Claude Code, Fleet Management (internal))
- [Aider, the open-source coding agent, writes most of its own code](https://agentsinthewild.co/use-cases/aider-writes-its-own-code): Aider's release notes report that Aider wrote up to 88% of the code in a release (v0.86.0, August 2025). (Evidence: company-reported; tools: Aider)
- [Goldman Sachs pilots Devin as an AI engineer next to its 12,000 developers](https://agentsinthewild.co/use-cases/goldman-sachs-devin-pilot): Pilot started with plans for hundreds of instances; the CIO expects 3–4x developer output on routine code updates. (Evidence: press-report; tools: Devin)
- [Cloudflare built an OAuth 2.1 library with Claude and published the prompts in git](https://agentsinthewild.co/use-cases/cloudflare-oauth-library-claude): A working OAuth 2.1 provider library for Cloudflare Workers, written mostly by Claude and reviewed by Cloudflare engineers. (Evidence: independent; tools: Claude)
- [Airbnb migrated 3,500 test files with an LLM pipeline in 6 weeks instead of 1.5 years](https://agentsinthewild.co/use-cases/airbnb-enzyme-rtl-migration): ~3,500 test files migrated in 6 weeks; about 97% fully automated and the rest finished by hand. (Evidence: company-reported; tools: LLM pipeline (custom))
- [Google uses LLMs for internal code migrations and reports about 50% less migration time](https://agentsinthewild.co/use-cases/google-llm-code-migrations): Across 39 migrations, about 74% of code changes were LLM-generated; developers estimated a 50% cut in total migration time. (Evidence: research-paper; tools: Gemini (internal))
- [Rakuten engineer let Claude Code work alone for 7 hours on a vLLM feature](https://agentsinthewild.co/use-cases/rakuten-claude-code-vllm): Seven hours of mostly autonomous coding in one run. Rakuten reports 99.9% numerical accuracy versus the reference method. (Evidence: vendor-case-study; tools: Claude Code)
- [Amazon upgraded 30,000 internal apps to Java 17 with Amazon Q Developer](https://agentsinthewild.co/use-cases/amazon-q-java-upgrades): 30,000 apps upgraded. Amazon estimates 4,500 developer-years saved and $260M a year from performance gains. (Evidence: company-reported; tools: Amazon Q Developer)
- [Nubank used Devin to split an 8-year-old ETL monolith into sub-modules](https://agentsinthewild.co/use-cases/nubank-devin-etl-migration): Cognition reports 8–12x engineering-time efficiency and over 20x cost savings on the data-class migrations. (Evidence: vendor-case-study; tools: Devin)

## Security

- [Google DeepMind's CodeMender agent upstreamed 72 security fixes to open-source projects](https://agentsinthewild.co/use-cases/google-deepmind-codemender): 72 security fixes upstreamed to open-source projects in six months, including projects as large as 4.5 million lines of code. (Evidence: company-reported; tools: CodeMender, Gemini Deep Think)
- [XBOW's autonomous pentesting agent reached #1 on HackerOne's US leaderboard](https://agentsinthewild.co/use-cases/xbow-hackerone-leaderboard): #1 on the HackerOne US leaderboard, with 1,000+ reports in a few months and 130+ confirmed and fixed by program owners. (Evidence: press-report; tools: XBOW)
- [Google's Big Sleep agent found an exploitable SQLite bug that fuzzing missed](https://agentsinthewild.co/use-cases/google-big-sleep-sqlite): Found a previously unknown, exploitable memory-safety bug in SQLite; developers patched it the same day, before any release. (Evidence: research-paper; tools: Big Sleep, Gemini 1.5 Pro)

## Sales & Marketing

- [Anthropic's growth marketer cut ad creation from 30 minutes to 30 seconds with Claude Code](https://agentsinthewild.co/use-cases/anthropic-growth-marketing-ads): Work that took about 30 minutes per ad now takes about 30 seconds; the plugin can make up to 100 ad variations at once. (Evidence: vendor-case-study; tools: Claude Code, Figma, Google Ads)
- [Amazon says its Rufus shopping assistant is on track for $10B+ in extra yearly sales](https://agentsinthewild.co/use-cases/amazon-rufus-shopping-assistant): Amazon expects over $10B in annualized incremental sales; shoppers who used Rufus were 60% more likely to buy. (Evidence: company-reported; tools: Rufus)
- [Vercel's lead agent took inbound qualification from a 10-person team to about one person](https://agentsinthewild.co/use-cases/vercel-lead-qualification-agent): Inbound SDR work went from about 10 people to 1 plus a part-time person; the agent costs about $5,000 a year to run (as reported). (Evidence: press-report; tools: Custom agent (Vercel AI SDK))

## Finance & Operations

- [BNY runs 100+ AI 'digital employees' with their own logins and human managers](https://agentsinthewild.co/use-cases/bny-digital-employees): Over 100 digital employees in production by October 2025, later reported as 130+, each with a human supervisor. (Evidence: press-report; tools: Eliza (internal), OpenAI)
- [Ramp's policy agent reviews expenses against company policy for 1,000+ finance teams](https://agentsinthewild.co/use-cases/ramp-policy-agent-expenses): Used by 1,000+ finance teams within six months; Ramp claims 99% approval accuracy and 7x more out-of-policy spend caught. (Evidence: vendor-case-study; tools: Ramp Policy Agent, OpenAI reasoning models)
- [Morgan Stanley's Debrief agent takes notes in advisor meetings and drafts the follow-up email](https://agentsinthewild.co/use-cases/morgan-stanley-debrief): Rolled out to the firm's financial advisors in 2024; the firm cites about 30 minutes saved per client meeting (testimonial). (Evidence: company-reported; tools: OpenAI GPT-4, Salesforce, Zoom)

## Data & Analytics

- [Uber's QueryGPT turns plain-English questions into SQL, cutting query time from ~10 to ~3 minutes](https://agentsinthewild.co/use-cases/uber-querygpt-text-to-sql): Uber estimates a query that took about 10 minutes to write by hand now takes about 3 minutes; the limited release reached ~300 daily active users. (Evidence: company-reported; tools: QueryGPT (internal), RAG, LLM)
- [Pinterest's text-to-SQL agent helps analysts find tables and write queries 35% faster](https://agentsinthewild.co/use-cases/pinterest-text-to-sql): Pinterest reports 35% faster task completion for SQL writing, and first-shot acceptance rising from 20% to over 40%. (Evidence: company-reported; tools: Querybook, RAG, LLM)

## Research

- [Sakana's AI Scientist-v2 wrote a paper that passed peer review at an ICLR workshop](https://agentsinthewild.co/use-cases/sakana-ai-scientist-peer-review): One of three fully AI-generated submissions cleared the acceptance bar at the ICBINB workshop (scores 6, 7, 6). (Evidence: research-paper; tools: The AI Scientist-v2)
- [Google's AI co-scientist proposed the same superbug hypothesis an Imperial College team spent years on](https://agentsinthewild.co/use-cases/google-ai-co-scientist): The system's top hypothesis matched the Imperial team's unpublished finding about how bacteria spread resistance genes. (Evidence: research-paper; tools: AI co-scientist, Gemini 2.0)

## Legal

- [Allen & Overy rolled out Harvey to 3,500 lawyers after a 40,000-query trial](https://agentsinthewild.co/use-cases/allen-overy-harvey-legal): About 3,500 lawyers asked around 40,000 questions during the trial, followed by a firm-wide rollout across 43 offices. (Evidence: company-reported; tools: Harvey, OpenAI GPT-4)

## Healthcare

- [Kaiser Permanente rolled out Abridge ambient AI notes to 24,000+ doctors](https://agentsinthewild.co/use-cases/kaiser-abridge-ambient-notes): One of the largest generative AI rollouts in healthcare: 24,000+ physicians across ~40 hospitals and 600+ offices. (Evidence: press-report; tools: Abridge)

## HR & Recruiting

- [IBM's AskHR agent resolves 94% of common HR questions without a person](https://agentsinthewild.co/use-cases/ibm-askhr-agent): 94% containment of common questions, 75% fewer HR support tickets since 2016, and a 40% cut in HR operating costs over four years. (Evidence: company-reported; tools: AskHR (watsonx Orchestrate))
- [LinkedIn's Hiring Assistant agent finds candidates and drafts outreach for recruiters](https://agentsinthewild.co/use-cases/linkedin-hiring-assistant): Launched to charter customers including AMD, Canva, Siemens, and Zurich Insurance, aiming at the 20+ hours a week recruiters spend on admin. (Evidence: company-reported; tools: LinkedIn Hiring Assistant)

## Internal Tools

- [McKinsey's Lilli platform is used regularly by over 70% of the firm's staff](https://agentsinthewild.co/use-cases/mckinsey-lilli-platform): More than 70% of the firm uses Lilli regularly, with millions of prompts answered (company figures). (Evidence: company-reported; tools: Lilli (internal))
- [Uber's Genie answers on-call questions in Slack and saved ~13,000 engineering hours](https://agentsinthewild.co/use-cases/uber-genie-on-call-copilot): 70,000+ questions answered across 154 Slack channels, with an estimated 13,000 engineering hours saved. (Evidence: company-reported; tools: OpenAI embeddings, RAG, Slack)

## Optional

- [How we verify](https://agentsinthewild.co/about): evidence labels and editorial rules
- [Categories](https://agentsinthewild.co/categories): Customer Support, Software Engineering, Security, Sales & Marketing, Finance & Operations, Data & Analytics, Research, Legal, Healthcare, HR & Recruiting, Internal Tools