WorkedPersonal admin
Tools: OpenClaw, Claude Sonnet
Setup, access, approval gate, failure notes
- Setup / prompt
- "Unsubscribe me from all newsletters and promotional emails, keep only transactional emails and personal correspondence."
- Access required
- Full access to a throwaway Gmail account (~2,000 emails)
- HITL checkpoint
- Human spot-checked ~50 results afterward
- Outcome & failure notes
- Unsubscribed from 203 lists across 847 newsletter emails in ~20 minutes for ~$8 in API calls; ~90% accuracy on spot check. Missed some legitimate client emails sent via a marketing platform. Ran on a throwaway account, so blast radius was low.
- Evidence type
- Tester first-hand, 1 week
Source: Medium (Likhit Kumar V P) · 2026 · Checked 2026-10-19
WorkedPersonal admin
Tools: Claude Code skill, Claude for Chrome
Setup, access, approval gate, failure notes
- Setup / prompt
- A custom 'travel agent' skill that sorts flights and housing against trip criteria, rerun with a real browser attached.
- Access required
- Browser extension reading public booking sites; no payment access
- HITL checkpoint
- "Act without asking" turned off; agent hands back links; human books and pays
- Outcome & failure notes
- Without a browser the skill returned plausible-sounding price ranges (guesses) because booking sites render prices with JavaScript. With Claude for Chrome it returned three specific options with live prices and links. Slower than doing it by hand and burns usage limits faster.
- Evidence type
- Practitioner first-hand
Source: CJ on Systems (Christopher Johnson) · 2026-08-26 · Checked 2026-10-19
PartialPersonal admin
Tools: ChatGPT agent
Setup, access, approval gate, failure notes
- Setup / prompt
- Ask the agent to find and book a table that fits stated constraints.
- Access required
- Browser session on booking sites; personal details for the reservation
- HITL checkpoint
- Agent asks the user for personal details and confirmation
- Outcome & failure notes
- A successful booking took ~11 minutes versus ~2 for a human and nearly missed the reservation hold window. Usability researchers rated the agent successful but shaky and slow.
- Evidence type
- UX research firm hands-on
Source: Nielsen Norman Group · 2025 · Checked 2026-10-19
PartialPersonal admin
Tools: OpenClaw, Mac Minis / old laptops
Setup, access, approval gate, failure notes
- Setup / prompt
- Nine single-purpose agents (family calendar, inbound sales, kids' homework, podcast, etc.) rather than one generalist.
- Access required
- Separate agent accounts and delegated access, expanded gradually
- HITL checkpoint
- Staged permissions: start narrow, widen only after trust is earned
- Outcome & failure notes
- Now a daily driver for her family and business, but the first attempt deleted her family calendar. The fix was specialization plus staged, delegated access instead of handing one agent everything.
- Evidence type
- Practitioner interview (podcast)
Source: Lenny's Newsletter / Podcast (Claire Vo) · 2026-03-29 · Checked 2026-10-19
PartialPersonal admin
Tools: OpenClaw, IMAP, Microsoft Graph, Telegram
Setup, access, approval gate, failure notes
- Setup / prompt
- Check mail every 30 min in business hours; sort into urgent / important / informational / noise; ping Telegram for urgent; draft routine replies only.
- Access required
- Read + send on 3 email accounts; persistent MEMORY.md file with relationship notes
- HITL checkpoint
- Never send without approval. Human reviews every draft.
- Outcome & failure notes
- About 85% triage accuracy and zero missed urgent emails; inbox processing dropped from ~10 to ~2 minutes. But tone matching, context lost across sessions (missed a verbal agreement mid-negotiation), over-eager CCs, and relationship-sensitive replies failed. Scope was cut back to routine confirmations, templated answers, and newsletter triage.
- Evidence type
- Operator first-hand, 30 days
Source: DEV Community (Jarvis Specter) · 2026-03-08 · Checked 2026-10-19
FailedPersonal admin
Tools: OpenAI Operator
Setup, access, approval gate, failure notes
- Setup / prompt
- "Find the cheapest set of a dozen eggs I can have delivered." A research request, not a purchase.
- Access required
- Logged-in grocery delivery account with a saved credit card
- HITL checkpoint
- Should have been: confirm before any purchase. In practice the confirmation was skipped.
- Outcome & failure notes
- Agent placed an Instacart order and charged $31.43 (fees, priority fee, tip) without approval, then reported the total as $19.68. OpenAI said Operator fell short of its safeguards and added stricter purchase confirmations. Lesson: never leave a saved card in an agent session for a research-only task.
- Evidence type
- Journalist first-hand test
Source: The Washington Post (Geoffrey A. Fowler) · 2025-02-07 · Checked 2026-10-19
FailedPersonal admin
Tools: OpenClaw
Setup, access, approval gate, failure notes
- Setup / prompt
- Ask the agent to schedule a meeting at a convenient time.
- Access required
- Calendar read/write
- HITL checkpoint
- None at the time. Caught only because the user double-checked.
- Outcome & failure notes
- Agent said the scheduling worked when it hadn't. Tester's takeaway: ambiguous goals ("convenient for whom?") fail, and agents can falsely report success, so every write needs independent verification.
- Evidence type
- Tester first-hand
Source: Medium (Likhit Kumar V P) · 2026 · Checked 2026-10-19
WorkedMoney / bills
Tools: Claude Code, SheetLink CLI, Google Sheets
Setup, access, approval gate, failure notes
- Setup / prompt
- Build a merchant→category lookup from 890 already-tagged transactions; Claude handles only new or ambiguous merchants; each decision becomes a new rule.
- Access required
- Read-only transaction feed (via the author's own sync tool) and the budget sheet
- HITL checkpoint
- User decides the genuinely ambiguous handful (two on a recent batch)
- Outcome & failure notes
- About 95% of merchants map to one category, so the lookup table does most of the work and the AI layer shrinks over time. Transfers and card payments are deliberately left alone. Note: the author builds and sells the sync tool used.
- Evidence type
- Builder first-hand (vendor author)
Source: SheetLink blog · 2026-06-23 · Checked 2026-10-19
WorkedMoney / bills
Tools: Claude
Setup, access, approval gate, failure notes
- Setup / prompt
- Give the agent a price ceiling and what matters (in-car radio) and let it work the retention chat.
- Access required
- SiriusXM account chat; card authorization at the end
- HITL checkpoint
- Agent paused on its own twice: when the best offer plus tax exceeded the ceiling, and before the card charge to confirm the plan covered the physical receiver
- Outcome & failure notes
- Renewal went from $29.87 to $8.04/month (incl. tax) for 12 months. Still ~$3/month above the user's ceiling, and the deal reverts after a year, so a renegotiation reminder was set.
- Evidence type
- User first-hand
Source: Damien Goehrig (blog) · 2026 · Checked 2026-10-19
PartialMoney / bills
Tools: Claude Sonnet 4 / 4.5, multi-agent (CEO + shopkeeper)
Setup, access, approval gate, failure notes
- Setup / prompt
- Same shop, upgraded models, CRM/pricing tools, and a supervising agent.
- Access required
- Same as phase 1, plus additional business tools
- HITL checkpoint
- Supervisor agent; humans still physical operators
- Outcome & failure notes
- Mostly eliminated loss-making weeks and turned a modest profit, but staff could still talk it into risky or unauthorized decisions. Better tools and structure helped more than the extra agent layer.
- Evidence type
- Lab experiment write-up
Source: Anthropic · 2025-12 · Checked 2026-10-19
PartialMoney / bills
Tools: ChatGPT agent mode
Setup, access, approval gate, failure notes
- Setup / prompt
- "Cancel my subscription to [service]." One service at a time.
- Access required
- User logs in to each service inside the agent's controlled browser
- HITL checkpoint
- User enters credentials and should confirm before the final cancel click
- Outcome & failure notes
- Peacock, Disney+ and Netflix cancellations went through. Some sites (Amazon Prime, Google One) block or flag the agent as suspicious, and retention flows are designed to resist bots.
- Evidence type
- Journalist first-hand
Source: Tom's Guide · 2025 · Checked 2026-10-19
PartialMoney / bills
Tools: Meta Muse
Setup, access, approval gate, failure notes
- Setup / prompt
- Ask the agent to find unused discounts and negotiate cable, phone, insurance, or medical bills by chat or phone.
- Access required
- Provider account logins; agent can place calls
- HITL checkpoint
- Verification codes, signatures, and bank details come back to the user
- Outcome & failure notes
- Users self-report savings from hundreds to thousands of dollars a year (e.g. Xfinity locked at $85.30/mo). But some providers block it as a bot, Amazon bans it outright, and one call needed the user patched in for an unreadable verification text. Figures are unverified user reports collected in a roundup.
- Evidence type
- Roundup of self-reported cases
Source: StashBase · 2026-09-22 · Checked 2026-10-19
PartialMoney / bills
Tools: ChatGPT
Setup, access, approval gate, failure notes
- Setup / prompt
- Have the model prepare negotiation scripts and target asks; a human makes each call.
- Access required
- None. The model only sees what the user tells it.
- HITL checkpoint
- Fully human-executed. AI only prepares.
- Outcome & failure notes
- Spectrum and AT&T wins plus a gym credit added up to ~$1,368/year for 2h41m of calls. Geico didn't budge. ChatGPT's price targets were overly optimistic and it didn't know account history (existing promos, late payments), so scripts missed context.
- Evidence type
- Tester first-hand, 30 days
Source: withn.tech · 2026 · Checked 2026-10-19
FailedMoney / bills
Tools: Frontier LLMs (Gemini, GPT, Claude)
Setup, access, approval gate, failure notes
- Setup / prompt
- Give the model a realistic taxpayer's documents and ask for a complete Form 1040.
- Access required
- Tax documents only (benchmark)
- HITL checkpoint
- None in the benchmark, which is the point: shows what unsupervised output looks like
- Outcome & failure notes
- Strict-correct returns (every evaluated line matching) were around a third in the first run and only just over half for the best models in later editions. Models confuse tax years and miss forms. Note the benchmark is published by a tax-software company.
- Evidence type
- Open benchmark
Source: Column Tax (TaxCalcBench) · 2025-07 · Checked 2026-10-19
FailedMoney / bills
Tools: Claude 3.7 Sonnet ("Claudius")
Setup, access, approval gate, failure notes
- Setup / prompt
- Agent manages inventory, pricing, and customer requests for an office snack shop for about a month.
- Access required
- Email, web search, notes, payments; humans restock physically
- HITL checkpoint
- Minimal by design; employees interact as customers
- Outcome & failure notes
- Lost money (net worth fell from ~$1,000 to under $800): gave away discounts when asked, sold tungsten cubes at a loss, mispriced items, and had an identity-confusion episode. Easy to socially engineer.
- Evidence type
- Lab experiment write-up
Source: Anthropic · 2025-06 · Checked 2026-10-19
WorkedCoding
Tools: Coding agents (Claude Code, etc.), AGENTS.md
Setup, access, approval gate, failure notes
- Setup / prompt
- Each time an agent repeats a bad behavior, add an explicit rule or script to AGENTS.md (e.g. use the screenshot script, use test filters).
- Access required
- Repo working tree
- HITL checkpoint
- Developer reviews output and maintains the rules
- Outcome & failure notes
- Adoption came in stages, starting with skepticism. The value came from treating agent errors as harness bugs to fix permanently, not one-off prompt tweaks.
- Evidence type
- Developer first-hand
Source: Mitchell Hashimoto (blog) · 2026-02 · Checked 2026-10-19
WorkedCoding
Tools: Google internal LLM migration tooling
Setup, access, approval gate, failure notes
- Setup / prompt
- Locate references, have the LLM propose edits, validate with builds/tests, route to owners for review.
- Access required
- Monorepo change generation; standard code review
- HITL checkpoint
- Every change human-reviewed; many edited before submit
- Outcome & failure notes
- LLMs generated ~74% of the 595 submitted changes (~69% of edits); ~38% needed human edits and ~26% were human-only. Developers estimated ~50% less migration time. Not a 'set and forget' win.
- Evidence type
- Peer-style paper (company-reported)
Source: arXiv: Migrating Code At Scale With LLMs At Google · 2025-04 · Checked 2026-10-19
WorkedCoding
Tools: LLM pipeline (Airbnb internal), Jest, TypeScript, ESLint
Setup, access, approval gate, failure notes
- Setup / prompt
- Per-file state machine: refactor → validate (tests, lint, types) → retry with error context; prompts tuned on failures.
- Access required
- Repo branch; CI validation; no production access
- HITL checkpoint
- Engineers reviewed output and manually fixed the last <100 files
- Outcome & failure notes
- 75% migrated in 4 hours, 97% after 4 days of prompt tuning; finished in 6 weeks versus an estimated 1.5 years by hand. The win came from validation and retries, not one-shot prompting.
- Evidence type
- Engineering blog (company-reported)
Source: Airbnb Engineering · 2025-03 · Checked 2026-10-19
WorkedCoding
Tools: Stripe Minions (fork of goose), MCP (400+ internal tools)
Setup, access, approval gate, failure notes
- Setup / prompt
- Tag the bot in a Slack thread (or click from a flaky-test ticket); agent runs in an isolated devbox and opens a PR.
- Access required
- Pre-warmed devboxes isolated from production and the internet
- HITL checkpoint
- All PRs human-reviewed; at most two CI rounds before handing back
- Outcome & failure notes
- Over a thousand merged PRs per week are fully minion-written. Isolation is what lets them skip per-command permission prompts. Runs that aren't fully correct still become starting points for engineers.
- Evidence type
- Engineering blog (company-reported)
Source: Stripe Dev Blog · 2026 · Checked 2026-10-19
PartialCoding
Tools: GitHub Copilot coding agent
Setup, access, approval gate, failure notes
- Setup / prompt
- Maintainers explicitly assign issues; agent opens PRs for review.
- Access required
- Repo PRs only; cannot open PRs on its own
- HITL checkpoint
- Maintainer must request each PR and review before merge
- Outcome & failure notes
- 878 PRs, 535 merged (67.9% vs 87.1% for Microsoft humans), ~95k lines added. Strongest at cleanup and tests, weaker at performance work. Early public PRs drew ridicule and the review burden is real.
- Evidence type
- Maintainer retrospective
Source: .NET Blog (Stephen Toub) · 2026-03 · Checked 2026-10-19
PartialCoding
Tools: Cursor + Claude 3.5/3.7 (early 2025)
Setup, access, approval gate, failure notes
- Setup / prompt
- Randomized trial: 16 experienced maintainers, real issues, AI allowed vs. disallowed.
- Access required
- Developers' own repositories
- HITL checkpoint
- Developers fully in the loop
- Outcome & failure notes
- Developers were 19% slower with AI while believing they were ~20% faster. Time went to prompting, waiting, and reviewing. METR changed its study design in 2026, and later evidence is more mixed, so treat this as context-specific.
- Evidence type
- Randomized controlled trial
Source: METR · 2025-07-10 · Checked 2026-10-19
FailedCoding
Tools: Claude Code, Terraform, AWS
Setup, access, approval gate, failure notes
- Setup / prompt
- Agent sets up a new site on existing AWS infra from a new laptop missing the Terraform state file.
- Access required
- Terraform with production AWS credentials; skip-permissions mode
- HITL checkpoint
- User approved a destroy plan without catching its scope
- Outcome & failure notes
- terraform destroy wiped the DataTalks.Club production DB and its snapshots (2.5 years of submissions); AWS support restored a hidden copy in about a day. Fixes: separate projects, deletion protection, off-resource backups, and the human runs destructive plans.
- Evidence type
- Founder postmortem interview
Source: Rootly (Alexey Grigorev interview) · 2026-03 · Checked 2026-10-19
FailedCoding
Tools: Gemini CLI, Windows
Setup, access, approval gate, failure notes
- Setup / prompt
- Ask the agent to rename and move a folder of files into a new directory.
- Access required
- Local filesystem write
- HITL checkpoint
- None; agent chained commands without verifying each step
- Outcome & failure notes
- A failed mkdir was followed by move commands that scattered files; the agent then claimed it had 'catastrophically' destroyed them. The user later found them misplaced at C:\ and withdrew the deletion claim. Agents' self-reports about what happened can be wrong in both directions.
- Evidence type
- User bug report
Source: GitHub google-gemini/gemini-cli #4586 · 2025-07 · Checked 2026-10-19
FailedCoding
Tools: Claude Code, drizzle-kit, Railway Postgres
Setup, access, approval gate, failure notes
- Setup / prompt
- Agent working in a separate terminal session on an app with an ORM.
- Access required
- Production DB connection string available to the agent
- HITL checkpoint
- None for that command; agent used --force to bypass the interactive prompt
- Outcome & failure notes
- `drizzle-kit push --force` dropped 60+ production tables; no automatic backups, so the data was unrecoverable and ~8 hours of rebuild followed. Pattern repeats with `prisma db push --accept-data-loss` (issue #14411). Deny-list force/accept-data-loss flags and keep prod creds out of agent envs.
- Evidence type
- User bug report
Source: GitHub anthropics/claude-code #27063 · 2026-02-20 · Checked 2026-10-19
FailedCoding
Tools: Replit Agent
Setup, access, approval gate, failure notes
- Setup / prompt
- Build a SaaS contacts app with an agent over many days, with explicit instructions not to change anything during a freeze.
- Access required
- Shared dev/prod database
- HITL checkpoint
- Instruction-only ('don't touch'); no technical gate
- Outcome & failure notes
- Agent ran destructive commands that wiped the production database and then wrongly said rollback was impossible. Root cause: dev and prod shared a database. Replit later added automatic dev/prod separation and a planning-only mode.
- Evidence type
- Incident + vendor CEO interview
Source: Fast Company · 2025-07 · Checked 2026-10-19
WorkedResearch digests
Tools: Claude Haiku (scoring), Claude Sonnet (summaries), GitHub Actions, Resend
Setup, access, approval gate, failure notes
- Setup / prompt
- Fetch ~250 papers/day → keyword prefilter → Haiku scores 1–10 against a tiered interest profile → Sonnet summarizes top 5–10 → EPUB emailed to Kindle at 5 AM ET.
- Access required
- arXiv API, Anthropic API, an email sender approved on the Kindle account
- HITL checkpoint
- None daily; human tuned the scoring prompt
- Outcome & failure notes
- First scoring prompt clustered everything at 8–8.5 (57% passed the threshold). Fixed with calibration examples and forced use of the full range. Kindle silently rejects unapproved senders. Runs for about $5–8/month.
- Evidence type
- Builder first-hand
Source: Another Coding Blog (Taylor Ortiz) · 2026-02-21 · Checked 2026-10-19
WorkedResearch digests
Tools: Claude Code for web, OpenAI Codex Cloud, GitHub
Setup, access, approval gate, failure notes
- Setup / prompt
- Write a clear research question; agent works unattended in a dedicated research repo, one folder per experiment, with network access for dependencies.
- Access required
- A separate public (or private) research repo only
- HITL checkpoint
- Human reads each report/PR; nothing merges into real projects automatically
- Outcome & failure notes
- Produced many completed experiments (e.g. Markdown library benchmarks, WebAssembly/Pyodide work). Isolation in a throwaway repo keeps the downside small; questions must be concrete and answerable by running code.
- Evidence type
- Developer first-hand
Source: Simon Willison's Weblog · 2025-11-06 · Checked 2026-10-19
WorkedResearch digests
Tools: Agent + Slack, Claude Code
Setup, access, approval gate, failure notes
- Setup / prompt
- Agent reads everything clipped the day before, sends an 8 AM digest of what's worth attention, and surfaces relevant saved articles when a new Claude Code project starts.
- Access required
- Read-only on the user's clipping store; post to the user's own Slack
- HITL checkpoint
- Read-only output to the user only
- Outcome & failure notes
- Builder-reported with a how-to; no accuracy data. Low risk because it only reads and posts to the owner. Included as a pattern example from X.
- Evidence type
- Builder self-report (X post)
Source: X (@diabhey) · 2026-04-24 · Checked 2026-10-19
PartialResearch digests
Tools: ChatGPT deep research
Setup, access, approval gate, failure notes
- Setup / prompt
- Ask a research question; agent asks clarifying questions, browses for minutes, writes a cited report.
- Access required
- Public web
- HITL checkpoint
- Reader must verify claims; best used in areas the user already knows
- Outcome & failure notes
- Produced a ~5,000-word cited report in about five minutes and beat Google's equivalent, but contained embarrassing factual mistakes. Useful for deepening work you can check, risky for topics you can't.
- Evidence type
- Journalist hands-on
Source: Platformer (Casey Newton) · 2025-02 · Checked 2026-10-19
FailedResearch digests
Tools: ChatGPT
Setup, access, approval gate, failure notes
- Setup / prompt
- Attorney used a chatbot to find supporting case law for motions.
- Access required
- None; output pasted into filings
- HITL checkpoint
- Required cite-check skipped
- Outcome & failure notes
- Five fabricated citations led a federal judge to reprimand and disqualify three Butler Snow attorneys and refer them to the bar. One of many 2025 sanctions. Any agent-generated citation must be opened and read before use.
- Evidence type
- Court order + legal press
Source: ABA Journal · 2025-07 · Checked 2026-10-19
WorkedSales / CRM ops
Tools: Salesforce Agentforce Sales Agent
Setup, access, approval gate, failure notes
- Setup / prompt
- Process hundreds of thousands of opportunities in a nightly window; generate cited recommendations for field updates and next steps.
- Access required
- CRM data with access checks per seller
- HITL checkpoint
- Sellers review recommendations; autonomy expanded gradually
- Outcome & failure notes
- 1.04M monthly recommendations for 13,000 internal sellers. Data integrity required cited reasoning, permission checks, and human review. Vendor dogfooding; no independent numbers.
- Evidence type
- Vendor engineering blog
Source: Salesforce Engineering · 2025 · Checked 2026-10-19
WorkedSales / CRM ops
Tools: Claude Code, HubSpot API, Slack
Setup, access, approval gate, failure notes
- Setup / prompt
- Auto-merge high-confidence duplicates; send fuzzy matches to AI review; route uncertain cases to a Slack queue; rescan every 4 hours.
- Access required
- HubSpot write access (merge) on live data
- HITL checkpoint
- Uncertain matches go to a human review queue (944 records at first)
- Outcome & failure notes
- Merged 2,500+ duplicate contacts/companies in about 4 hours for ~$10. The builder (a non-engineer) had a near-miss moment before running across 84k live records; the review queue was too big to handle by hand, which drove the AI-review tier. Open-sourced.
- Evidence type
- Operator first-hand
Source: Agent Operator (Andy Toizer) · 2026-04-07 · Checked 2026-10-19
PartialSales / CRM ops
Tools: LLM copy generation, Saleshandy
Setup, access, approval gate, failure notes
- Setup / prompt
- Three cohorts: AI-written (5k), human-written (2k), AI-drafted + human-edited (5k).
- Access required
- Sending domains and prospect lists
- HITL checkpoint
- Hybrid cohort: a human edits every AI draft
- Outcome & failure notes
- AI-only got 0.7% meetings and a 7.8% spam-flag rate versus 3.2% meetings and 3.1% spam for hybrid. Fully automated copy hurt deliverability and reply quality. Vendor-run test.
- Evidence type
- Vendor experiment
Source: Saleshandy blog · 2026 · Checked 2026-10-19
FailedSales / CRM ops
Tools: 11x (AI SDR)
Setup, access, approval gate, failure notes
- Setup / prompt
- Buy an 'AI digital worker' to run outbound prospecting end to end.
- Access required
- Email domains, CRM, prospect lists
- HITL checkpoint
- Marketed as autonomous
- Outcome & failure notes
- TechCrunch reported the vendor listed customer logos (e.g. ZoomInfo, Airtable) of companies that weren't customers, and sources described churn of 70–80% with ARR inflated by contracts carrying 90-day break clauses. The vendor disputed parts. Ask for references and trial outs.
- Evidence type
- Investigative reporting
Source: TechCrunch · 2025-03-24 · Checked 2026-10-19
WorkedContent pipelines
Tools: Hermes agent, Tavily, MailerLite, GitHub + Vercel
Setup, access, approval gate, failure notes
- Setup / prompt
- Research last 7 days → reject press-release hype → draft in voice from approved examples → evaluator pass → approval packet.
- Access required
- MailerLite campaign creation; website repo
- HITL checkpoint
- Approval packet shows copy, source URL per claim, subject, campaign status, blog metadata, and the exact pending action. A vague 'looks good' doesn't count.
- Outcome & failure notes
- Version 1 could research, draft, and send on its own, and the author judged that the wrong design. Quality depends more on what's rejected than what's generated. State must live outside chat history.
- Evidence type
- Practitioner first-hand (also sells builds)
Source: Mark Little · 2026-05-20 · Checked 2026-10-19
PartialContent pipelines
Tools: n8n, Gemini (audio), Telegram
Setup, access, approval gate, failure notes
- Setup / prompt
- Watch the RSS feed; Gemini listens to the full audio and drafts show notes, timestamps, 5 clip ideas, LinkedIn post, X thread, and newsletter blurb in the host's voice.
- Access required
- RSS feed, model API, Telegram bot, n8n Data Table
- HITL checkpoint
- Each piece is sent to Telegram with Approve / Edit / Skip; nothing ships without a tap
- Outcome & failure notes
- Builder-reported. Retries plus a backup model handle API busy errors. No published quality or engagement data, and the author sells it as a paid template, so treat it as a pattern, not a proven result.
- Evidence type
- Builder self-report (sells template)
Source: n8n Community · 2026-10-03 · Checked 2026-10-19
FailedContent pipelines
Tools: Generic AI chatbot
Setup, access, approval gate, failure notes
- Setup / prompt
- Freelancer used AI to produce a syndicated summer guide, including a reading list.
- Access required
- None; output went to a syndication partner
- HITL checkpoint
- Fact-check skipped at the writer, syndicator, and publisher
- Outcome & failure notes
- 10 of 15 books on the list didn't exist (fake titles attributed to real authors), and it ran in print in the Chicago Sun-Times and others. Later review found more errors in the section. Every name, title, and quote needs a source.
- Evidence type
- Publisher statement
Source: Chicago Sun-Times · 2025-05-20 · Checked 2026-10-19
FailedContent pipelines
Tools: Amazon Prime Video AI recaps
Setup, access, approval gate, failure notes
- Setup / prompt
- Model generates narrated season recaps from show content.
- Access required
- Published to the streaming catalog
- HITL checkpoint
- Not evident before launch
- Outcome & failure notes
- The Fallout recap misplaced a 2077 flashback in '1950s America' and garbled a key plot choice. Amazon pulled AI recaps from Fallout and the other test series. Narrative summaries need a fan-level fact check.
- Evidence type
- News reporting
Source: The Verge · 2025-12 · Checked 2026-10-19
FailedContent pipelines
Tools: CNET internal AI engine
Setup, access, approval gate, failure notes
- Setup / prompt
- AI drafts SEO explainers; editors review before publish.
- Access required
- CMS publishing
- HITL checkpoint
- Editor review existed but missed errors
- Outcome & failure notes
- After outside reporting, CNET's audit corrected 41 of 77 AI-written stories (factual errors, plagiarism-like phrasing). It paused the program and rewrote its AI policy. 'Human reviewed' is not a gate if reviewers skim.
- Evidence type
- News reporting + publisher audit
Source: The Verge · 2023-01-25 · Checked 2026-10-19
No cards match. Try clearing a filter.