Workflow library · updated 2026-10-23
AI agent systems and automations SaaS teams actually run
13 end-to-end agent systems and 44 building-block workflows that real companies and builders run today, each linked to its public source: Vercel, Zapier, Anthropic, Klarna, Intercom, Clay customers, Linear, Ramp, Stripe, Pylon customers and the n8n community. You get the architecture, tools, human checkpoints and pitfalls. Numbers are quoted from sources; anything else is labelled as our estimate.
7 frontier · 17 advanced · 20 starter · 60 cited sources · See the frontier builds →
Frontier systems: end-to-end AI agents
13 multi-stage systems that chain agents, automations and human checkpoints, each with a learning loop that improves it from results. Each one combines patterns from public builds and links to its sources and smaller building blocks.
Starts from your best customers and ends with booked meetings.
Every inbound lead is researched, qualified into a typed label, and gets a drafted reply approved in Slack.
Every sales call is prepped by an agent, turned into structured CRM fields afterwards, and added to a won/lost corpus that answers messaging and competitor questions in Slack.
Mines your CRM and public data for accounts you can win back or take from a competitor, researches each one, and writes outreach that references the real history.
A support agent that answers from your docs, follows policy, performs real actions like refunds and plan changes, collects pre-work for engineers, and hands off cleanly.
Usage, billing and support signals become account scores in the CRM.
Performance data in; hundreds of on-brand, limit-checked variants out, rendered into creative by a Figma plugin.
Finds real questions, drafts with focused editor agents, publishes after a human edit, then watches rankings and rewrites pages that stall.
Generates useful pages from structured data you own (integrations, templates, use cases), measures which earn traffic, and improves or prunes the template.
Watches competitor pricing, pages, ads and community chatter, diffs the changes, explains them, and sends a weekly brief to sales and product.
Sources candidates, screens out fraud, gives every applicant a short AI first-round interview, and hands the team a ranked shortlist with transcripts.
A Slack agent that answers from docs and past threads with sources, answers data questions with SQL, files tickets from conversations, and reports gaps in your docs.
Codes expenses, chases receipts, three-way-matches invoices, recovers failed payments and sends an AR digest.
Building blocks by category
Browse all building blocks
44 workflows
Every inbound form fill gets a deep-research pass on the person and company, a structured qualification verdict, and a drafted reply.
Who can build this →When a lead signs up, a voice agent calls them within a couple of minutes, asks one or two routing questions, and either texts a booking link or hands over to a rep.
Who can build this →Ops builds enrichment and contact-finding workflows once (waterfall across many data vendors); reps call them from the AI chat they already use to get account context, buying signals and verified contacts..
Who can build this →Pull last year's closed-lost deals and stalled meetings, have an LLM summarise why each prospect originally engaged from the pre-call notes, then send a short personal email that references the AE, the last meeting and that reason..
Who can build this →For every target account, an AI web-research step checks which of your service categories they use, why, and which competitor provides it, then writes the answers into CRM fields reps can filter on..
Who can build this →After each recorded call, the transcript is filtered by deal stage, summarised, and specific fields (pain, competitors, next steps, objections) are written to the opportunity.
Who can build this →Before each customer or prospect call, AI reads the HubSpot record, researches the company, lists the automation challenges they are likely facing, and posts a short brief to the rep in Slack..
Who can build this →When someone from an ideal-customer account engages with an executive's LinkedIn post, a webhook sends the lead into a workflow that enriches, filters on qualification criteria, and stores qualified leads for sales follow-up..
Who can build this →A chat bot grounded in transcripts from deals you won and lost upmarket.
Who can build this →New HubSpot contacts or Slack-posted domains are enriched with firmographics, scored against your ICP and replied to in-thread with a summary, so the right rep picks them up..
Who can build this →An agent reads a CSV of existing ads with performance, finds weak ones, and generates replacements that respect character limits (30 for headlines, 90 for descriptions) using two sub-agents.
Who can build this →Generate one useful page per integration, use case or app pair from structured data you already have (supported triggers, actions, templates), with AI only filling descriptive gaps..
Who can build this →A writer submits a feature name and help-doc links.
Who can build this →Merged pull requests are collected, grouped, and turned into a plain-language changelog draft for customers.
Who can build this →Scheduled scrapes of competitor pricing pages, sites, ads and news are diffed against the last run; an LLM explains what changed and why it might matter, and a weekly digest goes to Slack..
Who can build this →Monitor relevant subreddits for questions and sentiment about your category; cluster recurring questions and turn them into answer posts or FAQ entries, and alert the team to threads worth a genuine reply..
Who can build this →Load an open-source library of marketing "skills" (copywriting, SEO audits, CRO, pricing pages, email) into a coding agent such as Claude Code, Codex or Cursor, and run them against your own site and data..
Who can build this →English subtitle (SRT) files are added to a table; a workflow asks an LLM to translate them into ten languages while keeping timestamps intact, then stores the files for upload..
Who can build this →A support agent answers from your knowledge base in your tone, follows your policies, and takes actions (refunds, account changes) through APIs, handing over to people when unsure..
Who can build this →Before an engineer sees a ticket, an AI agent answers L1 questions from docs, runs structured runbooks for known L2 processes (e.g.
Who can build this →New tickets are classified, prioritised and labelled; similar resolved tickets are found and a suggested resolution is attached for the agent..
Who can build this →When a support person reacts to an internal question with a set emoji, AI generates a search term, fetches matching help docs, reviews related ticket history, and replies in the thread with linked steps..
Who can build this →Every new issue is compared against your history to suggest team, assignee, labels, project and likely duplicates, with the reasoning shown.
Who can build this →Daily, combine billing (failed payments, downgrades), support (ticket volume and tone) and CRM data into a health score; an LLM explains the main reasons and suggests a save play; alerts go to Slack and the owner in HubSpot..
Who can build this →Sync product usage from your warehouse to the CRM (seats, active users, key features, collaborators) and score accounts for expansion; reps and CSMs get prioritised lists and context..
Who can build this →A support specialist describes an issue in a short form; AI writes the status page message using your templates and rules, posts it to Slack for approval, and a webhook publishes the approved text..
Who can build this →Every morning, AI summarises yesterday's escalations, resolved issues and recurring themes into one formatted Slack post; ticket categories and reply counts also go to a sheet for metrics..
Who can build this →When a deal closes, create onboarding tasks, start a timed email sequence and alert the team; progress is tracked on the HubSpot record..
Who can build this →A Slack bot answers team questions (lead stages, processes, policies) from internal documentation, with links, so people stop interrupting the one person who knows..
Who can build this →Every application is enriched with IP and phone data; stated location is compared to real location, internet-based phone numbers are checked, and the applicant is rated no / moderate / high fraud risk.
Who can build this →Each CV is parsed and summarised against the role's must-haves, with evidence quoted, then written to a sheet for the hiring manager to sort..
Who can build this →A Slack conversation is summarised into a properly written ticket and added to the next planned sprint..
Who can build this →After a meeting, notes are turned into a summary and action items with owners and due dates in Google Workspace, and owners are reminded..
Who can build this →Team members ask questions in plain English; an agent writes SQL against your warehouse, runs it, checks the result and explains it, in several reasoning phases..
Who can build this →For platforms that host user content, reported URLs are analysed visually and textually for phishing, spam or copyright issues; the agent proposes an action and a security engineer makes the final call..
Who can build this →Query KPIs from BigQuery, have an LLM write commentary on what changed, build a Google Doc / PDF and send it by email and Teams..
Who can build this →Turn on ML-timed retries in your billing system, then add your own layer: personal emails for high-value accounts, a CSM alert, and a pause-not-cancel rule..
Who can build this →An agent codes each card transaction and invoice line to your GL, checks it against policy, approves low-risk items and escalates anything ambiguous, and answers employees' "can I expense this?" by text..
Who can build this →Snap a receipt in Telegram, or forward a card alert email; AI extracts vendor, amount and category and adds a row to the expense sheet..
Who can build this →Every week, pull open invoices, bucket them by age, render a PDF and email it to finance and account owners..
Who can build this →For dental and clinic software: a voice agent answers calls, checks calendar availability and books, reschedules or cancels appointments, then confirms by SMS..
Who can build this →For bookkeeping and accounting practice software: automatic client reminders for missing documents, filing of what arrives, and a status view per client..
Who can build this →Starter packs: where to begin
Not sure what to automate first? These sets are good first steps for common team shapes. Pick one, and start with the first item.
Founder-led sales, one support person, no ops hire yet.
You do everything. Most hours back for the least setup.
Building a repeatable pipeline with 2–5 reps.
Keep customers and answer faster.
Like Anthropic's growth team: one person, many channels.
The most ambitious builds we found running in production.
Is it worth automating? Quick estimate
A simple payback estimate. Use your own numbers; results are illustrative, not a promise.
Download the checklist
A one-page plain-text checklist to scope any automation before you talk to an agency: automation-scoping-checklist.txt.
Describe what you need in your own words. We turn it into a clear brief and match you with agencies from the directory that list this kind of work.
Get matched with 3 vetted agencies
Describe the project. We read every brief and introduce you to three agencies from this index that fit it, usually within 2 working days. Free for buyers; agencies can't pay to be picked.