How the work actually runs.
Each playbook is an illustrative composite. The workflow is real. The stack is real. The numbers are real, with names and identifying details anonymised.
Turned handwritten and typed application forms into an approve, reject or review decision for a vocational training provider
Pattern: handwritten + typed document OCR + per-field confidence + validation rules + threshold decision + review queue
- ~25 min → ~3 min Time to screen one applicant pack
- ~94%, rest sent to review Fields read correctly without a person
Gave every client a private Notion portal with only their own tasks, without duplicating a single database, for a solo business consultant
Pattern: linked databases + person-based page permissions + guest access + client portal views + plain-English handover
- 0 → all 9 Clients who can see only their own tasks
- ~9 → 0 Weekly status emails written by hand
Moved engaged Brevo readers into HubSpot newsletters every week, with a dry run first, for a regional news publisher
Pattern: scheduled segment export + CRM lookup + subscription check + three-way sort + dry-run report + list cleanup
- 0 → ~300 a week Engaged readers moved into a HubSpot newsletter
- ~1 in 6 → none Readers subscribed twice or emailed from both tools
Stopped duplicate contacts and same-day lead misses for a five-agent insurance agency with three Zaps
Pattern: web form intake + HubSpot create-or-update + rule-based owner assignment + follow-up task + Sheets lead log
- ~40 → 0 Duplicate contacts created per month
- ~1 in 3 → under 1 in 20 Web leads with no call the same day
Turned a 400-lead Pipedrive account into a forecast the founders trust for a B2B corporate training company
Pattern: pipeline restructure + stage exit criteria + required fields per stage + lead cleanup + weighted forecast
- ~35% → 100% Open deals with an expected close date and value
- off by ~40% → within ~10% Quarterly forecast vs actual revenue
Cut unanswered calls at peak hours from three in ten to almost none for a two-location physiotherapy clinic
Pattern: AI phone receptionist + FAQ answers + live booking + red-flag handoff + call summaries
- ~30% → under 3% Calls unanswered at peak hours
- +18 New patient bookings per month
Cut a clinical literature search from three weeks of manual exports to two days for a medical device regulatory team
Pattern: multi-database search + deduplication + open-access PDF retrieval + reference files + screening spreadsheet
- ~3 weeks → 2 days Time to a screened-ready literature set
- ~25% → under 1% Duplicate records reaching reviewers
Cut invoicing from a Sunday night chore to one tap after each job for a two-van removals company
Pattern: mobile job tracker + customer records + GST calculation + one-tap PDF tax invoice
- ~5 days → same day Time from finished job to sent invoice
- ~3 h/week → under 30 min Owner admin time on invoicing
Cut tenant requests left without a reply after a day from one in five to almost none for a property management company
Pattern: Gmail + WhatsApp intake + AI classification + routed task cards + drafted replies + reminders + activity log
- ~1 in 5 → under 1 in 50 Requests with no reply after 24 hours
- ~6 hours → under 30 minutes Time to first reply
Cut missed lead follow-ups from one in four to almost none for a two-owner HVAC and plumbing company
Pattern: lead email intake + daily WhatsApp follow-up list + voice note updates + job ledger
- ~1 in 4 → under 1 in 20 Leads with no follow-up in 48 hours
- ~5 h/week → minutes Owner time updating the sheet
Moved a home-built Sheets, Apps Script and AppSheet system into a new Google Workspace without losing a single automation for a window cleaning company
Pattern: system audit + dependency map + Workspace migration + script and trigger repair + AppSheet reconnection + handover
- 9 found in audit → 0 at handover Automations broken after the move
- 60+ → 1 config tab Hard-coded file IDs in scripts
Replaced an eight-page energy report with one daily list for the operators of 40 commercial buildings
Pattern: building data roll-up + daily prioritised issue list + weekly trend view
- ~40 minutes → ~5 Time to find what needs attention
- ~1 in 3 → most Energy anomalies acted on within a day
Cut drawing intake from two hours to ten minutes per project for a 30-person engineering firm
Pattern: CAD PDF extraction + instruction parsing + structured output + review queue
- ~2 hours → ~10 minutes Intake time per project
- ~90% Fields extracted without a person
Cut the time from loan inquiry to underwriting file from three days to one hour for a commercial loan broker
Pattern: lead capture + underwriting platform sync + status write-back + borrower updates
- ~3 days → under 1 hour Inquiry to underwriting file
- ~40 fields → 0 Data re-entered by hand per deal
Doubled chat-to-cart conversion with a product assistant that never invents a detail, for a home-goods store
Pattern: catalog-grounded shopping chat + budget recommendations + product cards + handoff
- ~9% → ~19% Chat sessions that add to cart
- ~70% Pre-sale questions answered without staff
Recovered 11 jobs from old estimates in one quarter for a residential roofing company
Pattern: estimate follow-up + 30/60/90 reactivation + post-job review request
- 11 in one quarter Old estimates turned into jobs
- ~3 → ~18 Google reviews per month
Cut inbox sorting from 90 minutes a day to 15 for a B2B industrial supplies distributor
Pattern: inbox categorisation + pipeline routing + drafted replies
- ~90 minutes → ~15 Daily inbox sorting time
- ~50% → 90%+ Quote requests answered same day
Halved the time from inquiry to first booking for a pet-care company with 25 caregivers
Pattern: inquiry response + meet-and-greet booking + caregiver assignment + invoicing + onboarding
- ~6 days → ~3 Inquiry to first booking
- ~20% → under 2% Inquiries with no reply after 24 hours
Caught slipping project milestones two weeks earlier for a 40-person commercial fit-out contractor
Pattern: milestone tracking + weekly slippage briefing + per-job timeline
- ~2 days → ~2 weeks Warning before a missed milestone
- ~45 minutes → read in 3 Time to answer "where are we on every job?"
Scored 1,200 independent tour guides for outreach in two days for a tourism marketing agency
Pattern: online presence scrape + 13-field scoring + opening line + sheet output
- ~6 weeks by hand → 2 days Time to research 1,200 guides
- ~2% → ~9% Reply rate on outreach
Cut missed client requests to near zero for a 15-person agency by logging every channel in one place
Pattern: multi-channel request capture + single request log + overdue list + ask-in-chat
- ~40% → 95%+ Requests logged automatically
- 15–20 a month → 2–3 Requests past their promised date
Turned two years of recorded meetings into an askable knowledge base for a 25-person consultancy
Pattern: recording ingest + transcription + auto-filing + cited answers in Slack
- 0 → every one, within the hour Recordings transcribed and filed
- 20+ minutes → under 1 minute Time to find what a client said
Booked 14 qualified meetings a month for a founder-led B2B software company without hiring an SDR
Pattern: lead research + enrichment + personalised sequence + reply classification + calendar booking
- 3–4 → 14 Qualified meetings booked per month
- ~10 → ~2 Founder hours on outbound per week
Turned 300 university websites into one clean program database for a student recruitment agency
Pattern: site crawl + structured extraction + dedupe + fit scoring + sheet output
- ~3 weeks by hand → 2 days Time to build the program list
- ~60% → 95%+ Fields filled per program
Answered every after-hours WhatsApp and website enquiry within a minute for a three-location auto repair group
Pattern: WhatsApp + website chat auto-reply + FAQ answers + booking link + CRM logging
- 0% → 100%, under 60 seconds After-hours enquiries answered
- ~35% Enquiries booked without staff
Replaced five disconnected spreadsheets with one work-order system for a 60-person contract manufacturer
Pattern: CRM-to-work-order handoff + shop-floor status tracking + daily exceptions report
- 2 days → same day Order-to-work-order handoff
- about 1 in 4 → 9 in 10 Late orders caught before the due date
Cut missed-lead loss by 80% for a 12-person landscaping company in 6 weeks
Pattern: missed-lead rescue + SMS qualifying + CRM logging
- 60 seconds (was 0%) After-hours response time
- 23 Booked appointments in 6 weeks
Reduced no-shows by 34% and recovered $48K in annual revenue for a two-location dental practice
Pattern: recall sequence + no-show recovery + PMS write-back
- 52% → 71% Recall capture rate
- 9.4% → 6.2% No-show rate