Mark Anthony Reyes

Accountants manually copy-pasted Xero transactions into Asana tasks and Sheets every week — hours lost, mismatched records, and missed reconciliations.

A self-healing Make pipeline pulls completed Asana tasks, hits the Xero API, iterates and aggregates the data into Sheets, then re-attaches the result to the right Asana card. Zero copy-paste.

Inboxes drowned in unnamed PDFs and screenshots. Staff dragged files into Drive folders by hand, often filing them wrong or losing them entirely.

Every attachment is sent to a vision model, intelligently renamed, and routed to the correct Drive folder automatically — turning chaotic inboxes into a searchable document library.

Calls went to voicemail after hours. Missed bookings, no-shows, and a receptionist tied up rescheduling appointments all day.

A voice agent answers 24/7 — checks availability, books, reschedules, and cancels end-to-end. Every confirmed booking syncs to calendar and CRM in real time.

Days spent scrolling job boards, tweaking the same resume, and writing each cold email by hand — burnout with low response rates.

A Slack command scrapes fresh roles, tailors the resume per posting via an AI agent with structured output, and drafts personalized application emails in minutes.

HR policies scattered across outdated PDFs, spreadsheets, and shared drives. Employees asked the same questions repeatedly, and finding the right policy meant digging through folders or waiting for HR replies. Every handbook revision required manual re-uploading with zero searchability.

An n8n pipeline monitors Google Drive for handbook uploads or edits, routes PDFs and spreadsheets through text extractors and JavaScript normalizers, then pushes structured chunks into a Supabase vector store via Google Vertex embeddings. Employees now query a live knowledge base and get instant, cited answers from the latest policies.

New property inquiries piled up in Google Forms with duplicates, inconsistent fields, and no prioritization. Agents manually copied leads into GoHighLevel, guessed which were hot, and lost deals while cold leads got the same generic email as serious buyers.

An n8n pipeline captures every Google Form lead, dedupes and standardizes the data, then runs it through Gemini to score Hot / Warm / Cold. Contacts and opportunities are auto-created or updated in GHL, lead-temperature-specific follow-up emails and Slack alerts fire instantly, and AI notes are attached to every opportunity — agents only touch leads that are ready to close.

Leads arrived through two separate channels — a Tally web form and a chatbot conversation — and each was handled differently. Reps eyeballed submissions, scored them inconsistently, and manually decided who to call vs. who to nurture. Hot leads slipped through the cracks, follow-ups were delayed hours or days, and no two leads got the same treatment.

Both Tally and the chatbot funnel into a single Shared Sub-Zap engine where AI scores every lead the exact same way. Hot leads instantly create/update a HubSpot contact, ping Slack, and generate a follow-up task due in ~1 hour. Nurture leads land in HubSpot with an automated outbound email. One brain, two entry points, zero inconsistency.

Inbound leads sat in a spreadsheet for hours. Hot prospects went cold while reps manually researched and prioritized them one by one.

Leads are enriched with Apollo, scored, and split into priority paths — hot leads get instant Slack pings and AI-drafted emails; the rest are nurtured automatically.

Sales reps chased every CRM stage manually — follow-ups forgotten, quotes delayed, welcome emails inconsistent across the funnel.

A 24-step engine reacts to every stage — Ready, No Response, Quoted, Approved, Paid — firing the right follow-up, quote, welcome, or recommendation. Reps close, the system chases.