AI automation

Problem: Service leads go cold when nobody replies quickly, and leads who already answered keep getting the same generic follow-ups.

What I built: A GoHighLevel sales funnel captures service requests, creates or updates pipeline opportunities, tags leads, and alerts the team. Service-based routing triggers email and SMS follow-ups, while replies move leads forward, flag hot leads, and stop the earlier follow-up workflow.

Problem: Booked calls get forgotten, and cancellations or no-shows rarely get a follow-up, leaving empty slots and lost leads.

What I built: A GoHighLevel automation confirms new bookings, sends reminders ahead of the appointment, and follows up automatically when a contact cancels or misses the slot, so the calendar stays full and no lead is left unanswered.

Problem: Moving won deals into the client pipeline is manual, and finished clients are rarely asked for a review at the right time.

What I built: Marking a lead as Won triggers Moved to Won and Create New Client, moving it from the Lead Pipeline into the Client Pipeline as a new client. When the job is done, Project Completed and Ask for Review sends three review SMS texts, and a reply of “done” stops them.

Problem: Missed calls turn into lost enquiries, and sent contracts sit unsigned with nobody chasing them.

What I built: A missed call gets an automatic text back after a minute, so no enquiry is lost. Three days after a contract is sent, a text reminds the contact if it is still unsigned and confirms once it is signed.

Problem: Support teams answer the same WhatsApp questions over and over, and harder questions need a person without losing the conversation.

What I built: A WhatsApp bot answers customer questions from a Notion knowledge base using Supabase vector search and Gemini. Questions it can't answer become HubSpot tickets with a Slack alert; when an agent closes the ticket, the answer is sent back on WhatsApp and can be saved to the knowledge base, which re-indexes itself automatically.

Problem: Reading and scoring every CV by hand is slow and inconsistent, so strong candidates wait days for a screening call.

What I built: An end-to-end hiring workflow that extracts and scores CVs with Gemini, routes candidates, generates questionnaires, books screenings, and updates Airtable.

Problem: Sales time goes into chasing every WhatsApp enquiry, including leads that were never a good fit.

What I built: A WhatsApp bot collects five qualification answers, scores leads with AI, and updates HubSpot contacts and deals. Lead scores route follow-ups to Notion, Slack, or a nurture list, while a companion workflow alerts Slack when errors occur.

Problem: Finding relevant job posts and tailoring a resume for each one takes hours of copying and editing.

What I built: A Slack message launches job search validation, listing scraping, and AI resume tailoring. Google Drive files are copied and updated with job-specific content.

Problem: Page messages go unanswered outside business hours, and generic bots can't answer questions about the business itself.

What I built: An AI agent retrieves business knowledge from Google Docs, maintains conversation memory with Gemini, and delivers contextual responses through webhooks.

Problem: Booking, changing and cancelling appointments by phone ties up staff and leads to double bookings and time-zone mistakes.

What I built: A voice AI workflow finds available slots, books, updates, and cancels appointments with timezone conversion, calendar checks, error handling, and Airtable logging.

Problem: Attachments pile up in the inbox and have to be downloaded, renamed and filed into Drive by hand.

What I built: A Make scenario that watches Gmail for unread messages, retrieves attachments, renames files with a generated response, uploads them to Google Drive, logs results in Sheets, and confirms back to the sender.

Problem: Every new client needs the same setup in Airtable, Xero, Notion and Gmail, and doing it by hand causes delays and missed steps.

What I built: A Make scenario watches Airtable records and checks Xero for an existing contact. The new-contact path creates a contact and invoice, adds a Notion item, sends Gmail emails, and updates Airtable to keep onboarding records connected.

Problem: Pulling bank transactions from Xero for a given date and sharing them with the team is a repetitive manual export.

What I built: A Make scenario that triggers when an Asana task completes, pulls bank transactions from Xero for a chosen date, routes and aggregates the data, and saves everything to Asana and Google Sheets.

Problem: New leads arrive with little company information, so the team can't tell which ones to call first.

What I built: A webhook catches each lead, resolves the company domain, and pulls company details — then splits high- and low-priority leads into Sheets, Slack alerts, draft notifications, and client emails.

Problem: Each stage of a lead in Asana needs different emails, folders and follow-ups, and doing it by hand means some slip through.

What I built: A 25-step CRM flow that reacts when a task updates and routes leads by stage — onboarding folders, follow-up nudges, quote emails, AI-drafted approval messages, and paid-and-closed wrap-ups.