Bilal M.
CRM & AI Automation Expert | n8n, Make, Zapier, Automated Workflows
CRM and AI automation. I build the CRM, then the automated workflows that run it. HubSpot, GoHighLevel, Zoho and Pipedrive wired into n8n, Make and Zapier, with AI agents that qualify inbound leads, answer customers and book appointments straight into the calendar. Most businesses don't lose leads because the leads are bad. They lose them because a form comes in at 9 PM and nobody replies until Thursday. A CRM on its own won't fix that. The CRM plus the workflow sitting behind it will, and building that part properly is the entire job. CRM SETUP, MIGRATION AND CLEANUP 1. Pipelines and deal stages mapped to how you actually sell, with the custom fields, tags and owner rules that keep the data usable six months in. 2. Migration and consolidation. Contacts, notes, tags and pipeline stages moved between HubSpot, GoHighLevel, spreadsheets or whatever you are on now, with the history of what happened intact. 3. Lead capture into the CRM from forms, landing pages, Jotform, paid ads and your website, so nothing gets typed in twice. 4. Deal routing and assignment. Round robin, territory or score based, running inside HubSpot off the enrichment pipeline rather than by hand each morning. 5. Lead scoring and enrichment. On one build, three enrichment sources run in parallel and merge by domain, and the score is calculated deterministically in code rather than left to an LLM. That is what lets sales ask why a lead was tiered the way it was and get a real answer. 6. Reporting a business owner can read. Source attribution, pipeline value by stage, and an honest read on where deals are actually dropping off. AUTOMATED WORKFLOWS IN N8N, MAKE AND ZAPIER 1. Builds with real error handling, conditional logic, webhooks and scheduled triggers. Separate error lanes so a retryable failure doesn't page anyone at 3am. 2. Speed to lead. A new lead gets enriched, scored, assigned and contacted by SMS or email in under a minute, and the follow up keeps running until someone replies. 3. Booking, reminders and no show recovery. One Twilio triggered agent handles appointment booking over SMS, MMS and WhatsApp against live availability held in Google Sheets. 4. Failed payment and churn recovery. A Stripe failure drives a state machine stored in Postgres so retries survive a restart. Silent retry first, then escalating contact, then a churn tag in HubSpot and a win back sequence. 5. Order and fulfilment flows. Shopify webhook, per line item inventory check, short lines fork to backorder with a customer notice instead of failing quietly. 6. Internal ops. CV screening and tiering into Notion, invoice and timesheet generation from emailed attachments using OCR, employee onboarding and offboarding access provisioning driven off one HRIS webhook with an append only audit log. 7. Fixing automations somebody else built. A Make scenario stopped pushing to Google Ads because a webhook wasn't receiving Airtable data, traced and fixed inside a week. A Jotform to Shopify flow was quietly creating broken pages from two unmapped fields, found in the execution history and fixed in two hours. AI AGENTS AND CHATBOTS 1. Agents on the OpenAI API with tool calling, RAG and memory, for the jobs a native CRM chatbot can't handle. 2. Retrieval over your own documents with LangChain and Qdrant, so answers come from your material instead of being invented. 3. Confidence scoring with human handoff. On one omnichannel support build, anything scoring under 0.80 goes to a person with a drafted reply attached rather than sending a guess to a customer. 4. Messaging agents across SMS, WhatsApp and web chat through Twilio, plus voice agents for outbound qualification calls. 5. Every agent writes back to the CRM. Qualification answers, call outcome and next step land on the contact record, not in a chat log nobody opens. INTEGRATIONS AND DATA 1. API and webhook integration across HubSpot, Airtable, Notion, Slack, Google Workspace, Stripe, Shopify, Odoo and Twilio. 2. Postgres, MongoDB or Supabase when the job needs a real database behind it instead of a spreadsheet pretending to be one. 3. ETL, scraping and two way sync, so Sheets and Airtable stay in step with the CRM instead of drifting apart. WHERE THIS PAYS OFF Real estate, home services like roofing, HVAC, electrical and landscaping, clinics and salons, coaching, ecommerce, SaaS, and agencies running CRM and automation for their own client base. If your business runs on inbound leads and booked appointments, this is the shape of work I do every week. HOW I WORK Every build ships with a Loom walkthrough and written documentation, so you know exactly what was made and can maintain it without me. No black boxes. Most projects land in 3 to 7 days. TO GET STARTED Message me with the process you want automated, the outcome you are after and your timeline. You will get a clear plan and an honest read on what is possible, usually within a few hours. If it is not a fit I will tell you that instead of taking the job.
Earnings and projects
Last 6 months ÷ 4.4 months active (since Jun 2026)
Last 6 months: $1k
$336 in the last 6 months
Earnings, last 30 days
Skill rankings: Top 31% in Chatbot Development
Ranked by MRR among everyone on Upwork with the skill.
- #124Agentic Workflows340 freelancers · 🇬🇧 #2Top 36%🇬🇧#2 of 6
- #273Pipedrive744 freelancers · 🇬🇧 #9Top 37%🇬🇧#9 of 21
- #323Lead Management Automation907 freelancers · 🇬🇧 #11Top 36%🇬🇧#11 of 22
- #863AI Automation2,024 freelancers · 🇬🇧 #17Top 43%🇬🇧#17 of 39
- #968Chatbot Development3,106 freelancers · 🇬🇧 #23Top 31%🇬🇧#23 of 76
- #981Zoho CRM2,793 freelancers · 🇬🇧 #13Top 35%🇬🇧#13 of 34
- #1,003CRM Development2,819 freelancers · 🇬🇧 #30Top 36%🇬🇧#30 of 69
- #1,830Automated Workflow5,123 freelancers · 🇬🇧 #42Top 36%🇬🇧#42 of 111