Freelancer - n8n Automation Expert – Gmail / Google Sheets / Slack / AI Email Classification

[soufiane_business]

Hi BenKo,

This is a strong fit because the requirements are already clear. I would split the six workflows into a clean first milestone instead of trying to ship everything in one big risky pass.

Best paid first milestone:

• Gmail trigger and message normalization
• AI classification for the 5 categories
• Sheets status row with unique reference number
• Slack urgent alert path
• loop protection for auto-replies
• error log and retry notes

After that, hotel forwarding, daily reporting, and booking-engine API steps can be added with less risk.

I can return the n8n export, the prompt/classification rules, test cases with sample emails, and handoff notes. A fixed price per workflow also makes sense here because each flow has a clear output.

[Serhii_Moshkivskyi]

Hi BenKo,

I’m interested in this project.

My closest match is the Gmail / Google Sheets / AI email workflow part. I have built an n8n proof-of-concept workflow with webhook input, Google Sheets status tracking, waiting step, Shopify order check logic, AI-generated email, Gmail sending, and status update.

For your project, I would suggest starting with one paid test workflow first, for example:

1. Automatic Email Labeling with Gmail + LLM
   or

2. Request Status Tracking with Google Sheets and unique reference numbers

If the first workflow is accepted and documented well, we can continue with the next workflows step by step.

I can help with n8n workflow setup, Gmail / Google Sheets workflow logic, LLM-based email classification, Slack notification flow, testing, and documentation.

I work best with clear written requirements and async communication. English communication is fine for me in writing.

I can share a screenshot of my n8n proof-of-concept workflow if helpful.

Could you share the requirements document or the first workflow specification, so I can estimate a fixed price for the first workflow?

[skoskr]

Hi BenKo,

This brief reads almost exactly like the production work I run on n8n daily — Gmail API, Google Sheets tracking, Slack briefings, LLM classification via Claude/OpenAI. Direct mapping:

• Email labeling with LLM: Email Re-engagement System PRO uses Groq + Claude fallback for content classification + dynamic generation in production (Gumroad).

• WhatsApp AI Invoice Agent: Gmail-adjacent — Twilio + Claude + Sheets + PDF.co + Drive, same orchestration shape.

• Status tracking + timestamps in Sheets: pattern from my Invoice Automation System (PDF gen + Drive + payment reminders, status stage progression).

• Slack daily digest: Marketing Automation Suite ships Google Ads anomaly briefings as a production feature.

Marketing Automation Suite - Setup Service

6-Month Email Re-engagement System PRO (AI-Powered)

[hellopaul]

Hey ,

This project aligns well with my background in AI automation, n8n workflows, API integrations, and scalable orchestration systems.

I run Blukaze Automations where we build enterprise-grade AI automation and workflow solutions.

Would love to discuss the requirements further — feel free to DM me or reach out at hello@blukaze.com

[Ghulam-e_Mustafa_Pak]

Interested

[mefadate]

If you want folks to throw their hat in, you might get more bites by sharing a tiny peek at the workflow pain points you’re hoping to fix. Even one rough example can help people see if their brain fits the puzzle. I’ve tinkered with n8n a lot for creative teams, and little context goes a long way in shaping better solutions and stronger challenge entries.

[maxim_makselyanov]

Hi BenKo. I can help build this as documented.

I understand the scope as 6 connected workflows: Gmail classification, hotel update forwarding, urgent Slack alerts, request/status tracking in Sheets, daily Slack reporting, and first-contact auto-reply with loop protection.

I would start with one paid workflow first: email-labeling schema, LLM guardrails for DE/EN/FR, Gmail labels, test cases with redacted samples, error handling, and a short handoff note. Rough fixed pricing after seeing the requirements doc: USD 180-250 for the first labeling workflow, then a per-workflow or package quote once the exact fields and API access are confirmed.

Relevant proof links:

CV: Please read

Sample deliverables: Please read

One question: do you want the fixed quote split per workflow, or one total package with the labeling workflow as milestone 1?

[Harrison_Pettigrew]

Hi @BenKo — fully understand if the review is done, but leaving this here in case any of the current applicants don’t work out.

The detail that would make or break this project is the 50-client account matching in WF1 (AI labeling). If that mapping is just a static keyword list, it’s brittle — client names in real emails rarely match exactly what’s in a lookup table, especially across DE/EN/FR. A more robust approach is a fuzzy-match layer before the LLM classification call, so “Muster GmbH” in the subject doesn’t go unmatched because the client is stored as “Muster” in the sheet. That’s the kind of thing that looks fine in testing and breaks in production on week two.

The build order I’d recommend: WF1 first, tested thoroughly against real email samples across all 5 categories and all 3 languages, before touching WF2-6. Everything downstream depends on WF1 firing correctly. Most developers will build all six in parallel and discover the misrouting issue only once clients start complaining.

For the loop protection in WF6, the standard approach is checking a sent-reply label or a reference number log in Sheets before triggering the auto-reply. The tricky case is re-opened threads where a client replies to the original confirmation email days later — that needs explicit handling.

Background: self-taught automation developer, production n8n workflows, comfortable with Gmail API + Google Sheets + Slack + Claude for multilingual classification. Available now.

If the current candidates don’t pan out, happy to send a structured proposal with fixed rates per workflow. One question: is the booking engine REST API already documented, or would that need to be scoped during the project?

[Suhail_Narot]

Hi,

This is a well-specified project and I’d like to build it.

On the 6 workflows:

1. Email labeling — AI classification via Gmail API + Claude is exactly what I use. I’d build a structured classification node that outputs consistent JSON labels and handles edge cases (multi-category emails, German-language content).
2. Hotel update forwarding — straightforward conditional routing post-classification, with error handling if no match.
3. Urgent escalation — keyword detection + Slack webhook notification, can include sender/subject/excerpt in the Slack message.
4. Request status tracking — Google Sheets intake with auto-generated reference numbers, timestamps, and deduplication logic.
   5-6. (Happy to review the full requirements doc for the remaining two.)

On German language: Claude handles German text natively — classification, extraction, and structured output all work well in German without translation.

I work from South Africa (UTC+2), English communication is my default. I can turn around a detailed quote within 24 hours of seeing your requirements document.

narotsuhail@gmail.com or DM here.

— Suhail Narot | Fajr AI | fajrai.net

[timai11]

Hi — this is well-scoped enough to start with a fixed pilot instead of a long discovery.

I’d split the first paid milestone into: Gmail label/classification for 2–3 categories → Google Sheets request tracker with unique ref/status timestamps → Slack urgent alert + error log. Once that path is stable, the remaining workflows can be cloned/extended safely.

I can help build the first slice in n8n with clean handoff docs and loop-protection for auto-replies.

Contact: travisofwork@gmail.com

[zynate]

Hey BenKo

This is more of a multiple API calling stuff. Most workflows fail because data is going and coming through different API and most people make it complex. We can do it simply and can provide you the full documentation, even you can handle it. We have been doing this thing for quite a time now and the stck youve mentioned is our core stack. Plot twist you dont have to blind buy. Visit our website, have a glance through our portfolio and let us know what you think!

[nxning_ai]

Hi zynate! Interesting use case with Freelancer - n8n Automation Expert – Gmail / Google Sheets /.

I build n8n automations for businesses - custom workflows, API integrations, AI platforms.

DM if you need support!

[zfahrny]

Hi,

This looks like an email-ops reliability build, not a vague AI experiment. The fragile parts are classification consistency, status tracking, and making sure urgent/booking-change messages do not silently disappear.

I would build it in a safe sequence:

1. lock the five-label classification schema and fallback rules,
2. run classification + Google Sheets logging in test mode,
3. add Slack escalation for urgent/keyword cases,
4. keep ambiguous or low-confidence emails in manual review,
5. document prompts, labels, failure states, and where the team can edit rules later.

I can work async from the requirements doc and sanitized sample emails. I would not auto-send or auto-forward anything irreversible until test output is approved.

Sample diagnostic/handoff artifact: Automation Failure Rescue sample diagnostic pack · GitHub

If useful, send 5-10 sample emails per category plus the target Sheets/Slack structure and I can confirm the smallest safe milestone.

[marcelobolla]

Hi BenKo,

This project is a clean fit for my stack. Gmail API, Google Sheets, Slack, and Claude for multilingual classification (DE/EN/FR) are all tools I work with directly.

Here’s what I’ve built that’s relevant:

-> AI agent connecting WhatsApp Business API + Google Maps API to classify and route delivery information in real time — same pattern as your email labeling + forwarding logic.

-> Bilingual (IT/EN) production website for a hospitality client, managed end-to-end.

For your 6 workflows specifically:

• Email classification via Claude API with fallback rules for ambiguous cases
• • Google Sheets tracking with unique reference IDs and status pipeline
  • • Slack escalation on keyword detection
    • • Auto-reply with loop protection on first-contact emails
    • I work from detailed requirements docs (no discovery needed), deliver clean documented workflows, and communicate in English or Italian. German communication on your side is fine.
  • I’d like to review your requirements document and come back with a structured proposal per workflow. Available to start immediately.
• Marcel Bolla
• marcelobolla3@gmail.com
• Sardinia, Italy (UTC+2)

[CarlosSilveira]

Hi BenKo, I can help with this package if the build is still open.

I am Carlos Silveira, based in Spain/CET. My fit is the integration and reliability layer: Gmail/Google Workspace APIs, Sheets or Postgres tracking, Slack alerts, structured LLM outputs, REST APIs, error handling and handoff documentation.

I would split it into fixed milestones:

1. Gmail classification and labels with JSON output, confidence rules and fallback handling.
2. 2. Hotel update forwarding, urgent escalation rules and Slack alerts.
   3. 3. Request status tracking with unique references, timestamps and daily reporting.
      4. 4. First contact auto-reply with loop protection, tests and documentation.
      5. Initial package estimate: 1,200-2,000 EUR depending on email volume, booking-engine API access and whether tracking stays in Sheets or moves to Postgres. I can also start with one paid workflow first if you prefer to validate fit before the full package.
   4. Example-wise, I can share an anonymized workflow outline/test export for email intake, classification, tracking and alerting without exposing client data.
3. Before locking the price I would need the requirements document, expected daily volume and booking API details. I work from written scope and a funded milestone/deposit before implementation or production access.

[tzlsau]

Hi BenKo,

I’m Kevin. I saw this project and it looks close to the kind of workflow I’m focusing on: Gmail / Google Workspace, Sheets-based tracking, Slack alerts, LLM classification, and clean handoff documentation.

The way I would approach it is to split the build into small, testable workflows instead of trying to make one large black-box automation:

1. start with the Gmail classification workflow and agree on the 5 labels, confidence threshold, and fallback rule;
2. write every important event into a Sheets tracker with a unique reference number and status fields;
3. add Slack escalation only after the classification and tracking path is reliable;
4. keep auto-replies guarded by first-contact detection and loop protection;
5. document each workflow with inputs, outputs, failure cases, and test examples.

For a first step, I would be comfortable preparing a fixed-price proposal after seeing the requirements doc. If you prefer a small paid trial first, I could build or prototype one workflow slice, for example Gmail classification → Sheets row → Slack urgent alert, before quoting the full package.

I have a visual preview of a similar Google Workspace assistant workflow that shows how I think about the review queue, AI summary, and human approval points. Happy to share it if useful.

Best,
Kevin

[Denish]

Hi there,

I can build this automated email classification system for you cleanly using n8n.

Connecting Gmail triggers to Google Sheets and routing alerts to Slack is a core pipeline structure I build regularly. For the AI Classification step, I use optimized system prompts inside n8n’s advanced AI nodes to ensure your incoming emails are accurately categorized (e.g., Support, Sales, Spam) before the data gets routed.

I use a local development sandbox to thoroughly test the data mapping and execution logs first, guaranteeing zero errors before handing the workflow over to you.

Let’s connect via DM or a quick chat to align on your specific spreadsheet layout and get this running for you this week.

Best regards,

Denish

[Work6]

This is a well-scoped build and right in my lane — Gmail-based AI email classification into your 5 categories, label-driven forwarding, keyword → Slack escalation, a Google Sheets status state-machine with reference numbers + full timestamp tracking (Pending → Confirmed), a daily dashboard + Slack digest, and first-contact auto-reply with loop protection.

The AI-classification + Sheets-state-machine core is something I’ve already built a version of (Claude scoring + routing into Sheets): I built an n8n + Claude lead-gen workflow (scrape -> AI-score -> auto-email) - sharing it . The two parts people get subtly wrong are the loop-protection (I handle it with a dedup/seen-key in the Sheet so first-contact only fires once) and the status tracking (an idempotent reference-number generator so a re-processed email never double-counts).

I don’t speak German, but you said English is fine on your side and the requirements doc is already done — so this is a pure build for me. Want me to do the first workflow (the email labeler) as a paid milestone so you can see how I work before committing to all six?

[amir_m]

Hi @BenKo,

Your project maps directly to what I build and maintain in production right now.

How I’d approach the 6 workflows

1. Email Classification (AI + Gmail API)

Single LLM call per email, structured JSON output with category + confidence score. For multilingual classification (DE/EN/FR), Claude handles this best in my experience — one prompt with explicit category definitions and language detection, no pre-translation step needed. For your 50 client accounts, I’d build a reference table (Sheets or Supabase) that the classifier matches against, so adding account #51 is a config change, not a workflow edit.

2. Hotel Update Forwarding

Downstream from workflow 1. Once classified as “hotel update”, n8n triggers a Gmail draft-and-send to the sales inbox. I add a deduplication check (Gmail Message-ID stored in tracking layer) so the same email never forwards twice, even if the workflow re-runs.

3. Super Urgent Escalation

Keyword detection layer runs parallel to AI classification. This is intentional — keyword matching is instant and deterministic, you don’t want urgent escalation waiting for an LLM response. Slack notification includes email subject, sender, matched keyword, and a deep link back to the Gmail thread.

4. Request Status Tracking

This is where I’d make the case for Supabase (PostgreSQL) over Google Sheets. Your status pipeline has 7 stages (Pending → Acknowledged → Offer Sent → Follow Up → Negotiation → Confirmed/Declined), timestamps at each transition, and unique reference numbers. Google Sheets works for v1, but at ~50 active clients with daily volume, you’ll hit row-lock conflicts and API quota limits within weeks. PostgreSQL handles this natively with proper indexing and zero throttling. If the team needs visibility, I build a simple read-only Sheets sync or a dashboard view.

That said, if you want Sheets first, I’ll build it clean and migration-ready.

5. Daily Reporting + Slack Digest

Cron trigger, pulls current pipeline state, generates a summary (X new requests, Y pending follow-up, Z confirmed this week), posts to Slack with a link to the dashboard. Formula-based dashboard stays in Sheets either way — it’s the right tool for that layer.

6. Auto-Reply + Loop Protection

First-contact detection via Gmail thread history check. Loop protection at three levels: exclude emails from own domain at trigger level, check In-Reply-To and X-Auto-Reply headers, and maintain a cooldown table (one auto-reply per sender per 24h). This prevents the classic auto-reply war between two automated systems.

All workflows include error handling with auto-retry and a dedicated error notification workflow that pings Slack with the exact workflow name, node, and error message.

About me

I’m an automation engineer working with self-hosted n8n, Supabase, and AI APIs (Claude, OpenAI, Gemini).

My primary production system is an AI-powered platform for a Central Asian marketplace — currently in integration phase, 600+ sellers onboarding. The architecture:

• Self-hosted n8n on DigitalOcean, two-workflow split (Collector + Generator)
• Gemini Vision API for product analysis, AI image generation for marketplace cards
• Supabase backend with PostgreSQL row locking for concurrent data processing
• Telegram bot interface + WebSocket-based widget embedded in the web app (adaptive desktop/mobile UI)
• Seller correction system, color variant pipeline, admin monitoring, error routing

This is not a side project. The same client has contracted me for three additional systems on the same infrastructure: content moderation, logistics automation, and operator workflow tools.

Pricing

Workflow	Fixed price
1. AI Email Classification	$350
2. Hotel Update Forwarding	$100
3. Urgent Escalation → Slack	$200
4. Request Status Tracking (Supabase)	$400
5. Daily Reporting + Slack Digest	$250
6. Auto-Reply + Loop Protection	$200
Total (individual)	$1,500
Package price (all 6)	$1,250

If you go with Google Sheets instead of Supabase for workflow 4, the price drops to $300 for that workflow ($1,150 package).

Each workflow delivered with documentation: workflow JSON, setup instructions, and a brief Loom walkthrough.

Happy to do a screen share and walk through my production n8n instance. I can show you the exact patterns I’d use for your email classification and status tracking.

Amir
[rrem69267@gmail.com]

[Raiym]

Hey,

Saw your post and this one actually matches what I’ve been building lately.

I just wrapped up an AI consultant for an industrial parts store — n8n + Pinecone + GPT-4o-mini, the bot searches through 7,000+ products and answers like a real sales rep. Before that I built a daily media monitoring pipeline — 27 political figures tracked across Telegram and Google News, results in Sheets every morning at 7AM.

Your setup (Gmail + Sheets + Slack + LLM) is the same stack, just different use case.

For the 6 workflows:

Email labeling with AI classification (DE/EN/FR) - $180
Hotel update forwarding - $60
Urgent escalation to Slack - $80
Request tracking in Sheets with status pipeline - $200
Daily Slack digest - $100
Auto-reply with loop protection - $80

Total $700, delivery 7-10 days after you share access. 50% upfront, 50% on delivery.

I work async only, no calls - everything in writing which keeps things cleaner anyway.

If the scope works for you, just reply here.

Raiym