You don’t need to be a developer anymore to build tools that actually work for your business.
AI gave us something we’ve never had before: the ability to recreate our own versions of expensive apps, completely customized for our needs, without writing a single line of code.
And when you combine that with automation platforms and AI agents, you can build systems that do what multiple SaaS products do. The ones you’re paying $30, $50, $100 a month for.
So today I’m showing you how to recreate a simplified version of Fyxer, an AI tool that drafts emails for you, using an n8n workflow that costs pennies instead of the +$30/month plan.
I built the whole thing in 30 minutes, but since I’m giving you my workflow to clone, yours will take 2 minutes to set up once you connect the two tools we’ll be using: Gmail (or Outlook) and OpenAI (or your preferred AI model).
So if you’re buried in customer emails, losing leads because you can’t respond fast enough, or spending hours every week writing the same replies over and over, this workflow will help.
The difference is that you own the system. You can change it, expand it, customize it however you want, and you’re not locked into another subscription. And because everything runs through your own accounts, you also keep full control over your data.
If email drafting isn’t a bottleneck for you, start here
If email drafting isn’t a main pain point for you right now, here are a few other automations I’ve built that you can clone and have running for yourself in minutes.
The Inbox Manager: How I Built a System That Keeps Email Under Control: Sorts your personal inbox with AI, summarizes newsletters, delivers a daily digest, and takes 15 minutes to set up.
Built for personal email right now, but I’m considering a business version that splits incoming messages into “awaiting reply”, “FYI”, “needs action” and routes them automatically. If you’d use that, let me know:
How I Built an AI Personal Assistant That Works 24/7: Manages email, calendar, notes, and ideas through one chat interface. Costs under $3/month. Works like an actual assistant.
How I Built a Reddit Market Intelligence AI Agent That Saves Me 3+ Hours Per Week: Weekly digests of market insights, trends, opportunities from Reddit, without manual scrolling. Useful if you do market research or track competitors.
Automate LinkedIn Posting With n8n: My Full Setup and Step-by-Step Guide: Repurpose content, schedule, and publish LinkedIn posts automatically.
Alright, back to our email auto-drafter.
The email response problem costing your business money
I know some of you might hesitate at the idea of letting AI write drafts.
But the truth is a lot of people already use AI to write emails, they’re just doing it manually, one prompt at a time.
This isn’t any different. We’re talking about repetitive questions, documented policies, and situations where speed helps customers or helps you catch leads faster.
You still review every draft. You’re not removing the human part, you’re removing the friction that slows you down.
Because every time you or your team opens a new email, it goes through the same sequence: scan the content, figure out what category it belongs to, decide if it’s urgent, think through an appropriate response, check any relevant context or company internal docs, then type it all out.
For businesses, this compounds fast:
Slower response times mean lower customer satisfaction scores.
Missed or delayed replies become missed opportunities.
An inquiry that sits for 48 hours while you’re buried in other work? That potential client probably already moved on to a competitor who responded in 20 minutes.
So if you’re reading this and thinking “yeah, but my situation is different”, you’re probably right. This automation isn’t for everyone.
It only makes sense if you’re in one of the specific scenarios where email volume is actually costing you money or customers.
When this workflow makes sense
You run a small service business where speed wins deals.
Veterinary clinics getting emergency calls and appointment requests. Dental offices juggling scheduling and insurance questions. Plumbers, electricians, HVAC contractors fielding quote requests. The client who gets a reply in fifteen minutes books with you. The one who waits three hours books with someone else. Same quality of work, different response time, completely different revenue.
You operate an e-commerce store and your support inbox never stops.
Product questions. Shipping status updates. Return requests. “Where’s my order?” messages. Most of these follow patterns. An AI trained on your policies drafts accurate responses instantly. You review and send instead of writing from scratch. Your team handles three times more tickets in the same amount of time.
You work in real estate where the first person to respond books the showing.
Property inquiries hit your inbox while you’re at another showing. By the time you follow up, they’ve already scheduled with someone else. Automated drafts mean you’re always first, even when you’re busy.
You’re running a SaaS product with support tickets, how-to questions, and onboarding issues coming in nonstop.
If most of what lands in the inbox is the same handful of questions in different shapes, let automation handle those. Free your team to focus on the issues that actually need judgment, context, and real conversation with your customers.
You manage a larger customer support operation processing hundreds or thousands of emails daily.
Even automating 50% of responses frees your team to focus on escalations and edge cases. They stop feeling like robots copying lines from internal docs into tickets and get to spend their time on work that actually needs their expertise.
You’re an NGO or nonprofit doing more with less.
Donor inquiries, volunteer applications, partnership requests, media outreach, all handled by a two-person team already stretched thin. Let automation cover the repetitive answers so the team can stay focused on real community work.
You’re a consultant or agency owner and your inbox is literally your pipeline.
You still need to personally manage relationships and handle the sensitive conversations yourself, but automation can help you get to every lead faster. It drafts the first version, sorts inquiries by priority, and pre-fills the things you repeat to every client like your fees, timelines, and questions. You stay in control of the relationship, you just stop losing time on the repetitive parts.
Here’s exactly what this automation does
The workflow is simpler than you think, but as powerful as most tools you pay for.
It watches your Gmail. The moment a new email arrives, the workflow picks it up.
It reads the entire message, including the subject, body, sender, and any extra details.
Then it sends that email to an AI classifier you control. This isn’t ChatGPT guessing. It’s a model trained on the categories you set (support emails, leads, order or product questions). Whatever buckets make sense for your business.
Once the email is classified, the workflow applies the right Gmail label automatically so your inbox stays organized without you doing anything.
Then it generates a draft reply using your system prompt. That prompt carries your voice, policies, and usual tone, so different types of messages get different types of replies.
It saves the reply as a draft inside the original email thread, with the correct subject and formatting.
You open Gmail, review the draft, edit if needed, and hit send. Instead of spending minutes writing each reply, you spend seconds approving it.
How far you can take this system
What I just described handles 80% of your email work. But the real advantage comes from customizing it for your actual business needs.
You can expand this into a complete email operations system that classifies, prioritizes, enriches, routes, responds, follows up, schedules, and learns, all without you micromanaging it.
Here’s some ways you could expand it:
Connect it to your actual company knowledge
Right now the AI drafts replies based on its training and your system prompt. But what if it could read your documentation first?
Your FAQ pages. Your troubleshooting guides. Your internal wikis. Your policy documents. Your best email responses from the last two years. Everything your team references when they’re responding manually.
This is RAG (Retrieval Augmented Generation). Instead of an AI that guesses answers, you get one that reads your resources first, finds the relevant information, then responds using that.
The customer gets the same answer either way. The difference is that one requires someone on your team to dig through internal docs to find it and the other finds it automatically.
Turn classified leads into CRM records automatically
When the workflow identifies a new lead, it doesn’t just label and draft, it creates a record in your CRM. Notion, Airtable, HubSpot, Pipedrive, Google Sheets, whatever you use. Contact email, name, company, classification, timestamp, source. Your pipeline builds itself.
Add lead scoring so you never miss a hot opportunity
After classification, another AI node assigns a score (1-5). High-score leads trigger a Slack notification to your sales channel and get a “🔥 Hot Lead” label in Gmail. You always know which conversations to prioritize.
Use sentiment analysis to catch fires before they spread
An angry VIP customer needs different handling than a general question. The AI detects sentiment, tags it, and routes accordingly. Negative sentiment from a high-value account? That goes straight to the account manager via Slack DM with high priority, so your team sees problems the moment they arrive.
Create support tickets automatically and route them to the right person
Support emails become tickets in Zendesk, Linear, Jira, HubSpot Service, or your tool of choice. The AI decides who should handle it based on product area, language, priority level. It sets SLA timers automatically, so your team knows exactly what needs attention and when.
Enrich inbound leads before you even respond
New business inquiry comes in. The workflow pings Clearbit, Apollo, or your own database to pull company information (size, industry, funding). Then the AI uses that to tailor its response. You’re not sending the same pitch to a solo founder and an enterprise VP. You’re speaking their language from message one.
The point is this: what starts as a simple draft-generator can evolve into a complete email operations system that classifies, prioritizes, enriches, routes, responds, follows up, schedules, and learns, all without constant human intervention.
If you want something this complex and custom-built for your specific business, let’s jump on a call.
How to build it: the complete setup
In this section you’ll see exactly how the automation works and how to set it up yourself.
I’m giving you a fast setup you can follow in a few minutes, plus a full node-by-node walkthrough if you want to understand every part of the workflow.
Use whichever path fits how hands-on you want to be.
Your quick-start setup
This is everything you need to do to get the workflow running in a few minutes:
Step 1: Download the JSON file, which contains the entire workflow, including my prompts.