WorkCue
Workflows7 min read

Automate Freelancer Lead Intake to a Proposal (Make or Zapier)

A practical Make/Zapier workflow from form or calendar lead intake to a human-approved proposal draft — filters, CRM hygiene, AI outline steps, and a ship checklist.

Most freelancers lose deals in the quiet gap between someone filled out your form and a proposal lands in their inbox. Not because the offer is weak — because intake sits in email, discovery notes never become a draft, and follow-up starts too late. This tutorial shows how to automate freelancer lead intake to a proposal with Make or Zapier without auto-sending commercial promises.

Sibling paths: full money loop in lead → invoice with Make; call-first drafting in meeting → proposal workflow. Platform choice: Make vs Zapier.

What you are building

StageAutomation jobHuman job
1. IntakeCapture form/calendar leadSpot spam / bad fit
2. NormalizeMap fields + dedupe by emailFix wrong company names
3. QualifyFilter by budget/typeDecide pursue / pass
4. HubCreate CRM/Notion deal stubRun discovery if needed
5. DraftAI outline → Doc/DraftsPrice, scope, send
6. RemindTask at +3–5 business daysEdit follow-up; stop when done

Hard rule: Automation prepares a draft proposal and reminders. You own price, timeline, guarantees, and the Send button.

Make or Zapier for this workflow?

SituationPrefer
Linear form → CRM → Slack → DocEither; Zapier is often faster day one
Budget bands, routers, array fields, error pathsMake
Non-technical teammate will maintain itZapier
You already like visual scenariosMake

If you are unsure, prototype the first three modules in whichever UI you will actually open this week. Rebuild later if cost or branching hurts — that pattern is normal. Broader context: Zapier alternatives and AI automation tools for small business.

Prerequisites (30 minutes before the canvas)

  1. Intake source — Tally, Typeform, Google Form, or Calendly/Cal.com booking with questions.
  2. System of record — Notion database, Airtable, HubSpot free, or a Sheet you respect (CRM patterns).
  3. Status list — at minimum: new, qualified, needs_discovery, proposal_draft, proposal_sent, won, lost.
  4. Proposal skeleton — a Google Doc / Notion template with placeholders, not a blank page.
  5. Rate card offline — sticky note or sheet the model cannot edit.
  6. Consent + privacy — only fields you need; no passwords or sensitive attachments in the automation.

Draw statuses on paper first. Fuzzy status design creates more “automation debt” than missing modules.

Form fields that make automation sane

Required:

  • Name, work email, company (or “solo”)
  • Project type (controlled list)
  • Budget band (bands beat free text)
  • Target start / deadline
  • Short problem statement (2–5 sentences)
  • How they found you
  • Consent checkbox for email follow-up

Optional: link to brief, industry, timezone. Skip twenty vanity fields that tank conversion.

Scenario blueprint (modules)

Build in slices — ship 1–5 before AI drafting:

  1. Trigger — form submit or booked call with answers
  2. Normalize + dedupe — map fields; update by email, do not clone
  3. Filter — hard disqualifiers (budget floor, no email)
  4. Create/update deal + Slack/email one-screen summary
  5. Router — Path A draft stub vs Path B discovery booking
  6. Optional AI — CRM snapshot + proposal outline into Doc/Drafts
  7. Notify you to review & price; on proposal_sent, schedule a follow-up task (not auto-email)

Step-by-step build

1. Capture clean inputs

Connect the form or scheduler. Map every required field. If a required field is empty, halt — do not invent values.

2. Deduplicate before you celebrate

Search the CRM by email. Same person often arrives from Instagram, a referral, and your site form. Duplicates break proposal history and make you look disorganized.

3. Qualify with filters, not vibes

Examples:

  • Budget band below your floor → status lost_auto + polite pass template (draft only, or a single canned reply you already approve)
  • Enterprise domain + high band → priority Slack
  • “Need it yesterday” + unclear scope → Path B discovery, not a rushed proposal

Automation should protect your calendar, not flood it.

4. Create the deal stub

When qualified:

  • Create Notion/Airtable/HubSpot record with source + UTM if present
  • Attach the problem statement and budget band
  • Create an empty discovery notes section
  • Link your proposal template copy for this deal

5. AI outline — draft only

Only after filters pass. Use an AI module (Make/Zapier OpenAI/Claude step) or a later paste into ChatGPT/Claude with a locked prompt. Pattern from ChatGPT proposal prompts:

Using ONLY the intake fields below, draft a proposal outline:
Understanding, Approach, Deliverables bullets, Assumptions, Open questions.
Do not invent prices, timelines, case studies, or guarantees.
Flag missing info as questions for me.
Intake:
"""
{{mapped fields}}
"""

Paste the outline into your Doc template. Generators overview: best AI proposal generators for freelancers.

6. Human commercial gate

You still:

  • Set package options and dollars
  • Set revision limits and timeline you can hit
  • Remove invented proof points
  • Send from your real email / proposal tool

If discovery is incomplete, do not let the scenario invent scope. Route to a booking link instead.

7. Proposal-sent reminders

When you flip status to proposal_sent (manual or “I sent it” button):

  • Create a task at +3 to +5 business days: “Follow up proposal — ”
  • Optional: generate a Gmail/Outlook draft follow-up for you to edit

Do not auto-send commercial nudges until you trust the copy cold. Drafts beat regrets.

Weekend MVP (four hours)

If you only have an afternoon:

  1. Form → CRM row + email/Slack notify
  2. Filter: budget below floor → stop
  3. Qualified → copy of proposal Doc + task “Draft proposal today”
  4. Manual proposal_sent → reminder task in Todoist/Notion

That alone kills the “I forgot they applied” failure mode. Add AI outline and fancy routers in week two.

Testing checklist

  • Fake lead appears once after two submits with the same email
  • Disqualified lead never gets a proposal Doc
  • Missing budget band does not invent a price
  • AI outline contains no dollar amounts you did not provide
  • Reminder fires only after proposal_sent
  • Error path (CRM down / empty email) alerts you instead of silent fail

Privacy and trust

Prefer work emails and short project descriptions — not personal IDs or payment data in AI modules. Check NDAs before sending decks into third-party AI steps. Keep API keys in Make/Zapier connections. After you win, continue with client onboarding and lead → invoice.

FAQ

Should I automate sending the proposal?

Not at first. Automate prep and reminders. Auto-send only when amounts, tax, and scope come from fields you typed and have run for weeks without errors.

Is Make required, or is Zapier enough?

Zapier is enough for a linear intake → CRM → Doc → task path. Choose Make when you need routers, richer data shaping, or better economics on chatty scenarios. Compare: Make vs Zapier.

Where does a discovery call fit?

If the form is thin, Path B should book a call and then use meeting → proposal. Do not force a priced proposal from three vague sentences.

Can AI set my prices?

No. Keep a rate card outside the model. AI may label option tiers (Essential / Standard) from a fixed menu you provide — you assign dollars offline.

What if I only get a few leads a month?

Start when missed follow-ups cost real deals — not because a thread said every solopreneur needs twelve zaps.

Bottom line

Automating lead intake → proposal is about status hygiene, dedupe, and a human gate on commercials — not a fully lights-out closer. Ship capture, qualify, stub, and reminder first; add AI outlines second. Prefer Make when the path branches; Zapier when you want the shortest linear build. Then connect the rest of the revenue loop with lead → invoice once proposals are boringly reliable.

Get the wider stack on one page via the Freelancer AI Stack checklist.

Related guides