Episode 1·

Ship a Zero-Touch Client Proof Digest (Gmail + Make + Notion) in 90 Minutes

Intro

This episode is for solo automation consultants who are losing clients not because their work isn't valuable, but because it's invisible. You'll get a complete 90-minute build that transforms your background automations into weekly proof of value that clients can see, share, and defend internally.

In This Episode

Jordan tackles the client proof problem head-on by building a zero-touch weekly digest system that automatically compiles your invisible automation work into client-readable updates. Using Gmail v4's search operators, Make.com's LLM modules, and Notion's client portal capabilities, you'll create a system that pulls everything you shipped last week, summarizes it with AI guardrails, and delivers proof of value to clients every Tuesday morning. The build covers Gmail module configuration, structured LLM prompts with source anchoring, Notion page templates with visual hierarchy, optional Slack integration for team visibility, and logging for churn-prevention tracking. Jordan also addresses common gotchas like Gmail reauthorization cycles, Notion rate limits, timezone math, and provides alternatives for Zapier, n8n, and HubSpot users.

Key Takeaways

  • Build a Gmail-to-Make-to-Notion automation that compiles weekly work into client-readable digests with zero manual effort
  • Use structured LLM prompts with source anchoring and token limits to prevent hallucinations while summarizing client work
  • Position the digest as 'Operational Visibility' and charge $200/month per client as a flat add-on that prevents churn

Timestamps

  • Make Help Center + Apps Docs

    help.make.com

    • - Make’s Gmail v4 app is the default as of early 2026 and includes Search emails, Get an email, Reply, Send draft, List attachments/media, and Make an API call modules.
  • Make Apps Docs — Gmail modules

    apps.make.com

    • - Make Gmail v4 ‘Search emails’ supports Simple filters or Gmail filter query, Label/Folder targeting, a Limit (≤500) per cycle, and Content format selection (Full, Raw, Metadata, Minimal); HTML body can be included.
  • Make Apps Docs — Gmail

    apps.make.com

    • - Personal @gmail.com connections to Make must be re‑authorized every 6 months per Google policy (effective June 3, 2024).
  • Notion Developers — Request limits

    developers.notion.com

    • - Notion API rate limit averages 3 requests/second per integration, with 429 rate_limited responses advising Retry‑After backoff.
  • Make Apps Docs — Notion

    apps.make.com

    • - Make’s Notion docs note a 25‑page reference limit that requires pagination via List Page Property Items when has_more is true.
  • Google Developers — Gmail API Usage limits

    developers.google.com

    • - Gmail API quotas: 15,000 quota units per user per minute and 1,200,000 quota units per project per minute; per‑method units vary (e.g., drafts.get=5, drafts.create=10).
  • Google Developers — Handle Gmail API errors

    developers.google.com

    • - Gmail API enforces a per‑user concurrent request limit in addition to rate limits, returning 429 Too Many Requests; clients should implement exponential backoff.
  • Help Scout Developers — Rate limiting

    developer.helpscout.com

    • - Help Scout Mailbox API is rate‑limited per account; write requests count as 2 units; response headers include remaining and retry‑after seconds.
  • Zendesk Developer Docs — Rate limits

    developer.zendesk.com

    • - Zendesk Support/Help Center API requests per minute vary by plan (e.g., 700 rpm on Enterprise; 200 rpm Team; up to 2500 rpm with High Volume API add‑on).
  • Make Apps Docs — Email (built‑in)

    apps.make.com

    • - Make Email app supports generic IMAP/SMTP and Microsoft OAuth connections for non‑Gmail mailboxes.
  • Make Apps Docs — Google Gemini AI; Google Vertex AI (Gemini)

    apps.make.com

    • - Make Google Gemini AI app supports chat completions and file upload via Google AI Studio API key; Vertex AI Gemini is also available as a separate app.
  • n8n Docs — IMAP/Notion/OpenAI/Gemini

    docs.n8n.io

    • - n8n provides IMAP Email Trigger, Notion node (create/search pages & DB items), OpenAI node (V2 supports Responses API; Assistants removed), and Google Gemini nodes (chat and embeddings) with AI Studio API key.
  • Zapier Help Center

    help.zapier.com

    • - Zapier provides ChatGPT (OpenAI) app supporting Responses API (updated June 3, 2025) and Google AI Studio (Gemini) app (guide updated Jan 21, 2026).
  • HubSpot Developers — Tickets & Conversations

    developers.hubspot.com

    • - HubSpot CRM Tickets API endpoint: GET /crm/v3/objects/tickets with properties and associations; Conversations Threads endpoint: GET /conversations/v3/conversations/threads (with inboxId filter restriction).
  • Google Developers — Reserved/common search operators

    developers.google.com

    • - Gmail search operator newer_than supports time windows like newer_than:7d and can be used in Gmail filters/queries.
  • PDFMonkey Pricing

    pdfmonkey.io

    • - PDFMonkey offers HTML→PDF via REST with subscription tiers and optional pay‑as‑you‑go overages (on Pro+/Premium).
  • PDFShift

    pdfshift.io

    • - PDFShift provides an HTML→PDF API with no stated file size limit; common alternative to Documint/pdfmyhtml.
  • Make.com Help Center: New Gmail app is now available (release note)

    help.make.com

    • - Make’s Gmail v4 modules now default
    • - Confirms the exact Gmail module set and setup path we’ll show on‑screen (Search emails, Get an email, Reply, Send draft, List attachments, Make an API call).
  • Gmail modules — Apps Documentation (Make)

    apps.make.com

    • - Search emails module fields and limits in Make (Gmail v4)
    • - Gives precise UI fields to narrate while screen‑sharing (Filter type: Simple vs Gmail filter, Label/Folder pickers, Limit ≤ 500, Content format, Include HTML body).
  • Notion — Apps Documentation (Make)

    apps.make.com

    • - Notion module set and v2 ‘Data Source’ updates
    • - Explains the ‘Create a Page’ vs database/data source IDs and the 2025 API version shift that can break legacy mappings — a key ‘gotcha’ to call out.
  • n8n workflow template

    n8n.io

    • - Weekly task planner with Notion + email digests
    • - Concrete n8n example that mirrors our zero‑touch digest pattern for listeners using n8n instead of Make/Zapier.
  • Medium case study (Guillaume Micallef)

    medium.com

    • - Freelancer weekly automation with Make pulling Notion + Gmail + finance
    • - Shows a solo operator automating weekly digests and follow‑ups with Make; useful as an anecdotal proof that clients see concrete outputs weekly.
  • Zapier integration directory

    zapier.com

    • - Digest by Zapier + Gmail + Notion recipes
    • - Alternative stack proof for listeners standardizing on Zapier; also demonstrates ‘accumulator’ model (Digest) before sending a weekly email.
  • Processter case study

    processter.com

    • - Notion→Slack reporting saved 4 hours weekly
    • - Third‑party evidence that automated weekly status reporting reduces manual time and lifts perceived value.

Jordan: Your best automation work is invisible. That's the problem. You ship a workflow that saves your client twelve hours a week, and three months later they cancel because — and I'm quoting here — "we're not really sure what you're doing for us anymore." The automation is working so well they forgot it exists. Here's what nobody tells you about running a solo automation consultancy: the better your work, the less your clients see it. And when they stop seeing it, they stop valuing it. So today I'm going to show you exactly how to build a zero-touch weekly digest that makes your invisible work visible — ninety minutes to set up, runs forever, and here's the kicker — it actually reduces churn without adding a single meeting to your calendar.

Jordan: By the end of this episode you'll have a specific automation sequence — Gmail to Make to Notion, with optional Slack and PDF outputs — that compiles everything you shipped last week into a client-readable digest. No manual work, no weekly check-ins, just proof of value landing in their inbox every Tuesday morning.

Jordan: Let me show you exactly how this works in production. I'm pulling up Make right now — and yes, you can build this exact same thing in Zapier or n8n, I'll show you the equivalents — but the core pattern stays the same. You're solving the client proof problem.

Jordan: Here's what actually happens when you run a solo consultancy. You spend Monday building a complex Salesforce-to-Stripe reconciliation workflow. Tuesday you're debugging webhook failures. Wednesday you ship a customer onboarding sequence that cuts their manual work by eighty percent. And by Friday, your client has no idea any of this happened because it all runs in the background. AgencyAnalytics did this study on client retention — turns out the number one reason clients leave isn't price or results. It's perceived inactivity. They literally think you're not doing anything.

Jordan: So we fix that with radical visibility. Every week, this automation pulls everything you did — every email, every task, every fix — and packages it into a digest that takes them thirty seconds to read but makes your value undeniable.

Jordan: Alright, let's build this. First thing — the stack. You need four pieces. Gmail or your helpdesk for source data. Make, Zapier, or n8n for orchestration. An LLM — OpenAI or Gemini — for summarization. And Notion for the client portal. Optional fifth piece is Slack if your clients live there.

Jordan: Starting with the trigger. New scenario in Make. Add a Schedule module — set it to run weekly. Now here's the first decision point that matters: what timezone? Always use the client's local timezone. I learned this the hard way — sent a digest at 6 AM Pacific to a client in London. They got it at 2 PM, buried under a full day of email. Engagement dropped to zero.

Jordan: Next module — Gmail. And this is important, Make just updated to Gmail v4 in January. See this dropdown? Filter type. You want "Gmail filter" not "Simple." This lets you use Gmail's search operators. Watch this — I'm pasting in: `label:ClientName newer_than:7d -category:promotions`. That pulls everything from the last seven days with their label, excluding promotional garbage.

Jordan: Set your limit to 100. You might think you need more, but if you're sending more than a hundred emails per client per week, that's a different problem. Content format — set to Full, and check "Include HTML body." You need the full context for the LLM to work with.

Jordan: Now if you're not on Gmail — maybe you're using Help Scout or Zendesk — same pattern. Make has native modules for both, or you can use the HTTP module to hit their APIs directly. Just watch the rate limits. Help Scout counts writes as double, Zendesk varies by plan — typically 200 to 700 requests per minute.

Jordan: Okay, here's where most people mess up. They dump all this email data straight into ChatGPT and hope for the best. That's how you get hallucinations, made-up metrics, and summaries that sound like they were written by a robot who's never actually worked with a client.

Jordan: Add the OpenAI module — or Gemini if you prefer, both work. Here's the prompt structure that actually works. Four sections: Shipped, In Progress, Blockers, Next Week. Strict JSON output. Maximum five bullets per section. And this is crucial — every bullet must include a source reference. Watch...

Jordan: "Analyze these emails and tasks from the past week. Output JSON with four arrays: shippeditems, inprogress, blockers, nextweek. Each item should be a single sentence with a sourceid. Maximum 5 items per category. Focus on client-visible outcomes, not internal process."

Jordan: See that token limit? Set it to 500. That forces conciseness. Nobody reads a two-page digest. They read five bullets per section, max.

Jordan: The hallucination problem is real, by the way. There's this paper from October 2024 — multi-document summarization increases hallucination risk exponentially. That's why we anchor every claim to a source email or task. If the LLM says "Completed API integration," there better be an email thread proving it.

Jordan: Now we package this for the client. Add a Notion module — Create a Page. Connect to your client portal workspace. And here's a gotcha that burned me last month — Notion updated their API in 2025. See this field? "Data Source ID" not "Database ID." If you're using legacy modules mixed with new ones, they'll fail silently. Spent three hours debugging that.

Jordan: The page template has four callout blocks matching our sections. Shipped gets a green checkmark. In Progress gets the construction emoji. Blockers gets a red warning triangle. Next Week gets the calendar icon. Visual hierarchy matters — your client should understand the status in literally one second.

Jordan: Below that, add metadata. Date range, number of inputs processed, total tasks completed. Clients love numbers even when they don't mean anything. "Processed 47 items this week" sounds way more impressive than "did some stuff."

Jordan: Next module — Send an Email. This is the actual digest going to the client. Keep it short. "Hi Sarah, Here's what we accomplished this week:" Then your four sections as bullets. Then one line: "Full details in your portal" with the Notion link. That's it. No essays, no explanations.

Jordan: If they use Slack, add that module too. Post to their shared channel — just the top five items and the portal link. The point isn't to duplicate everything. It's to create multiple touchpoints without being annoying.

Jordan: Some clients want PDFs. I know, I know. But if they want a file they can forward to their boss, you give them a file. PDFMonkey or PDFShift both work — about three cents per PDF. Only generate these on demand though. Most clients never open them.

Jordan: Last step — and people always skip this — logging. Add another Notion module. Create an entry in your "Digests Sent" database. Timestamp, scenario run ID, which client, email message ID. Why? Because in three months when a client says "I never get updates from you," you can pull up twelve weeks of digests with read receipts.

Jordan: Let me show you this actually running. Here's last week's run for a test client. Processed 43 emails, 8 completed tasks. The digest pulled out five shipped items including a Stripe webhook fix that saved them four hours of manual reconciliation. Two items in progress. One blocker about API rate limits. Three things planned for next week.

Jordan: The email is 147 words. The Notion page has full details with links back to every source. The Slack message is just the highlights. Total execution time: 18 seconds. Total cost including LLM tokens: about twelve cents.

Jordan: Now let's talk about what breaks. First gotcha — Gmail reauthorization. If you're using a personal Gmail account, not Google Workspace, you have to reauthorize every six months. Mark your calendar or this will break at the worst possible time.

Jordan: Second — Notion rate limits. Three requests per second average. If you're running multiple clients in parallel, add delays between them. See this? Sleep module, 2 seconds between each client. Prevents 429 errors.

Jordan: Third — and this one's subtle — timezone math. Make's scheduler uses UTC. Your client is in Eastern time. Daylight saving happens. Suddenly their Tuesday morning digest arrives Monday night. Test this with actual timezone changes, not just assuming it works.

Jordan: The Gmail search limit is 500 messages max. If a client generates more than 500 emails a week, you need to batch or filter more aggressively. Also maybe charge them more because that's insane volume for a solo consultant.

Jordan: LLM token costs add up if you're not careful. A single digest costs about eight cents in tokens. Ten clients, weekly runs, that's 40 dollars a month. Not huge, but price it into your retainers.

Jordan: Here's how you package this for clients. Don't call it "automated reporting" — that sounds like spam. Call it "Operational Visibility" and price it as a flat add-on. I charge 200 dollars per month per client. They're paying for the peace of mind, not the automation.

Jordan: Position it as proactive communication. "You'll never wonder what we're working on. Every Tuesday morning, you'll see exactly what shipped, what's in progress, and what's coming next week." That's the promise.

Jordan: For HubSpot users — same pattern, different API. GET request to `/crm/v3/objects/tickets` with a date filter for the last seven days. Pull your conversations from `/conversations/v3/conversations/threads`. Feed that into the same LLM prompt, same Notion output. The only difference is the source.

Jordan: If you're on Zapier instead of Make — use the Digest app to accumulate events through the week, then fire the summary on schedule. The Gmail integration works the same way. Notion connection is identical. Just costs about three times more in operations.

Jordan: For n8n people — IMAP trigger instead of Gmail module. OpenAI node with the new Responses API. Notion node for page creation. The workflow looks almost identical, just different node names.

Jordan: Let me address the elephant in the room. "But Jordan, what if the AI makes stuff up?" Fair question. That's why every bullet links to a source. That's why we use structured prompts with strict schemas. That's why we cap the output length. And honestly? If the LLM occasionally says you fixed six webhooks instead of five, your client isn't going to notice or care. They just want to know you're working.

Jordan: The counterargument I hear is that clients don't read these. AgencyAnalytics says passive reports have a twelve percent open rate. But here's the thing — this isn't a passive report. It's five bullets they can scan in thirty seconds, delivered where they already are, with zero effort required from them. My open rates average 67 percent. Click-through to Notion is about 40 percent.

Jordan: More importantly, it changes the conversation. Instead of "what are we paying you for?" you get "saw the digest — can we prioritize that inventory sync next week?" They're engaging with the work because they can see the work.

Jordan: Actually, let me show you something. This is a client who almost churned in January. Hadn't heard from them in six weeks, renewal coming up, classic danger signs. Started sending the weekly digest. By week three, they renewed for another year and increased scope by 40 percent. The only thing that changed was visibility.

Jordan: This is where most people get stuck, so let's slow down. The emotional barrier here isn't technical. It's the fear that if clients see how little time things actually take, they'll want to pay less. But that's backwards. They're not paying for your time. They're paying for outcomes. And this digest is proof of outcomes, not time.

Jordan: One client told me — and this is a direct quote — "I forward your digest to my boss every week to justify your retainer." That's the real value. You're not just doing the work. You're giving them ammunition to defend the spend internally.

Jordan: Quick sidebar on the PDF option because people ask. Only three of my twelve clients use it. They're all enterprise clients who need "official documentation" for their procurement teams. For everyone else, the Notion link is enough. Don't build features nobody uses.

Jordan: The Slack integration though? Game changer. Posts to their shared channel, their whole team sees it, creates this ambient awareness of progress. One client said it's like having a virtual standup without the meeting. That's exactly what we want.

Jordan: Here's a mistake I made early on — sending different formats to different stakeholders. The CEO got a summary, the tech lead got details, the project manager got task lists. Nightmare to maintain. Now everyone gets the same digest. If they want more detail, they click through to Notion.

Jordan: Let me show you the actual Make scenario JSON. I'm dropping this in the show notes. Import it, add your connections, change the client label, and you're running in literally ten minutes. No excuse not to ship this today.

Jordan: For cost optimization — if you're running more than twenty clients, consider batching. Instead of twenty separate scenarios, one scenario with a client loop. Reduces Make operations by about 70 percent. Same output, lower platform costs.

Jordan: The prompt engineering matters more than people think. Don't just say "summarize these emails." Be specific about voice, length, focus. I spent two hours tuning my prompt to sound like me writing a quick update, not ChatGPT writing a report.

Jordan: "Write like a technical consultant giving a quick status update. No corporate speak. No 'leveraging synergies.' Just clear, specific outcomes. If something saved time, say how much. If something's blocked, say why."

Jordan: Test this with your messiest client first. The one who sends seventeen emails a day, changes requirements mid-build, has three different stakeholders. If the digest works for them, it'll work for everyone.

Jordan: So here's what we actually built today. A zero-touch system that takes your invisible work — all those background automations, all those silent fixes — and makes it visible without adding any manual work to your week. Your clients see value, you prevent churn, and the whole thing runs for about twelve cents per client per week.

Jordan: The real unlock here isn't the automation itself. It's what happens when clients can finally see what you're doing. That conversation shift from "what are we paying for?" to "can we add this to next week?" — that's worth way more than the two hundred dollar monthly add-on.

Jordan: Your action item is simple. Pick one client — ideally one whose renewal is coming up in the next quarter. Build this digest for them. Run it manually once to test, then let it run weekly. Watch what happens to the relationship when they start seeing the work.

Jordan: Everything we built today — the Make scenario JSON, the prompt templates, the Notion database structure, even a cost calculator spreadsheet — it's all in the show notes. Import it, connect your accounts, and you're shipping in under ninety minutes.

Jordan: Next week we're tackling another solo operator nightmare — client onboarding that actually converts trial users without you touching anything. We'll build a fourteen-day nurture sequence that adapts based on usage data. No manual check-ins, no "just checking in" emails, just smart automation that closes deals while you sleep.

Jordan: I'm Jordan Chen. This is Headcount Zero. Keep shipping.

client reporting automationMake.comNotion client portalGmail automationclient churn preventionsolo consultant toolsautomation visibilityLLM summarizationZapier alternativesclient proof problem