---
title: "Automate a Weekly Status Report With AI"
description: "Automate a weekly product status report with AI: a Zapier schedule trigger, Jira or Sheets inputs, a fixed-format AI step, a human approval, then Slack."
canonical_url: "https://builderscamp.com/guides/tools/automate-weekly-status-report-with-ai"
date_published: "2026-11-09"
date_modified: "2026-11-09"
author: "Nuno Vilaça"
publisher: "Builders Camp"
guide_class: "tools"
---

# How to automate a weekly product status report with AI

**TL;DR:** Automating a weekly product status report takes five steps in Zapier: a weekly schedule trigger, a search step that pulls the week's Jira issues or sheet rows, an AI step with fixed output fields, a human approval that can stop the run, and a Slack or email send. The approval step is the one to keep, because the model drafts the report and you still decide what it says.

A weekly status report automation is a five-step Zap: a schedule fires, a search pulls the week's inputs, an AI step fills a fixed template, a person approves or edits it, and the result posts to Slack or email. The build needs a paid Zapier plan, because Zapier's [Free plan page](https://help.zapier.com/hc/en-us/articles/32337438839565-What-s-included-in-Zapier-s-Free-plan) limits free accounts to "Two-step Zap workflows" with one trigger and one action, and Zapier's [AI by Zapier docs](https://help.zapier.com/hc/en-us/articles/8496342944013-Create-AI-powered-actions-with-AI-by-Zapier) say: "On a Free plan, you cannot add tools or publish a Zap with an AI by Zapier step".

This page covers the scheduled machine around the report. Writing the report well is a separate skill, and the guide to [AI for stakeholder updates](https://builderscamp.com/guides/tools/ai-for-stakeholder-updates) covers the three edits a model will not make for you: cutting the workstream that did not move, naming the decision you need, and putting the bad news first. The workflow below is built so you still make those edits, every week, in the approval step.

Every Zapier and Make step, setting and limit on this page was checked against the vendors' own help pages and pricing page on 9 November 2026.

## What does the finished workflow look like?

The finished workflow is five steps, and each one maps to a built-in Zapier tool or a standard app action. The table shows the build for a team that tracks work in Jira and keeps metrics in a Google Sheet.

| Step | Zapier tool or action | What it does | What breaks it |
|---|---|---|---|
| 1. Trigger | Schedule by Zapier, Every Week | Starts the run on a set day and time | Wrong account timezone |
| 2. Inputs | Jira: Find Issues by Filter; Google Sheets: Lookup Spreadsheet Rows (Advanced) | Pulls the week's issues, metrics and last week's report | A filter nobody maintains |
| 3. Draft | AI by Zapier with output fields | Writes the report into fixed sections | No output fields, so free text |
| 4. Review | Human in the Loop: Request Approval | Pauses the run until a person approves or edits | Timeout set to continue |
| 5. Send and store | Slack: Send Channel Message; Google Sheets: Create Spreadsheet Row | Posts the approved text and saves a copy | Mapping the draft instead of the edit |

Step 5 has two jobs on purpose. The copy saved to the sheet becomes next week's input, which is how the report learns what changed.

## How do you set the weekly schedule trigger?

Schedule by Zapier is the trigger, set to Every Week with a day and a time. Zapier's [Schedule help page](https://help.zapier.com/hc/en-us/articles/8496288648461-Schedule-Zaps-to-run-at-specific-intervals) says the tool lets a Zap run every month, week, day or hour, that the weekly option asks for a Day of the Week and a Time of Day, and that "Using this tool does not count towards your task usage".

Two limits on that same [Schedule page](https://help.zapier.com/hc/en-us/articles/8496288648461-Schedule-Zaps-to-run-at-specific-intervals) shape when you schedule it. Zapier writes that it "does not guarantee your Zap will trigger at the precise minute you set" and that a Zap "should run within a few minutes of your scheduled time". The page also says schedule triggers use the timezone set in your Zapier account rather than the one set in the Zap, and that Zap history shows run timestamps in UTC.

So schedule the draft, not the send. Fire the trigger a few hours before stakeholders expect the report, for example Friday morning for a Friday afternoon post, which leaves time for the review step to wait for you. If the report lands a few minutes late, nobody notices. If it lands before you have read it, everybody does.

## Where should the report pull its inputs from?

Pull inputs from the systems that already hold the truth, with search actions placed right after the trigger. Zapier's [Jira Software Cloud integration page](https://zapier.com/apps/jira-software-cloud/integrations) lists a Find Issues by Filter action that "Executes a saved Jira filter and returns the filtered issues for automated reporting and monitoring workflows", alongside Find Issues (Via JQL), which searches for issues using JQL.

Use a saved filter rather than JQL typed into the Zap. The filter lives in Jira where the team can see and fix it, and when someone renames a status, one person updates one filter instead of hunting through Zap settings. A filter for issues in this project updated in the past week is the usual starting point; add a second filter for anything flagged blocked.

For metrics and notes, a Google Sheet works well. Zapier's [Google Sheets integration page](https://zapier.com/apps/google-sheets/integrations) lists Get Many Spreadsheet Rows (Advanced), which can "Return up to 1,500 rows as a single JSON value or as line items", and Lookup Spreadsheet Rows (Advanced), which can "Find up to 500 rows based on a column and value as line items". Keep one tab for the week's numbers and a notes column you fill during the week, and a second tab where step 5 saves every approved report.

That second tab matters because the AI step has no memory. Zapier's [AI by Zapier docs](https://help.zapier.com/hc/en-us/articles/8496342944013-Create-AI-powered-actions-with-AI-by-Zapier) say "AI by Zapier does not learn from previous actions", and that each time the action runs, "the AI knowledge is reset to its initial state". Pull last week's approved report from the sheet on every run and hand it to the model as context.

## How do you make the AI step return the same structure every week?

Define output fields on the AI by Zapier step, one per section of the report. Zapier's [AI by Zapier docs](https://help.zapier.com/hc/en-us/articles/8496342944013-Create-AI-powered-actions-with-AI-by-Zapier) say you can map output fields to later steps in the Zap, mark a field as required with "Is this output field required?", and warn that "If you do not add any output fields, AI by Zapier will return one combined output as the result".

A combined output is the format drift problem: one week a bulleted list, the next week three paragraphs, the week after a list with a new heading. Fixed fields stop that. A field set that works for most product teams:

- **headline**: one sentence on what changed since last week, required
- **shipped**, **slipped_or_at_risk** and **next_week**: short lists, each item tied to an issue key
- **decision_needed** and **decision_by**: the one call you need from a reader, and the date, required (allow "none")
- **inferred**: anything the model guessed rather than found in the inputs

The prompt then only has to describe the job. Type `/` in the prompt field to insert data from earlier steps, as the docs describe, and keep the instructions short:

```text
You write a weekly product status report for [team]. Use only the Jira issues,
metrics rows and notes below. Last week's approved report is included: report
only what changed since it. Tie every shipped or slipped item to its issue key.
If you are unsure whether something happened, put it in "inferred", not in the
main sections. If no decision is needed, write "none" in decision_needed.
```

The model tier decides what each run costs. The same [AI by Zapier docs](https://help.zapier.com/hc/en-us/articles/8496342944013-Create-AI-powered-actions-with-AI-by-Zapier) list Standard at 1x tasks, Advanced at 3x and Premium at 5x, and note that Premium "is the default selection for new steps". Summarising a week of tickets into fixed fields is a simple job, so switch to Standard first and move up a tier only if the draft keeps missing items.

## Where does the human review step go, and what happens if nobody approves?

The review step goes between the AI draft and the send, and it should be able to stop the run. Zapier's [Human in the Loop page](https://help.zapier.com/hc/en-us/articles/38733184458765) describes the tool as a way to "pause a Zap at a certain step so a human can review before your workflow continues", and lists "Reviewing AI-generated content" first among its uses.

The Request Approval action holds the settings that matter. Zapier's [Request Approval help page](https://help.zapier.com/hc/en-us/articles/38731463206029) lets you map content fields for review, choose whether the reviewer can edit, choose Stop run if the reviewer declines, and pick an On timeout behaviour of Skip and continue or End run. Notifications can go to email, Slack, or another Zap, and a reminder can be sent after a set wait.

Set it up like this:

1. Map each AI output field into Content to review, so you see the report section by section.
2. Set Let Reviewer edit content to Yes. Editing is the point of the step.
3. Set the decline action to Stop run and the timeout to End run. A report nobody read should not go out.
4. Send the notification to Slack and turn on a reminder.

Two details trip people up. Edits come back as new fields named "Edited Content" plus the field name, so the send step has to map those, not the original AI fields, or your corrections never reach the channel. And the [Request Approval help page](https://help.zapier.com/hc/en-us/articles/38731463206029) lists, as a prerequisite, "A paid Zapier account. Human in the Loop is a premium app on Zapier", and notes that on a Pro account you can only send requests to yourself, while on a Team or Enterprise account the reviewer must be logged in as one of the Zapier users you select.

## How do you send the approved report to Slack or email?

The send step is a normal app action that maps the approved, edited fields into a message. Zapier's [Slack integration page](https://zapier.com/apps/slack/integrations) lists Send Channel Message, which can "Post a message or thread reply to a channel", and the [Gmail integration](https://zapier.com/apps/gmail/integrations) offers Send Email for readers who want the report in their inbox.

Build the message from the fields in a fixed order: headline, decision needed with its date, slipped or at risk, shipped, next week. Decision and bad news sit above the good news on purpose, for the reasons the [stakeholder updates guide](https://builderscamp.com/guides/tools/ai-for-stakeholder-updates) sets out. Leave the inferred field out of the message; it exists for you, in the review step.

Then add the final action: Google Sheets Create Spreadsheet Row, writing the approved text and the date to the history tab. Next week's run reads that row as last week's report.

## How much does running the workflow cost?

The workflow needs a paid Zapier plan, and each step that runs uses tasks. Zapier's [pricing page](https://zapier.com/pricing) lists the Free plan with two-step Zaps and 100 tasks a month, multi-step Zaps from the Professional plan, and Professional "Starting from $19.99 /month" when billed yearly, which the page shows by default. The same page says AI steps follow the same task-based pricing as other steps and draw from one shared task allocation.

Those prices and limits are as of 9 November 2026 and change often, so check the pricing page before you budget.

As a rough illustration, not a quote: if the Jira search, the sheet lookup, a Standard-tier AI step, the approval, the Slack post and the sheet write use about one task each, a run costs around six tasks, or roughly 25 to 30 a month at one run a week. The schedule trigger adds nothing, and Zapier's Request Approval page says only successful Human in the Loop actions count toward task usage. Moving the AI step to Premium multiplies that step by five.

## Can you build the same workflow in Make?

Make can run the same pattern as a scheduled scenario. Make's [schedule help page](https://help.make.com/schedule-a-scenario) lists Weekly among the schedule options, says a scenario runs every 15 minutes by default until you change it, and notes that the minimum interval depends on your plan. Make's [credits page](https://help.make.com/credits) says that by default 1 operation equals 1 credit, and that on paid plans, with a custom AI provider connection such as OpenAI or Anthropic Claude, you pay Make credits for operations and pay the provider for tokens.

The difference that matters for this build is the review step. Zapier has a built-in approval action; in Make, design the review before you build the rest, for example by posting the draft to a private Slack channel and having a second scenario send the version you approve. The [Zapier vs Make comparison](https://builderscamp.com/guides/tools/zapier-vs-make-for-product-managers) covers the wider trade-offs between the two tools.

## When should you not automate the status report?

Skip the automation when the report is mostly judgment and the inputs are thin. If your week's real news lives in conversations rather than in Jira or a sheet, the workflow will faithfully summarise the tickets and miss the story, and you will rewrite it in the review step every Friday. You have then automated the easy half of the job and kept the hard half, plus a Zap to maintain.

Dirty inputs are the other reason. A model asked to report on a board where half the tickets are stale reports the stale tickets with confidence. Fix the board first; [which tasks to automate](https://builderscamp.com/guides/tools/which-tasks-to-automate) has a test for whether a task is stable enough to hand over, and the [workflow automation best practices](https://builderscamp.com/guides/tools/workflow-automation-best-practices) guide covers error handling once it runs.

The automation pays off when the report is regular, the inputs are structured, and the drafting is the part that eats your time. For most product teams with a maintained board, that is true.

## Where can you learn to build reliable AI workflows like this?

Builders Camp's Automate Workflows with AI bootcamp covers identifying repetitive product work, redesigning it into a workflow and automating it with modern AI tools, and its syllabus names Zapier, Make.com and GPT. It runs 1 week with 2 live sessions and 8 self-paced microlessons, is rated intermediate, and is led by its own programme director, a working practitioner. Its curriculum covers triggers, actions and error handling, structured outputs and guardrails for AI steps, human-in-the-loop approvals for high-stakes steps, and monitoring. The practical challenge hands you a [broken Make.com support-ticket automation](https://builderscamp.com/guides/challenges/automate-workflows-with-ai-broken-support-automation) to diagnose and fix under a 200-token prompt limit.

[See the Automate Workflows with AI bootcamp](https://builderscamp.com/bootcamps/automate-workflows-with-ai?utm_source=guide&utm_medium=organic&utm_campaign=automate-weekly-status-report-with-ai)

Before you build any of it, run the AI step by hand once: paste last week's Jira export and notes into a chat with the same prompt and fields, and compare the draft with the report you actually sent. Every gap between the two is a field, an input or an instruction the Zap needs before its first Friday.

## Frequently asked questions

### Can I automate a weekly status report on Zapier's Free plan?

Not the full workflow. Zapier's help center says the Free plan allows two-step Zap workflows, one trigger and one action, and that on a Free plan you cannot publish a Zap with an AI by Zapier step. A scheduled report with an input, an AI step, an approval and a send step needs a paid plan.

### Will the report go out at exactly the time I schedule?

No. Zapier's Schedule help page says it does not guarantee the Zap will trigger at the precise minute you set and that it should run within a few minutes of the scheduled time. The schedule uses the timezone in your Zapier account settings, and Zap history shows run times in UTC.

### How do I stop the AI from changing the report format every week?

Define output fields on the AI by Zapier step, one per section of the report, and mark the ones that must always be filled as required. Zapier's docs say that without output fields the step returns one combined output, which is the free text that drifts from week to week.

### What happens if nobody approves the draft?

You choose. Zapier's Human in the Loop Request Approval action has an On timeout setting: Skip and continue sends the run on without a response, and End run stops it. For a status report, pick End run, so an unreviewed draft never reaches stakeholders.

### Does the AI remember last week's report?

No. Zapier's docs say AI by Zapier does not learn from previous actions and resets to its initial state on every run. Store each approved report in a Google Sheet and pull the latest row into the next run, so the model can report what changed since last week.

### Which AI model tier should a status report use?

Zapier lists three tiers for AI by Zapier: Standard at 1x tasks, Advanced at 3x and Premium at 5x, with Premium selected by default for new steps. Summarising a week of tickets into fixed fields is a simple task, so test Standard first and move up only if the output misses things.

### Can I build the same workflow in Make instead of Zapier?

Yes. Make's help center lists a Weekly option in a scenario's schedule settings, and its credits page says that by default 1 operation equals 1 credit. On paid Make plans, AI modules that use your own provider connection, such as OpenAI or Anthropic Claude, bill credits per operation while the provider bills you for tokens.

### How is this different from writing the update with ChatGPT?

Writing with a chat assistant still means you collect the inputs and start the draft every week. A scheduled workflow does the collecting and drafting on its own, then waits for you. The editing judgment stays the same, and the guide to AI for stakeholder updates covers those edits.

## Sources

- [Zapier Help: Schedule Zap workflows to run at specific intervals](https://help.zapier.com/hc/en-us/articles/8496288648461-Schedule-Zaps-to-run-at-specific-intervals)
- [Zapier Help: Use AI by Zapier to analyze and return data](https://help.zapier.com/hc/en-us/articles/8496342944013-Create-AI-powered-actions-with-AI-by-Zapier)
- [Zapier Help: Request approval to keep your workflow running with Human in the Loop](https://help.zapier.com/hc/en-us/articles/38731463206029)
- [Zapier Help: Use Human in the Loop to pause Zap workflows pending human review](https://help.zapier.com/hc/en-us/articles/38733184458765)
- [Zapier Help: What's included in Zapier's Free plan?](https://help.zapier.com/hc/en-us/articles/32337438839565-What-s-included-in-Zapier-s-Free-plan)
- [Zapier: Plans and Pricing](https://zapier.com/pricing)
- [Zapier: Jira Software Cloud integrations](https://zapier.com/apps/jira-software-cloud/integrations)
- [Zapier: Google Sheets integrations](https://zapier.com/apps/google-sheets/integrations)
- [Zapier: Slack integrations](https://zapier.com/apps/slack/integrations)
- [Zapier: Gmail integrations](https://zapier.com/apps/gmail/integrations)
- [Make Help Center: Schedule a scenario](https://help.make.com/schedule-a-scenario)
- [Make Help Center: Credits](https://help.make.com/credits)

## How this guide was made

Researched from Builders Camp's bootcamp, track and masterclass material and the sources listed on this page, drafted with AI, and fact-checked against every source cited.