mirror of
https://github.com/khoaliber/n8nworkflows.xyz.git
synced 2026-04-28 00:29:22 +00:00
creation
This commit is contained in:
nusquama
+388
@@ -0,0 +1,388 @@
|
||||
Monitor daily HR risks and standup summaries with Monday.com and GPT-4o-mini
|
||||
|
||||
https://n8nworkflows.xyz/workflows/monitor-daily-hr-risks-and-standup-summaries-with-monday-com-and-gpt-4o-mini-12269
|
||||
|
||||
|
||||
# Monitor daily HR risks and standup summaries with Monday.com and GPT-4o-mini
|
||||
|
||||
## 1. Workflow Overview
|
||||
|
||||
**Purpose:**
|
||||
This workflow runs every morning to monitor HR tasks stored in a Monday.com board, detect operational risks (overdue, stuck, unassigned tasks), and automatically email either:
|
||||
- an **escalation risk report** (if risks exist), or
|
||||
- a **normal daily standup summary** (if everything is on track).
|
||||
|
||||
**Target use cases:**
|
||||
- Daily HR operations pulse-check (task progress, blockers, ownership gaps)
|
||||
- Leadership-ready escalation when HR delivery is at risk
|
||||
- Automated standup-style summary without manual reporting
|
||||
|
||||
### 1.1 Trigger & Data Retrieval
|
||||
Runs on a schedule and fetches all HR items from a Monday.com board/group.
|
||||
|
||||
### 1.2 Task Filtering & Normalization
|
||||
Removes completed tasks (“Done”), and reshapes Monday.com items into a simplified task list.
|
||||
|
||||
### 1.3 Risk Detection & Metrics
|
||||
Computes counts and detailed lists for:
|
||||
- overdue tasks (more than 2 days late),
|
||||
- “Stuck” tasks,
|
||||
- unassigned tasks (no owner).
|
||||
|
||||
### 1.4 Decision + AI Content Generation + Email Delivery
|
||||
If any risk exists → generate AI risk report and email escalation.
|
||||
Else → generate AI daily standup summary and email it.
|
||||
|
||||
---
|
||||
|
||||
## 2. Block-by-Block Analysis
|
||||
|
||||
### Block 1 — Trigger & Data Source
|
||||
**Overview:** Runs daily at a fixed time and loads HR tasks from Monday.com for analysis.
|
||||
**Nodes involved:** `Daily Trigger`, `Get HR Tasks`
|
||||
|
||||
#### Node: Daily Trigger
|
||||
- **Type / role:** Schedule Trigger (`n8n-nodes-base.scheduleTrigger`) — entry point.
|
||||
- **Configuration (interpreted):**
|
||||
- Runs every day at **08:30** (server / n8n instance timezone).
|
||||
- **Inputs / outputs:**
|
||||
- **Input:** none (trigger)
|
||||
- **Output:** to `Get HR Tasks`
|
||||
- **Version notes:** typeVersion **1.3**.
|
||||
- **Potential failures / edge cases:**
|
||||
- Timezone mismatch (instance timezone vs expected local HR timezone).
|
||||
- Workflow is currently **inactive** (`active: false`), so it will not run until activated.
|
||||
|
||||
#### Node: Get HR Tasks
|
||||
- **Type / role:** Monday.com node (`n8n-nodes-base.mondayCom`) — reads board items.
|
||||
- **Configuration (interpreted):**
|
||||
- **Resource:** Board Item
|
||||
- **Operation:** Get All
|
||||
- **Board ID:** `5025720609`
|
||||
- **Group ID:** `new_group29179`
|
||||
- **Return all:** enabled
|
||||
- Uses Monday.com API credential: **“Monday.com account 3”**
|
||||
- **Inputs / outputs:**
|
||||
- **Input:** `Daily Trigger`
|
||||
- **Output:** list of Monday.com items → `Filter Active Tasks`
|
||||
- **Version notes:** typeVersion **1**.
|
||||
- **Potential failures / edge cases:**
|
||||
- Invalid board/group IDs, or group renamed/removed.
|
||||
- Monday API permission errors (token scope/user access).
|
||||
- Large boards: returning all items can be slow or hit API rate limits.
|
||||
- Column ordering is assumed later (owner uses `column_values[1]`), which can break if board columns change.
|
||||
|
||||
**Sticky note coverage (for this block):**
|
||||
- “### Trigger and Data Source
|
||||
This workflow runs daily and fetches active HR tasks from Monday.com.”
|
||||
|
||||
---
|
||||
|
||||
### Block 2 — Task Filtering & Preparation
|
||||
**Overview:** Filters out completed tasks and converts the Monday.com payload into a compact array of task objects used downstream.
|
||||
**Nodes involved:** `Filter Active Tasks`, `Transform Tasks`
|
||||
|
||||
#### Node: Filter Active Tasks
|
||||
- **Type / role:** Code node (`n8n-nodes-base.code`) — removes tasks with status “Done”.
|
||||
- **Configuration (interpreted):**
|
||||
- JavaScript filters incoming items:
|
||||
- Reads `status` from `item.json.column_values` where `c.type === 'status'`
|
||||
- Keeps items where status exists and is **not** `"Done"`
|
||||
- **Key logic (variables/expressions):**
|
||||
- `item.json.column_values.find(c => c.type === 'status')?.text`
|
||||
- **Inputs / outputs:**
|
||||
- **Input:** `Get HR Tasks`
|
||||
- **Output:** filtered items → `Transform Tasks`
|
||||
- **Version notes:** typeVersion **2** (modern Code node behavior).
|
||||
- **Potential failures / edge cases:**
|
||||
- If Monday response structure changes or `column_values` missing → could throw if not an array (current code assumes it is).
|
||||
- If “Done” status label is different (e.g., “Completed”) → tasks won’t be filtered correctly.
|
||||
- If multiple status columns exist, `find(c.type==='status')` may select an unexpected one.
|
||||
|
||||
#### Node: Transform Tasks
|
||||
- **Type / role:** Code node (`n8n-nodes-base.code`) — normalizes items into `{ tasks: [...] }`.
|
||||
- **Configuration (interpreted):**
|
||||
- Outputs a **single item** with a `tasks` array derived from all input items.
|
||||
- Each task object includes:
|
||||
- `name`: item name
|
||||
- `owner`: `column_values[1]?.text` (second column in array)
|
||||
- `status`: first column with `type === 'status'` → `.text`
|
||||
- `due_date`: first column with `type === 'date'` → `.text`
|
||||
- **Key logic (variables/expressions):**
|
||||
- Owner extraction is **positional**: `i.json.column_values[1]?.text || null`
|
||||
- Status/date extraction is by type: `find(c => c.type === 'status'/'date')?.text`
|
||||
- **Inputs / outputs:**
|
||||
- **Input:** `Filter Active Tasks`
|
||||
- **Output:** one aggregated item → `Build HR Metrics`
|
||||
- **Version notes:** typeVersion **2**.
|
||||
- **Potential failures / edge cases:**
|
||||
- **Column position dependency** for owner is brittle; if board columns reorder, owner may become wrong/null.
|
||||
- Date parsing later depends on the format in `due_date` text; Monday date text may be locale-formatted.
|
||||
|
||||
**Sticky note coverage (for this block):**
|
||||
- “### Task filtering and preparation
|
||||
Filters out completed tasks and prepares HR task data for analysis.”
|
||||
|
||||
---
|
||||
|
||||
### Block 3 — Risk Detection & Metrics
|
||||
**Overview:** Computes risk metrics and detailed risk lists (overdue >2 days, stuck, unassigned). Produces a single structured object used for branching and AI prompts.
|
||||
**Nodes involved:** `Build HR Metrics`, `Any HR Risks?`
|
||||
|
||||
#### Node: Build HR Metrics
|
||||
- **Type / role:** Code node (`n8n-nodes-base.code`) — calculates risk counts and arrays.
|
||||
- **Configuration (interpreted):**
|
||||
- Reads `items[0].json.tasks` from previous node (expects exactly one aggregated item).
|
||||
- Builds:
|
||||
- `unassigned`: tasks where `owner` is null/empty
|
||||
- `stuck`: tasks where `status === 'Stuck'`
|
||||
- `overdue`: tasks where due date exists and is **more than 2 days in the past**
|
||||
- Adds `days_overdue` to overdue entries.
|
||||
- **Key logic (variables/expressions):**
|
||||
- `const days = Math.floor((today - due) / 86400000); if (days > 2) ...`
|
||||
- Output shape:
|
||||
- `json.tasks`
|
||||
- `json.metrics.{ overdue_count, stuck_count, unassigned_count, overdue, stuck, unassigned }`
|
||||
- **Inputs / outputs:**
|
||||
- **Input:** `Transform Tasks`
|
||||
- **Output:** → `Any HR Risks?`
|
||||
- **Version notes:** typeVersion **2**.
|
||||
- **Potential failures / edge cases:**
|
||||
- **Date parsing risk:** `new Date(t.due_date)` may return Invalid Date if Monday’s `.text` is not ISO-compatible (common in non-US locales).
|
||||
- Timezone effects: “days overdue” can be off by 1 depending on time of day/timezone.
|
||||
- Assumes `items[0]` exists; if no tasks returned at all, it will throw.
|
||||
|
||||
#### Node: Any HR Risks?
|
||||
- **Type / role:** IF node (`n8n-nodes-base.if`) — branches based on combined risk count.
|
||||
- **Configuration (interpreted):**
|
||||
- Condition: `overdue_count + stuck_count + unassigned_count > 0`
|
||||
- Uses strict type validation.
|
||||
- **Key expressions:**
|
||||
- `{{ $json.metrics.overdue_count + $json.metrics.stuck_count + $json.metrics.unassigned_count }}`
|
||||
- **Inputs / outputs:**
|
||||
- **Input:** `Build HR Metrics`
|
||||
- **True output (risk exists):** → `AI Risk Report`
|
||||
- **False output (no risk):** → `AI Daily Summary`
|
||||
- **Version notes:** typeVersion **2**.
|
||||
- **Potential failures / edge cases:**
|
||||
- If `metrics` missing (due to upstream error/shape change), expression evaluation fails.
|
||||
- If counts are strings (unlikely here) strict validation could fail comparisons.
|
||||
|
||||
**Sticky note coverage (for this block):**
|
||||
- “### Risk Detection and Metrics
|
||||
Checks for overdue, stuck, and unassigned tasks and builds HR risk metrics.”
|
||||
|
||||
---
|
||||
|
||||
### Block 4 — AI Analysis + Email Delivery
|
||||
**Overview:** Uses OpenAI (GPT-4o-mini) to generate either a risk escalation report or a daily standup summary, then emails the resulting text via Gmail.
|
||||
**Nodes involved:** `AI Risk Report`, `Escalation Email`, `AI Daily Summary`, `Daily HR Summary Email`
|
||||
|
||||
#### Node: AI Risk Report
|
||||
- **Type / role:** OpenAI (LangChain) node (`@n8n/n8n-nodes-langchain.openAi`) — generates escalation report text.
|
||||
- **Configuration (interpreted):**
|
||||
- Model: **gpt-4o-mini** (selected via model list resolver)
|
||||
- System message: “You are an HR operations analyst.”
|
||||
- User prompt requests “HR Risk Report” with:
|
||||
1) Critical blockers
|
||||
2) Overdue tasks with days overdue
|
||||
3) Unassigned tasks
|
||||
4) Recommended HR actions
|
||||
- Injects data: `JSON.stringify($json.metrics, null, 2)`
|
||||
- Credentials: **“OpenAi account 3”**
|
||||
- **Inputs / outputs:**
|
||||
- **Input:** True branch from `Any HR Risks?`
|
||||
- **Output:** to `Escalation Email`
|
||||
- **Version notes:** typeVersion **2.1**.
|
||||
- **Potential failures / edge cases:**
|
||||
- OpenAI auth/quota/model access errors.
|
||||
- Prompt may exceed token limits if metrics lists are large (many tasks).
|
||||
- Output mapping assumes a specific response structure used later (`$json.output[0].content[0].text`).
|
||||
|
||||
#### Node: Escalation Email
|
||||
- **Type / role:** Gmail node (`n8n-nodes-base.gmail`) — sends escalation email.
|
||||
- **Configuration (interpreted):**
|
||||
- To: `user@example.com`
|
||||
- Subject: `🚨 HR Risks Detected – Immediate Action Required`
|
||||
- Body (text): `{{ $json.output[0].content[0].text }}`
|
||||
- Credentials: **“Gmail credentials”** (OAuth2)
|
||||
- **Inputs / outputs:**
|
||||
- **Input:** `AI Risk Report`
|
||||
- **Output:** none (terminal)
|
||||
- **Version notes:** typeVersion **2.2**.
|
||||
- **Potential failures / edge cases:**
|
||||
- Gmail OAuth token expired/insufficient scopes.
|
||||
- If OpenAI node output schema differs, body expression may be undefined and send blank email or fail.
|
||||
- Deliverability/recipient restrictions in Google Workspace environments.
|
||||
|
||||
#### Node: AI Daily Summary
|
||||
- **Type / role:** OpenAI (LangChain) node (`@n8n/n8n-nodes-langchain.openAi`) — generates normal daily summary.
|
||||
- **Configuration (interpreted):**
|
||||
- Model: `gpt-4o-mini`
|
||||
- System message: “You are an HR operations assistant.”
|
||||
- Prompt: “Create a concise HR daily standup summary grouped by status.”
|
||||
- Injects tasks: `JSON.stringify($json.tasks, null, 2)`
|
||||
- Credentials: **“OpenAi account 3”**
|
||||
- **Inputs / outputs:**
|
||||
- **Input:** False branch from `Any HR Risks?`
|
||||
- **Output:** to `Daily HR Summary Email`
|
||||
- **Version notes:** typeVersion **2.1**.
|
||||
- **Potential failures / edge cases:**
|
||||
- Token/size issues if the tasks array is very large.
|
||||
- Same response-structure assumption as above for the email body mapping.
|
||||
|
||||
#### Node: Daily HR Summary Email
|
||||
- **Type / role:** Gmail node (`n8n-nodes-base.gmail`) — sends daily summary email.
|
||||
- **Configuration (interpreted):**
|
||||
- To: `user@example.com`
|
||||
- Subject: `Daily HR Action Summary`
|
||||
- Body (text): `{{ $json.output[0].content[0].text }}`
|
||||
- Credentials: **“Gmail credentials”**
|
||||
- **Inputs / outputs:**
|
||||
- **Input:** `AI Daily Summary`
|
||||
- **Output:** none (terminal)
|
||||
- **Version notes:** typeVersion **2.2**.
|
||||
- **Potential failures / edge cases:**
|
||||
- Same Gmail OAuth and expression-shape risks as escalation email.
|
||||
|
||||
**Sticky note coverage (for this block):**
|
||||
- “### AI analysis and decision
|
||||
Uses AI to decide whether there are HR risks or a normal daily update. And
|
||||
sends escalation emails if risks exist or a daily summary if everything is on track.”
|
||||
|
||||
---
|
||||
|
||||
## 3. Summary Table
|
||||
|
||||
| Node Name | Node Type | Functional Role | Input Node(s) | Output Node(s) | Sticky Note |
|
||||
|---|---|---|---|---|---|
|
||||
| Daily Trigger | Schedule Trigger | Daily workflow entry point at 08:30 | — | Get HR Tasks | ### Trigger and Data Source<br>This workflow runs daily and fetches active HR tasks from Monday.com. |
|
||||
| Get HR Tasks | Monday.com | Fetch all items from HR board/group | Daily Trigger | Filter Active Tasks | ### Trigger and Data Source<br>This workflow runs daily and fetches active HR tasks from Monday.com. |
|
||||
| Filter Active Tasks | Code | Remove tasks with status “Done” | Get HR Tasks | Transform Tasks | ### Task filtering and preparation<br>Filters out completed tasks and prepares HR task data for analysis. |
|
||||
| Transform Tasks | Code | Normalize/aggregate items into `tasks[]` | Filter Active Tasks | Build HR Metrics | ### Task filtering and preparation<br>Filters out completed tasks and prepares HR task data for analysis. |
|
||||
| Build HR Metrics | Code | Compute overdue/stuck/unassigned metrics | Transform Tasks | Any HR Risks? | ### Risk Detection and Metrics<br>Checks for overdue, stuck, and unassigned tasks and builds HR risk metrics. |
|
||||
| Any HR Risks? | IF | Branch based on total risk count > 0 | Build HR Metrics | AI Risk Report (true), AI Daily Summary (false) | ### Risk Detection and Metrics<br>Checks for overdue, stuck, and unassigned tasks and builds HR risk metrics.<br>### AI analysis and decision<br>Uses AI to decide whether there are HR risks or a normal daily update. And sends escalation emails if risks exist or a daily summary if everything is on track. |
|
||||
| AI Risk Report | OpenAI (LangChain) | Generate escalation-focused HR risk report | Any HR Risks? (true) | Escalation Email | ### AI analysis and decision<br>Uses AI to decide whether there are HR risks or a normal daily update. And sends escalation emails if risks exist or a daily summary if everything is on track. |
|
||||
| Escalation Email | Gmail | Email escalation report to stakeholders | AI Risk Report | — | ### AI analysis and decision<br>Uses AI to decide whether there are HR risks or a normal daily update. And sends escalation emails if risks exist or a daily summary if everything is on track. |
|
||||
| AI Daily Summary | OpenAI (LangChain) | Generate normal daily standup summary | Any HR Risks? (false) | Daily HR Summary Email | ### AI analysis and decision<br>Uses AI to decide whether there are HR risks or a normal daily update. And sends escalation emails if risks exist or a daily summary if everything is on track. |
|
||||
| Daily HR Summary Email | Gmail | Email daily summary to stakeholders | AI Daily Summary | — | ### AI analysis and decision<br>Uses AI to decide whether there are HR risks or a normal daily update. And sends escalation emails if risks exist or a daily summary if everything is on track. |
|
||||
| Workflow overview | Sticky Note | Documentation / overview | — | — | ## Workflow Overview<br>…(see note in section 5) |
|
||||
| Sticky Note | Sticky Note | Visual grouping label | — | — | ### Trigger and Data Source<br>This workflow runs daily and fetches active HR tasks from Monday.com. |
|
||||
| Sticky Note1 | Sticky Note | Visual grouping label | — | — | ### Task filtering and preparation<br>Filters out completed tasks and prepares HR task data for analysis. |
|
||||
| Sticky Note2 | Sticky Note | Visual grouping label | — | — | ### Risk Detection and Metrics<br>Checks for overdue, stuck, and unassigned tasks and builds HR risk metrics. |
|
||||
| Sticky Note3 | Sticky Note | Visual grouping label | — | — | ### AI analysis and decision<br>Uses AI to decide whether there are HR risks or a normal daily update. And sends escalation emails if risks exist or a daily summary if everything is on track. |
|
||||
|
||||
---
|
||||
|
||||
## 4. Reproducing the Workflow from Scratch
|
||||
|
||||
1. **Create a new workflow**
|
||||
- Name it: **“Monitor daily HR actions, blockers, and risks using Monday.com and AI”** (or your preferred name).
|
||||
- Ensure workflow settings use **Execution Order: v1** (Workflow settings → Execution).
|
||||
|
||||
2. **Add trigger**
|
||||
- Add node: **Schedule Trigger**
|
||||
- Configure: run daily at **08:30**
|
||||
- Connect: `Daily Trigger` → `Get HR Tasks`
|
||||
|
||||
3. **Add Monday.com retrieval**
|
||||
- Add node: **Monday.com**
|
||||
- Resource: **Board Item**
|
||||
- Operation: **Get All**
|
||||
- Set **Board ID** to your HR board (example in workflow: `5025720609`)
|
||||
- Set **Group ID** to the target group (example: `new_group29179`)
|
||||
- Enable **Return All**
|
||||
- Credentials:
|
||||
- Create/select **Monday.com API** credentials (token/user with board access)
|
||||
- Connect to next: `Get HR Tasks` → `Filter Active Tasks`
|
||||
|
||||
4. **Filter out completed tasks**
|
||||
- Add node: **Code**
|
||||
- Paste logic equivalent to:
|
||||
- Find status column (`type === 'status'`) and exclude `"Done"`
|
||||
- Connect: `Filter Active Tasks` → `Transform Tasks`
|
||||
|
||||
5. **Transform into normalized `tasks[]`**
|
||||
- Add node: **Code**
|
||||
- Configure it to output **one item** with `json.tasks = items.map(...)`
|
||||
- Map at minimum:
|
||||
- `name` from item name
|
||||
- `owner` from the appropriate owner/person column
|
||||
- `status` from status column text
|
||||
- `due_date` from date column text
|
||||
- Important: avoid positional mapping if possible; prefer selecting the correct column by `id` or `type` to reduce breakage when board columns move.
|
||||
- Connect: `Transform Tasks` → `Build HR Metrics`
|
||||
|
||||
6. **Build metrics (overdue/stuck/unassigned)**
|
||||
- Add node: **Code**
|
||||
- Compute:
|
||||
- `unassigned` where owner is empty
|
||||
- `stuck` where status equals `"Stuck"`
|
||||
- `overdue` where due date is more than **2 days** behind “today”
|
||||
- Output:
|
||||
- `tasks` (pass-through)
|
||||
- `metrics` object with counts and lists
|
||||
- Connect: `Build HR Metrics` → `Any HR Risks?`
|
||||
|
||||
7. **Add risk decision branch**
|
||||
- Add node: **IF**
|
||||
- Condition: Number comparison
|
||||
- Left value expression: `metrics.overdue_count + metrics.stuck_count + metrics.unassigned_count`
|
||||
- Operation: `>`
|
||||
- Right value: `0`
|
||||
- Connect:
|
||||
- **True** → `AI Risk Report`
|
||||
- **False** → `AI Daily Summary`
|
||||
|
||||
8. **Add OpenAI node for risk report**
|
||||
- Add node: **OpenAI (LangChain)**
|
||||
- Model: **gpt-4o-mini**
|
||||
- Messages:
|
||||
- System: “You are an HR operations analyst.”
|
||||
- User: request a risk report and inject `metrics` as JSON
|
||||
- Credentials:
|
||||
- Create/select **OpenAI API** credentials
|
||||
- Connect: `AI Risk Report` → `Escalation Email`
|
||||
|
||||
9. **Add Gmail node for escalation email**
|
||||
- Add node: **Gmail → Send**
|
||||
- To: escalation recipient(s)
|
||||
- Subject: “HR Risks Detected – Immediate Action Required”
|
||||
- Email type: **text**
|
||||
- Body expression:
|
||||
- Map from the OpenAI node output (match your node’s actual output structure). In this workflow it is:
|
||||
- `{{$json.output[0].content[0].text}}`
|
||||
- Credentials:
|
||||
- Create/select **Gmail OAuth2** credentials (ensure send permissions)
|
||||
- This is a terminal node on the True branch.
|
||||
|
||||
10. **Add OpenAI node for daily summary**
|
||||
- Add node: **OpenAI (LangChain)**
|
||||
- Model: **gpt-4o-mini**
|
||||
- Messages:
|
||||
- System: “You are an HR operations assistant.”
|
||||
- User: “Create a concise HR daily standup summary grouped by status.” + inject `tasks` as JSON
|
||||
- Connect: `AI Daily Summary` → `Daily HR Summary Email`
|
||||
|
||||
11. **Add Gmail node for daily summary**
|
||||
- Add node: **Gmail → Send**
|
||||
- To: daily recipient(s)
|
||||
- Subject: “Daily HR Action Summary”
|
||||
- Email type: **text**
|
||||
- Body expression: same mapping approach as escalation branch (in this workflow: `{{$json.output[0].content[0].text}}`)
|
||||
|
||||
12. **Validate & activate**
|
||||
- Run once manually to validate:
|
||||
- Monday.com returns expected column structures
|
||||
- dates parse correctly for overdue logic
|
||||
- OpenAI output mapping resolves correctly for Gmail body
|
||||
- Activate workflow.
|
||||
|
||||
---
|
||||
|
||||
## 5. General Notes & Resources
|
||||
|
||||
| Note Content | Context or Link |
|
||||
|---|---|
|
||||
| “## Workflow Overview … (How it works / Setup steps)” including: connect Monday.com, ensure status values include “Done” and “Stuck”, connect OpenAI, configure recipients, adjust schedule timing. | Sticky note: “Workflow overview” |
|
||||
| Disclaimer: *Le texte fourni provient exclusivement d’un workflow automatisé réalisé avec n8n…* | Provided by user (workflow context) |
|
||||
Reference in New Issue
Block a user