Every Monday at 8am, this n8n workflow pulls live data from Meta, Google Ads and LinkedIn, sends it through GPT-4o for analysis, builds a branded slide deck and emails it to every client before they open their laptop.
12 hrs
saved per week
vs. manual reporting
2 min
to generate report
after workflow runs
100%
on-time delivery
every single Monday
3×
more accounts
handled per analyst
Typical build: 2–3 week sprint · Fixed price · Zero delivery risk
Trigger
Cron schedule
Avg runtime
~2 minutes
Error handling
Auto-retry ×3
Pulling from 3 platforms, building a deck, writing a summary, formatting, emailing each client. A senior analyst burns an entire Monday on reporting that could be automated.
Manual copy-paste from different dashboards leads to formula errors, mismatched date ranges and slide decks that don't match the data. Clients notice.
Adding a new client means another 2 hours of weekly reporting. Agencies hit a ceiling where they can't take on more clients without hiring more analysts just for reporting.
This is the actual workflow Kovil AI builds and deploys — not a diagram. Here's what runs inside every node.
n8n's Schedule node is configured to run every Monday at 8:00 AM in the client's timezone. No manual trigger, no spreadsheet reminder, no human needed — the workflow starts automatically. The trigger can be reconfigured for any cadence (daily, bi-weekly, end of month).
Three separate HTTP Request nodes run in parallel. Each calls the respective platform's API (Meta Marketing API v19, Google Ads API v16, LinkedIn Marketing API v2) and returns campaign-level data: spend, impressions, clicks, conversions, ROAS, CPC, CPM. All date ranges are calculated dynamically from the trigger date.
n8n's Merge node combines the three platform datasets into a single unified object. A custom JavaScript Function node then normalises field names (e.g. "spend" vs "cost" vs "amount_spent"), converts currencies to a base currency, and flags any campaigns where data is missing or an API returned an error.
A structured prompt sends the full normalised dataset to OpenAI's GPT-4o API. The prompt instructs the model to identify: top 3 performing campaigns, bottom 3 underperformers, week-over-week budget efficiency changes, anomalies (e.g. CPM spikes >30%), and to write 3 concrete recommendations in plain English. Output is returned as structured JSON.
Using the Google Slides API, n8n creates a new presentation by copying a master template (stored in the client's Drive). It then programmatically replaces placeholder text across 5 slides: Slide 1 (executive summary), Slides 2–4 (platform breakdowns with charts), Slide 5 (GPT-4o recommendations). The final report is exported as both a Google Slides link and a PDF.
Gmail API sends a personalised email to each stakeholder on the client's distribution list. The subject line includes the brand name and date range. The body contains a 3-sentence AI-written summary of the key insight from GPT-4o. The PDF report is attached and the Google Slides link is embedded. Sent from the agency's email address via Gmail OAuth.
A Slack webhook posts a formatted message to the agency's #weekly-reports channel. The message includes: total spend across all platforms, blended ROAS, best-performing campaign (with metric), top GPT-4o insight of the week, and a link to the full Google Slides report. Emojis and Slack block kit formatting make it easy to scan in seconds.
Workflow orchestration
The backbone of the automation. All triggers, data merges, API calls, and conditional logic live in n8n. Self-hosted or n8n Cloud.
Ad data source
Pulls campaign-level performance data including spend, impressions, clicks, ROAS and reach. Uses long-lived user access tokens.
Ad data source
Returns search, display and Performance Max campaign metrics via Google Ads Query Language (GAQL). Requires Google Cloud OAuth setup.
Ad data source
Pulls sponsored content and lead gen campaign data. Requires LinkedIn App credentials and organisation admin access.
AI analysis engine
Receives the merged dataset and returns structured JSON: top performers, anomalies, recommendations. Temperature set to 0.3 for consistent, factual outputs.
Report builder
Clones a master template and populates slide placeholders programmatically. No manual slide editing required.
Client delivery
Sends branded HTML emails with PDF attachments from the agency's own Gmail address via OAuth 2.0.
Team notification
Posts a formatted Block Kit message to the internal reporting channel with headline metrics and a report link.
Kovil AI scopes, builds, tests and deploys this workflow end-to-end. You don't touch n8n until it's live and running.
Book a 30-minute discovery call. We'll scope the workflow for your platforms, clients and reporting cadence — fixed price, zero delivery risk.
Typical sprint: 2–3 weeks · Fixed-price · Fully managed delivery · Post-launch support included