Revenue GPS Dashboard from Stripe, Zoho Books, GoHighLevel and ClickUp

Project description

A growing agency keeps its money in four places: retainers in Stripe, project invoices in Zoho Books, the sales pipeline in GoHighLevel and client agreements in ClickUp. The founder's question is simple, "how much is coming in over the next 90 days and what needs my attention today", and answering it means four logins and a spreadsheet. This n8n workflow answers it every morning at 7:00. I built it as a working demo for an agency's job post, on sandbox data in each API's real response format, and recorded it running end to end.

Tools & Deliverables

• n8n workflow on a 7:00 schedule pulling Stripe subscriptions, Zoho Books invoices, GoHighLevel opportunities and a ClickUp client agreements list
• Source health check on every run: each pull continues on error, so a failed sync is reported and its numbers marked stale instead of silently showing zero
• One tested calculation module: MRR and its movement this month, ARR, invoice aging and collections, pipeline by stage with probabilities, and 30/60/90 day forecasts split into confirmed and speculative
• Cross-system reconciliation: agreements with no billing, fee mismatches between ClickUp and Stripe, overdue invoices, deals with no owner, late follow-ups, scheduled churn and unconfirmed renewals
• Gemini daily brief written from verified figures only, then a guardrail step that extracts every number it wrote and matches it against the data before publishing
• Founder alert on critical issues, and a dashboard page built from each run's snapshot

The problem

• The numbers live in four tools that do not talk to each other. Stripe knows what is billed, Zoho knows what is invoiced and unpaid, GoHighLevel knows what might close, and ClickUp knows what was agreed. None of them knows what the others say.
• Forecasts mix money you have with money you hope for. A signed retainer and a deal at the proposal stage end up in the same total, so the 90 day number looks better than it is.
• The expensive mistakes are the silent ones. A client marked active in ClickUp but never set up in Stripe is revenue that is simply not being billed, and nobody notices until the month end.
• Syncs fail quietly. When an API token expires, the dashboard shows zero or yesterday's number and still looks normal, which is worse than showing nothing.
• An AI summary of financial data can invent figures. A model asked to summarise revenue will produce a fluent paragraph, and one wrong number in it is enough to make the founder stop trusting the whole report.

Approach

Each tool is used for what it is the source of truth for. Recurring revenue, new and churned clients come from Stripe subscriptions. Agreement terms, fees and end dates come from the ClickUp agreements list. Outstanding invoices, due dates and collections come from Zoho Books. Pipeline, stages, owners and follow-ups come from GoHighLevel. The four pulls run in sequence and each one continues on error, so the calculation step knows exactly which sources answered and can mark the rest as stale.

All the money logic lives in one small code module, because forecasts and reconciliation need to be repeatable and testable. Confirmed revenue is retainer billing that falls inside the agreement term plus open invoices due in the window. Speculative revenue is the weighted pipeline, using a probability per stage, plus renewals that are not yet confirmed, counted at 70%. Overdue invoices are kept out of the forecast and listed separately for chasing. The same module then compares the tools against each other and turns every mismatch into a ranked alert. AI comes last: Gemini receives only the verified figures and writes the morning brief, and a guardrail step extracts every number from its text and checks it against the data. If one does not match, the brief is held back and the dashboard shows the numbers alone.

Solution

In the recorded run, all 12 steps went green in about 22 seconds. The dashboard showed $42,900 in monthly recurring revenue, up $9,100 this month after two new clients and one churn, $37,070 in outstanding invoices with $6,820 overdue, and $126,500 of weighted pipeline across seven open deals. The 30 day forecast came out at $57,250 confirmed plus $20,175 speculative, and the 90 day forecast at $123,050 confirmed plus $54,055 speculative, with the split visible line by line.

The reconciliation is where it earned its keep. It flagged a client active in ClickUp at $3,600 a month with no Stripe subscription, which is revenue not being billed, and sent that to the founder as a critical alert. It also caught a $200 a month fee mismatch between ClickUp and Stripe, a 40 day overdue invoice, two deals with no owner and two late follow-ups. The AI brief covered all of it, and the guardrail checked 32 numbers in its text against the data with 0 mismatches before it was published.

FAQ