Hey everyone,
Welcome back for another bite to chew on.
There is a category of work in every brand that everybody agrees matters, but nobody wants to own.
Reconciliation.
Chasing receivables.
Cash flow forecasting.
It gets done badly, late, or never. So, I just handed the whole category to Perplexity Computer. And in today’s edition, I have receipts to show you how it went and if it’s worth trying for your brand.
This is the first build in a series. One operator with one system, taking over the seats your brand is struggling to fill.
We’re starting with finance ops because it is the least glamorous and the most expensive to ignore.
Everything below is replicable in an afternoon.
On the Menu:
The finance work every lean brand skips, and what it costs
The exact build: two connectors, three standing jobs, and the prompts to copy
Week one receipts, the honest caveats, and the math for a sub-$10M org chart
FREE Trial of Perplexity Computer
You already know reconciliation, receivables, and cash forecasting matter. Nobody on your team owns them, because the hire costs $150K and the work is invisible right up until it’s expensive.
That gap is exactly what this issue is about. Perplexity Computer is how we closed it.
You describe an outcome in plain English, and it deploys subagents to browse, pull data, draft documents, and run the workflow on a schedule, whether your laptop is open or not.
400+ connectors: One-click OAuth into Gmail, Slack, Shopify, QuickBooks, and your banking stack through Plaid, Ramp, and Stripe.
Standing scheduled jobs: Describe the task and the cadence. Computer runs it in the cloud on hourly, daily, or weekly schedules, no machine required.
Subagents that finish the job: Computer breaks the outcome into tasks, solves problems as they come up, and checks in only when it actually needs you.
The Problem
Every DTC brand has an unglamorous middle.
Ad spend that never gets reconciled against what actually hit the card. Zero forward visibility on cash, so every big decision gets made on a gut check of the bank balance.
Lean teams skip this work, and for good reason: the finance hire costs $150K, and nobody celebrates a clean reconciliation - you only notice the function when it fails.
Here is how expensive that failure can get:
In week one, Computer found $3,847 in discrepancies sitting in our own books.
Money we had already lost and did not know about. Keep that number in mind for the rest of this.
The Build
The whole setup took one afternoon. Two connectors, three standing jobs.
Step 1: Connect the stack.
Gmail plus banking. Each connector is one-click OAuth and takes about a minute. Gmail gives Computer the invoices and the AR thread history. The banking connection gives it the actual transactions.
Step 2: Set the three standing jobs.
These run in the cloud on their own schedule. The prompts, verbatim:
Job 1, reconciliation (weekly): "Every Friday, pull every ad platform invoice from Gmail for the past 7 days. Match each one against the corresponding card and bank transactions. Flag any mismatch over $50 with the invoice, the transaction, and the delta, in one table."
Job 2, AR chase (weekly): "Every Monday, review our open invoices. For anything past terms, draft a polite chase email that escalates based on how overdue it is. Queue every draft for my review. Do not send anything."
Job 3, 13-week cash flow (weekly): "Every Monday at 7am, build a 13-week cash view from bank balances, expected receivables, payroll, ad spend at current run rate, and upcoming inventory POs. Flag any week where projected cash dips below $50,000."
That is the entire build.
No spreadsheet templates.
No Zapier chains.
No finance hire.
The Receipts
Week one, with receipts:
Discrepancies caught: $3,847 across six flagged transactions. The biggest one: a Meta invoice billed twice, $2,112 hitting the card in two identical charges nine days apart.
Chase emails drafted: nine drafts queued against $46,200 in receivables past terms. I reviewed, tweaked two, and sent them all in under ten minutes.
The Monday cash view: a 13-week forecast in my inbox before my first coffee. First time this brand has ever had one.
Everything still gets human review.
It flagged two false positives, both legitimate charges that looked like duplicates (one Meta invoice settled across two card transactions).
The first cash forecast missed a $6,000 annual software renewal until I fed it our contracts folder.
And there are things I would not trust it with yet: sending money, sending emails unreviewed, or anything where a wrong action costs more than a wrong answer.
Drafts and flags only. I stay the approver.
The Operator Math
The point is where the return comes from.
Growth-side automation competes with things you already do.
Back-office automation competes with things that never happen at all.
That is why this returns faster than another creative-testing tool ever will.
The bigger unlock is tempo.
Cash decisions that used to happen quarterly, when the accountant surfaced a problem, now happen every Monday. For a sub-$10M brand, that changes the real org chart.
The $150K finance seat becomes a standing system plus one founder hour a week of review.
Sum It Up
The most consequential seat in a lean brand right now might not be the most shiny, but the longer it goes ignored, the deeper the hole gets dug.
On the problem: Unreconciled spend, unchased receivables, and zero cash visibility are not small leaks. They cost us $3,847 in one week's worth of catches.
On the build: Two connectors, three standing prompts, one afternoon. Copy the prompts above and you have a finance function by Friday.
On the math: Back-office automation beats growth-side automation because the baseline is zero. Weekly decisions instead of quarterly ones.
The creator pipeline build is already live, and the retention autopsy is coming next. Same experiment, same rule: one operator, one system.
Chew on This readers get a free trial to Perplexity Computer to run this exact build.
This issue is sponsored by Perplexity. The build, the results, and the screenshots are from our own account.
Let us know how we did...
All the best,
Ron & Ash