Armand Labs
Builds · Contract operations

Our DocuSign renewal penciled to $43K. So we built the signing system.

The signature was the easy part. The hard part was everything that had to happen before and after it — correctly, every time.

5 min read · August 15, 2026 · by AJ Barragan
Armand Labs signing-system build: DocuSign renewal penciled to $43K a year
The build receipt, published August 15, 2026.
≈4,100Annual send pace
36 / 36Bad test contracts caught
00:51Signature to sealed copy

I opened two DocuSign renewal offers. Both put $25,096.50 on the page. That looked like the annual number. It wasn't.

We had sent 3,787 envelopes in 11 months — roughly 4,100 a year at that pace. One offer included 1,500 sends, then charged $6.90 for every send over the limit. That put another roughly $18,000 on top. The alternate 2,000-send option still penciled to about $39,800.

The renewal math
Base offer$25,096.501,500 sends included
+
Expected overage≈$17,940≈2,600 sends × $6.90
Modeled annual cost at our measured volume≈$43,000

The bill exposed the wrong problem

The obvious reaction was to shop for a cheaper signature tool. But the signature itself was only one stop in the operation.

A contract started as a PDF in our event system. Before anyone could sign it, every field had to be present and in the right place. After signing, the final PDF had to be sealed, the signing history had to become an audit certificate, both records had to be filed, and the finished copy had to reach the client. The expensive product handled the middle. Our team still owned the handoffs.

That changed the assignment. We did not need another place to draw a signature. We needed one chain of custody from contract creation to delivery.

The missing layer
Validate every field
Reject bad documents
Authenticate the signer
Capture signatures
Seal the PDF
Generate the certificate
File and deliver

So we built the whole chain

The new system takes the contract, detects the required fields, and refuses to send a malformed document. It creates the signing session, records what happened, seals the finished PDF, generates the certificate, files the records in SharePoint, and emails the signed copy.

That refusal step matters more than the signature screen. A polished signing experience is worthless if the wrong person gets the document, a required field is missing, or a box lands in the wrong place. The quality gate had to catch the problem before the client ever saw it.

Then we tried to break it

We created 36 deliberately broken test contracts. A naive validator reported nine bad documents as fine. The new gate caught all 36.

36 deliberately broken contracts

Naive check

Nine bad files escaped.

27 caught9 passed

New gate

Every bad file was stopped.

36 tested36 caught

That is the difference between a feature and an operating system. The feature lets a person sign. The system makes sure the right document gets signed, preserves the record, puts it where the business needs it, and finishes the client handoff.

The end-to-end receipt

In one full test, the contract completed 14 fields and captured two drawn marks. The system sealed the signed PDF, generated the audit certificate, filed the records, and delivered the client copy.

00:51From completed signature to sealed client copy
14Fields completed
2Drawn marks captured
1 chainSigned · sealed · filed · delivered

The entire post-signature handoff took 51 seconds. More important, nobody had to remember the next step or move the same contract between three systems by hand.

Most software stops at the signature.
Ours finishes the job.

Validated · Signed · Sealed · Filed · Delivered

Built with Next.js, Supabase, Microsoft 365, and SharePoint.

Something in your business work like this should but doesn't?

Tell me what's broken and I'll tell you what I'd build.

Tell us what's hurting →
← All builds