Why We Stopped Using DocuSign (And Built Our Own E-Signature System)
How we went from paying $15/user/mo for DocuSign to building ContractClaw Sign — with WhatsApp delivery, AI review, and zero per-doc fees.
The $2,160 Problem
In early 2025, we were a small team building BusinessClaw — an AI agent platform for Indian businesses. Three of us needed to send contracts regularly: service agreements to new clients, NDAs with partners, and vendor agreements with BSP providers.
DocuSign Personal plan. $15/user/month. Three users. $540/year.
That did not sound terrible until we looked at what we actually used versus what we were paying for. We sent roughly 20-30 documents per month. We needed signatures, audit trails, and something professional-looking on the receiving end. We did not need PowerForms. We did not need Payments. We did not use Bulk Send. We never touched the advanced field types. We used maybe 15% of what DocuSign offered.
But the pricing was not the real problem. The real problem was what DocuSign could not do at any price.
Three Problems DocuSign Could Not Solve
1. No WhatsApp Delivery
This was the dealbreaker that eventually forced our hand.
Over 80% of our client communication happened on WhatsApp. Every single contract we sent via DocuSign required an immediate follow-up WhatsApp message: "Hey, I just sent you a contract via email — can you check your inbox and sign it?" Half the time, the client would reply: "Can you send it here on WhatsApp? I do not check email regularly."
We could not. DocuSign delivers via email and only email. In 2026, they still do not support WhatsApp delivery.
One client — a property management company in Mumbai that was about to sign a Rs. 3 lakh annual service agreement — literally told us: "I do not check email. Send it on WhatsApp or I will sign it when I come to your office next month." We printed the contract, drove to their office, and got a wet signature. That was the moment we knew DocuSign was wrong for our market.
Average signing turnaround with DocuSign: 2.3 days. Most of that was just waiting for the client to open their email.
2. No AI Integration
We were building an AI agent platform. Our agents could carry on multi-turn conversations, qualify leads, handle objections, schedule meetings, and close deals. The entire experience was conversational and intelligent — until the moment a deal needed a signature.
At that point, the AI agent had to hand off to a human. The human would open DocuSign, select the right template, manually fill in the client details, and send the envelope. Then wait 2-3 days for the client to check their email and sign.
The gap between "AI closes the deal in a 10-minute conversation" and "human manually sends contract, client signs 3 days later" was absurd. We wanted the AI agent to detect buying intent, generate the appropriate contract, and send it for signing — all within the same conversation, with zero human intervention.
DocuSign has an API. We could have built a custom integration. But we would still be paying per-envelope on top of the subscription, and we would still have the email-only delivery problem.
3. $15/User/Month Adds Up
$15/user/month for 3 people was $540/year. Annoying but manageable. But we were planning to grow. At 10 people on the Business plan ($25/user/month), we would be paying $3,000/year for e-signatures. At 25 people: $7,500/year.
For a startup in India, $7,500/year for a feature that is fundamentally a commodity — putting signatures on documents — is a hard number to justify to yourself when there are alternatives at a fraction of the cost. The per-user model specifically penalizes growing teams.
We Evaluated Three Alternatives Before Building
Before deciding to build our own system, we spent three weeks seriously evaluating existing alternatives. We were not eager to build yet another product — we had enough to build already.
DocuSeal: Good API, Hidden Costs
DocuSeal is an open-source e-signature platform with a clean, developer-friendly API. We tested the self-hosted version (free) and evaluated the cloud API.
What worked: The REST API was well-documented and straightforward. Self-hosting via Docker was simple. The form field auto-detection was better than expected — it correctly identified signature, date, and name fields in about 80% of our test PDFs.
What did not work: The cloud API pricing was $20/month base plus $0.20 per document for API-driven signing. For our volume (30 docs/month), that was $26/month — cheaper than DocuSign. But for the kind of volume we anticipated at scale (500-1,000 docs/month across our platform's tenants), we were looking at $120-220/month. The per-document cost made it unpredictable.
Self-hosting eliminated the per-document cost but required ongoing DevOps attention: Docker updates, SSL certificate management, backup scripts, monitoring, and the inevitable 2 AM incident when the container crashes during a signing flow.
And neither option solved our core problem: no WhatsApp delivery.
Verdict: Good for developers embedding basic signing into a product. Not a complete solution for our needs.
Documenso: Four Critical Bugs in Testing
Documenso positions itself as "the open-source DocuSign" — a bold claim. Built on Next.js and Prisma, the codebase was familiar to our stack. The community was active. The vision was compelling. We tested both the cloud and self-hosted versions.
We found four critical bugs during our two-week evaluation:
-
Blank signatures on mobile Safari. The signing canvas rendered correctly but the captured signature data was occasionally blank — the signing pad would freeze mid-signature on iOS Safari, producing an empty signature field in the final PDF. We reproduced this on iPhone 13 and iPhone 15 running Safari 17+.
-
PDF field spoofing. When a PDF contained form fields with names matching Documenso's internal field identifiers, the pre-fill data could overwrite signer input. This was a security issue — a malicious PDF could pre-fill signature fields with arbitrary data.
-
Broken pre-fill for templates. Template merge fields (like
{{client_name}},{{company}}) intermittently failed to populate when the template was used via the API. The web UI worked fine; the API endpoint had a race condition where the template was cloned before merge field values were applied. -
Missing Chromium dependency in self-hosted Docker image. The PDF rendering pipeline required Chromium for HTML-to-PDF conversion, but the official Docker image did not include the Chromium binary. This caused silent failures — documents appeared to process successfully but generated corrupted PDFs. The fix was documented in a community GitHub issue, not in the official deployment guide.
These were not theoretical edge cases. The mobile Safari bug alone would have hit us weekly given that 60%+ of our users sign on iPhones. The PDF spoofing issue was a security concern we could not deploy with.
Verdict: A promising project with an active community, but not production-ready for our reliability and security requirements at the time of our evaluation.
Zoho Sign: Best Existing Option, Still Missing WhatsApp
Zoho Sign was the strongest off-the-shelf alternative. Proven platform. Indian company. INR pricing. Aadhaar eSign support on Professional plans. SOC 2 and ISO 27001 certified.
What worked: Stable. Reliable. Good template system. The free tier (5 docs/month) was genuinely useful for testing.
What did not work:
- No WhatsApp delivery. Email only.
- No AI integration. Zia's role in Zoho Sign is limited to field detection, not contract review.
- Per-user pricing ($12/user/month) had the same scaling problem as DocuSign.
- We were not in the Zoho ecosystem, so the strongest selling point — native Zoho CRM auto-population — was irrelevant to us.
Verdict: If we had been a Zoho CRM shop, we would have picked Zoho Sign and lived without WhatsApp delivery. We were not.
What We Built and How
After three weeks of evaluation, we made the call: build our own e-signature system as a feature of the BusinessClaw platform. Not because we wanted to be in the e-signature business, but because signing was a critical part of our AI agent workflow and no existing tool connected the dots.
Core Technical Decisions
PDF generation: ReportLab. We evaluated WeasyPrint (HTML-to-PDF, struggled with complex layouts), FPDF2 (lightweight but weak Unicode support), and ReportLab (mature, excellent Unicode for Hindi/Tamil/Telugu, reliable output). ReportLab won on Indian language support and production stability. Server-side PDF generation means no client-side Chromium dependency.
Identity verification: OTP from day one. Every signer receives a one-time password via SMS (Twilio/MSG91) or email before they can access the signing interface. This provides a stronger identity link than email-click-only approaches and satisfies Section 3A of the IT Act 2000 for electronic signature validity. We made OTP available on the free tier, not just paid plans.
Document integrity: SHA-256 hashing with RFC 3161 timestamps. Every document is hashed at creation and after each signing event. The hash chain makes post-signing tampering detectable. We integrated RFC 3161 trusted timestamping through a third-party Timestamp Authority (TSA), providing cryptographic proof of signing time that is independent of our servers. This matters in legal proceedings where the timing of a signature is disputed.
WhatsApp delivery: Our own WhatsApp Business API. We already had WhatsApp Business API integration for our AI agent platform's communication infrastructure. Extending it to deliver signing requests was straightforward engineering: create a WhatsApp message template with a "Review and Sign" button, send via the Business API, track delivery and read receipts, and handle the signer's click-through to the signing interface.
AI contract review: Our existing agent framework. We built a contract review skill for our AI agents using Google ADK. When a document is uploaded for signing, the AI scans it for risky clauses (unlimited liability, auto-renewal traps, non-compete overreach), flags missing protections (confidentiality, indemnification, termination rights), and provides a risk score. This was not a separate product — it was a natural extension of the agent platform we were already building.
What Surprised Us
Building the core signing flow was not the hard part. Upload PDF, place fields, send link, capture signature, hash document, generate certificate — this is well-understood engineering with clear specifications. The hard parts were peripheral:
Mobile signing UX on low-end Android. India has hundreds of millions of Android users on devices with 2-3GB RAM and 5.5-inch screens. Getting the signature pad, pinch-to-zoom document view, and field navigation to work smoothly on a Rs. 8,000 Redmi phone took more iteration than we expected. We ended up building a simplified mobile signing interface that loads the document as progressive images rather than rendering the full PDF in the browser.
Email deliverability. Getting signing request emails to land in inboxes (not spam or Promotions) required SPF, DKIM, DMARC configuration, a dedicated sending domain, a 4-week warm-up period, and ongoing monitoring of bounce rates and spam complaints. This was exactly the kind of friction that makes WhatsApp delivery so valuable — it bypasses the entire email deliverability problem.
State-specific Indian compliance. Rental agreements in Maharashtra have different stamp duty rates than Karnataka. Employment agreements in different states have different notice period norms. We built template variants for major Indian states — this was not technically complex, but researching the state-specific requirements and getting them reviewed by legal counsel took longer than the engineering.
The Result
| Metric | Before (DocuSign) | After (ContractClaw Sign) | |--------|-------------------|--------------------------| | Monthly cost | $45/mo (3 users) | $0 (built into platform) | | Avg. signing turnaround | 2.3 days | 4.2 hours | | WhatsApp delivery | Manual follow-up message | Automatic, built-in | | AI contract review | None | Built-in | | Client complaints about signing process | Weekly | Rare | | Documents requiring human follow-up | ~70% | ~10% |
The 2.3-day to 4.2-hour turnaround improvement was almost entirely attributable to WhatsApp delivery. When a signing request arrives on WhatsApp — the app your Indian client already has open and checks dozens of times per day — they sign promptly. When it arrives in email, it competes with 50 other unread messages and gets deferred to "later."
Should You Build Your Own?
Probably not. We built ContractClaw Sign because e-signatures are a core part of our AI agent platform — the signing flow needed to be deeply integrated with our agent conversation system, our WhatsApp infrastructure, and our multi-tenant architecture. For most businesses, buying a tool is the right choice.
But you should not overpay for DocuSign either.
If you are a business in India sending fewer than 50 documents per month, communicating with clients primarily via WhatsApp, and paying $15-25/user/month for DocuSign, you are paying too much for a tool that does not match how your market communicates.
We built the tool we wished we had: free for up to 5 documents per month, Rs. 499/month for unlimited documents, with WhatsApp delivery and AI contract review included. No per-document fees. No per-user scaling traps. No annual lock-in with difficult cancellation.