MessageCustomersDocument.com
Document.comLegal documents

From 11-hour reply times to 6 minutes, on a single inbox.

Document.

12 staff, ~40k MAUsLegal documents · Consumer SaaS
11 hr → 6 min
Median first-response time
73%
AI-resolved without a human
$1,781
Saved on tooling each month
4 → 1
Tools the team now opens
We were paying Intercom $1,800 a month and still firing our team to keep up. We replaced four tools with Message. Our reply time dropped from 11 hours to 6 minutes, and our bill dropped to $19.
MP
Mara Patel
Head of operations, Document.com
The problem

Before Message.

Document.com runs an automated state-document generator. When a customer cannot find the right form, they need help in under 10 minutes or they bounce to a competitor. The legacy stack (Intercom for chat, Zendesk for tickets, Aircall for phone, plus a separate help center) meant agents pivoted across four tabs and lost context every time.

The switches

4 things that changed.

Side by side, what the old stack did and what Message does instead.

Before
Intercom + Zendesk + Aircall + Notion KB
After
Message inbox (chat + ticket + phone + KB)
Before
$1,800 / mo combined
After
$19 / mo Business plan
Before
11-hour median first response
After
6-minute median first response
Before
Agents pivoting across 4 tabs
After
One queue, one screen, one billing line
The story

How Document.com got here.

01

What broke first

The phone integration. Aircall's webhook to Zendesk would drop one call in five on Saturday afternoons. Customers calling about a notarization deadline got dropped, churned, and left a one-star review that Mara had to answer Monday morning. That was the trigger.

02

Why message.com won the bake-off

Document.com tested four tools over a weekend: HelpScout, Front, Intercom Fin, and Message. Message was the only one that handled chat, phone, and ticket from a single queue without three separate billing lines. Mara's words: 'We did not want a vendor diagram, we wanted an inbox.'

03

Migration in 4 days

Mara's team imported 18 months of Zendesk tickets and 9 months of Intercom conversations through the Message migration tool. The widget went live on a Wednesday at 4pm. By Thursday morning the AI was answering 31 percent of questions without escalation, grounded in their own 80-article knowledge base.

04

The reply-time crash

Inside three weeks, the median first-response time fell from 11 hours to 6 minutes. Agents stopped re-asking customers for context. The AI handles 73 percent of the volume now, and the human agents are doing the work humans should do: tricky residential-vs-commercial document edge cases.

In the build

The stack they run.

Products on the Message platform plus the integrations they connected.

Message products
Unified inboxLive chatPhoneKnowledge basemessage.ai
Connected integrations
StripeWordPressGoogle Analytics
Your study is next

Want results like Document.com?

Fourteen days free. No credit card. Same product, same support, same AI as every team in this collection.