# WhatsApp leads into the CRM in seconds, built in one week

Page: https://pixelteh.com/work/alltrust-whatsapp-lead-intake · Last updated: 2026-11-16

_Mortgage brokerage · AI automation · n8n_

Alltrust Mortgage received leads as WhatsApp texts, forwards and screenshots, and an assistant retyped each one into Zoho. One developer built an n8n and Gemini workflow in a week that turns them into clean CRM records.

## Challenge

Partners sent leads as WhatsApp texts and screenshots, and an assistant retyped each one into the CRM, two or three minutes per lead.

## Outcome

An n8n and Gemini workflow, built in a week, turns each message into a clean CRM record the broker approves with one tap, with referral partners linked automatically.

## Key results

- **1 week**: to build
- **Seconds**: from WhatsApp to CRM, down from 2 to 3 min of retyping
- **~30**: leads per week
- **Auto**: referral partner matching

## Facts

- **Client**: Alltrust Mortgage
- **Build**: 1 week, part of an ongoing partnership
- **Team**: 1 developer
- **Status**: Live

## The situation

Partners sent leads all day over WhatsApp: a text, a forwarded chat, a screenshot with half the phone number cut off. An assistant retyped every one into the CRM, two or three minutes each. Typos crept in, co-borrowers and referral sources went missing, and leads cooled while they waited.

## How it works

The broker forwards the message or screenshot. Gemini reads it and pulls out the lead details. The workflow then cleans the data: phone numbers in the CRM's format, emails checked, names split, uncertain fields flagged. A card comes back on WhatsApp with Confirm, Edit and Discard. Edit takes a correction in plain English. Confirm checks for duplicates, creates the record, attaches the original screenshot and adds a note to the timeline.

Alltrust's CRM is heavily customised, so we built the workflow around their real setup rather than a standard one. If a message names the partner who referred the lead, the workflow finds that partner in the CRM and links them automatically. When a name could match more than one partner, it shows the options on the approval card instead of guessing. Co-borrowers land in the right fields, and all settings sit in one place, so the client can adjust the workflow without touching code.

## Results

Leads now reach the CRM seconds after they arrive, and the broker approves each one with a tap. The assistant who used to retype them moved to other work, and typos and dropped fields mostly disappeared. Referral credit is automatic, so Alltrust can see which partners bring in business, reward them properly and put its time into the right relationships.

## What the client said

> The intake was an addition to the original project we were working on. They quickly understood the idea and delivered within a week, which was 4x faster and cheaper than any other offer I had at the time.
>
> Nick S., CEO, Alltrust Mortgage

## Stack

n8n, Google Gemini (vision and text), WhatsApp Cloud API, Zoho CRM REST API v8 (OAuth2)

## Related

- [Add AI to an existing system](https://pixelteh.com/services/ai-integration)
- [Mortgage and lending](https://pixelteh.com/industries/mortgage-lending)
- Next case study: [From a landing page to technical co-founder of a Swiss marketplace](https://pixelteh.com/work/swiss-fintech-advisor-marketplace): From a landing page to technical co-founder: launched in 6 weeks
