- Building AI Agents
- Pages
- Sales CRM Assistant Agent
🤖 AGENT OF THE WEEK (10/16/25)
Welcome back to Agent of the Week:
Normally, your agent has to hit Google Sheets or Airtable every time it needs data.
Now? n8n Tables store everything inside n8n — fewer API calls, faster runs, cleaner builds.
So to test this, I built a CRM Sales Assistant Agent — it keeps client and deal data in n8n Tables, then lets me chat to pull summaries, list won deals, and even sum revenue by month — answering all my sales analytics questions.
One agent, one native CRM table. Simple, but seriously powerful for everyday client questions.
Detailed Guide
Step 1: Create Your Table
Head over to your n8n workspace and click the new Data Tables (Beta) tab.
Create a new table — let’s call it leads_example
.
Then, add your own custom columns. These will serve as the core of your CRM:
Column Name | Type | Description |
---|---|---|
name | string | Client’s name |
company | string | Company name |
source | string | Where the lead came from (LinkedIn, email, etc.) |
lead_stage | string | Stage of the deal (qualified, in progress, closed) |
assigned | string | Team member managing the project |
string | Client email address | |
deal_value | number | Deal amount |
contact_date | date | When the deal was created or updated |
status | string | Deal outcome (won / lost / pending) |
You’ll also notice 3 system-generated columns (id
, createdAt
, updatedAt
) that n8n manages automatically — you don’t need to edit or delete them.
This is now your agent’s native database.
Step 2: Import Data from Google Sheets
If you already have a CRM spreadsheet in Google Sheets, you can migrate it directly.
Create a simple workflow:
Manual Trigger → Google Sheets (Get Rows) → Data Tables (Insert Rows)
When you run it, you’ll see for the example:
“11 items out, 11 items in.”
That’s your entire dataset — moved from Google Sheets into n8n Tables in seconds.
No extra APIs. No sync lag. Just local, structured data inside your agent’s environment.
Step 3: Sales CRM Assistant Agent — Query by Name, Status, or Date
This agent can handle complex questions:
“Show me all projects for Orion Motors.”
“Which deals did we win in September?”
“What’s the total revenue we collected last month?”
Workflow:Chat Trigger → Router (Name / Status / Date Match) → Data Table (Query) → Calculator (Sum) → AI Agent
The logic is simple:
If the message includes a company name, it does a name match.
If it mentions won/lost/pending, it filters by status.
If it mentions a month or year, it does a date match.
Example:
“We completed 4 paid projects in September totaling $16,500 — including Orion Motors, Alpha Health, and Crest Interiors.”
The agent uses n8n’s calculator node to sum deal values and outputs both a breakdown and a final total.
Step 4: Add Guardrails
Agents can hallucinate.
To keep things accurate, always include system-level instructions like:
“Never invent missing data.”
“Only use table fields shown.”
“Show a calculation breakdown when using totals.”
This prevents overconfident text generation and gives you explainable, auditable results.
Step 5: Real Use Cases
Here are a few ways you can apply this CRM Agent today:
Sales teams: Ask “What’s the total revenue this month?”
Project managers: Ask “Which projects are still pending?”
Client success: Ask “Summarize the latest project for Orion Motors.”
Executives: Ask “Give me a weekly pipeline summary by stage.”
Each question can now be answered instantly — no dashboards, no Excel sheets, no delay.