AI agents for Salesforce
The enterprise system of record for revenue. AGNT connects to Salesforce so agents can read what happens there, reason about it, and act — with a receipt for every run.
Lead Researcher is a ready-made agent that works in Salesforce — free, and it runs on your machine.
Salesforce and AGNT, in brief
| What it does | Lets an AGNT agent read from and act in Salesforce. |
|---|---|
| Connects with | OAuth 2.0 Bearer — held in the local vault |
| AGNT node | custom-api, or the zapier-action bridge |
| Runs when | A lead or opportunity changes |
| Typical build time | Ten minutes |
| Where data goes | Nowhere by default — AGNT runs on your machine and credentials never leave it. |
What you can automate
Reacts when…
- a lead or opportunity changes
- a task is due
- a report cycle begins
Agents can…
- update records and fields
- create tasks and notes
- extract report data
Teams use it for
- hygiene enforced across the pipeline
- opportunity summaries before every call
- forecast data pulled into real reports
Copy this Salesforce workflow
Most Salesforce builds begin the same way. A run starts when a lead or opportunity changes. The agent reads what arrived, judges it against criteria you wrote in plain language, and acts only where that judgment says it should — typically to update records and fields. That is what makes hygiene enforced across the pipeline something you can hand over rather than merely schedule.
Three nodes carry that: something happens in Salesforce, an agent reads it and forms a view, and it acts only where that view says it should. The definition below is the whole thing — paste it into a new workflow and it loads exactly as drawn.
{
"id": "df52e241-c85a-4c91-a807-b3349ba5af26",
"name": "Salesforce — read, decide, act",
"nodes": [
{
"id": "bc3a541c-bc0c-4c62-a000-a4aa82361ed4",
"text": "Run on a schedule",
"x": 512,
"y": 144,
"isEditing": false,
"type": "trigger-timer",
"icon": "clock",
"category": "trigger",
"isSelected": false,
"parameters": {
"interval": "15",
"unit": "minutes"
},
"description": "Checks Salesforce every fifteen minutes.",
"error": null,
"isActive": false,
"output": null,
"outputs": {}
},
{
"id": "24494f24-85e4-4913-ada6-008d44e98e9e",
"text": "Work out what should change",
"x": 512,
"y": 336,
"isEditing": false,
"type": "agnt-agent",
"icon": "agnt",
"category": "action",
"isSelected": false,
"parameters": {
"instructions": "Decide what needs to happen in Salesforce. Explain your reasoning."
},
"description": "Decides what needs to happen in Salesforce, and why.",
"error": null,
"isActive": false,
"output": null,
"outputs": {}
},
{
"id": "eb244c00-280c-4115-ae41-af0fcd078252",
"text": "Call the Salesforce API",
"x": 512,
"y": 528,
"isEditing": false,
"type": "custom-api",
"icon": "connect",
"category": "action",
"isSelected": false,
"parameters": {
"url": "https://{instance}.my.salesforce.com/services/data/v59.0/sobjects/Contact",
"method": "POST",
"authType": "Bearer",
"headers": "{\"Content-Type\": \"application/json\"}"
},
"description": "POST against the Salesforce API, authenticated with oauth 2.0 bearer.",
"error": null,
"isActive": false,
"output": null,
"outputs": {}
}
],
"edges": [
{
"id": "0d481a8e-7a2a-4a00-ad3e-2a3ea894cafc",
"start": {
"id": "bc3a541c-bc0c-4c62-a000-a4aa82361ed4",
"type": "output"
},
"end": {
"id": "24494f24-85e4-4913-ada6-008d44e98e9e",
"type": "input"
},
"startX": 800,
"startY": 168,
"endX": 512,
"endY": 360
},
{
"id": "9319c1ee-007f-4430-a9ed-37181a295e00",
"start": {
"id": "24494f24-85e4-4913-ada6-008d44e98e9e",
"type": "output"
},
"end": {
"id": "eb244c00-280c-4115-ae41-af0fcd078252",
"type": "input"
},
"startX": 800,
"startY": 360,
"endX": 512,
"endY": 552
}
],
"zoomLevel": 1,
"canvasOffsetX": 0,
"canvasOffsetY": 0,
"isTinyNodeMode": false
}
Install Lead Researcher — it runs on Salesforce
Lead Researcher is a working agent that uses Salesforce. Installing it drops the whole workflow onto your canvas — connected, editable, and yours to change.
Runs on your own machine · See what it does
The system of record needs a librarian
Salesforce holds the revenue truth and depends on people maintaining it. Hygiene enforcement, note-writing and field updates are exactly the repetitive reading-and-writing work agents absorb well, and the payoff is a forecast built on data somebody can trust.
Preparation before every conversation
An opportunity summary assembled before a call — history, open items, what changed since last time — is a small piece of work that rarely gets done under time pressure and reliably improves the conversation when it does.
How to reach Salesforce from a workflow
There is no first-party Salesforce node yet, and a page implying otherwise would cost you an afternoon. What works today is the generic custom-api node pointed at Salesforce’s own API, authenticating with oauth 2.0 bearer.
| What you need | Value |
|---|---|
| Endpoint | https://{instance}.my.salesforce.com/services/data/v59.0/sobjects/Contact |
| Method | POST |
| Authentication | OAuth 2.0 Bearer |
| Header | Authorization: Bearer {oauthAccessToken} |
| Reference | Salesforce API documentation |
Worth knowing: The instance host is returned by the OAuth exchange and differs per org — never hard-code login.salesforce.com for data calls.
Where Salesforce integrations usually go wrong
The instance host is returned by the OAuth exchange and differs per org — never hard-code login.salesforce.com for data calls.
That single detail accounts for most of the time people lose on a first Salesforce build. With it out of the way the endpoint above is all you need, and the workflow below wires it up.
Connect Salesforce in two minutes
- Download AGNT Community Core — free, local-first, no account needed to run.
- Connect with one OAuth sign-in from AGNT’s vault — tokens are stored encrypted on your machine and never leave it.
- Drop the Salesforce node into a workflow or hand it to an agent — the first run produces a receipt you can read line by line.
Salesforce + AGNT — common questions
Can agents update Salesforce records?
Yes — fields, tasks and notes, within the permissions of the connected account.
Can it extract report data?
Yes, which is how forecast numbers reach a narrated weekly report.
Is this safe for a regulated pipeline?
Every write is receipted and consequential actions can sit behind approval gates.