Salesforce Intercom Integration: Connect Them for Your AI Agent (2026)
Support and revenue live in two systems, and your agents feel it every day: a conversation lands in Intercom, but the customer's account, plan, open opportunity, and account owner are all in Salesforce — one tab over. A Salesforce Intercom integration is one of the most-requested connections in customer support, and there's more than one way to do it depending on what you actually need.
This guide covers both: the native Salesforce app for Intercom (and where it stops), and how to give your AI agent live Salesforce lookups and actions right inside a conversation using a Custom Tool.
Framing first. Macha runs on top of the Intercom you already use — it's an AI-agent layer, not an Intercom or Salesforce replacement. Fin and your Inbox keep working as they do; Macha reads the conversation, reasons, and reaches into Salesforce to answer or act.
Want the broader picture on connecting any app to Intercom? See our pillar on building custom Intercom integrations — this piece focuses on giving your AI agent live Salesforce access.
The native option — and where it stops
Intercom does offer a native Salesforce app in the App Store. It surfaces Salesforce data — accounts, contacts, leads, opportunities — in the Inbox so a human agent can see who they're talking to. If all you need is passive context on the screen, that may be enough. (Worth knowing: Salesforce is disabling the creation of new Connected Apps by default in its Spring '26 release, moving toward External Client Apps — existing integrations keep working, but if you're wiring up a fresh Salesforce connection today, the setup steps in older guides may differ.)
Where it stops:
- It's mostly display, not action. It shows fields; it doesn't let an AI agent decide to look something up mid-conversation, pass the right parameters, and reason over the result.
- It's fixed. You get the objects and fields the app exposes — not the specific query or write your workflow needs (e.g. "is this account past due?" or "log this conversation to the opportunity").
- It's context, not capability. The app helps a human read Salesforce; it doesn't give your AI agent a live lookup or the ability to act.
For a lot of teams, the sidebar view is fine. For anyone who wants their AI agent to actually use Salesforce, there's a lighter, more flexible path.
What you actually want: the agent using Salesforce live
The high-value version of this integration isn't a displayed field — it's the agent doing the work. A customer messages Intercom; the agent recognizes them, looks up their live Salesforce account by email, and answers "yes, your Enterprise plan includes that" or "I can see your renewal is next month" — grounded in Salesforce, not a stale copy. That's what a Custom Tool enables.
How to connect them with a Custom Tool
A Custom Tool is a single action you define once and point at Salesforce's REST API. No marketplace app, no code.
1. Point it at the Salesforce API. Salesforce exposes a REST API (with OAuth 2.0); you give Macha the endpoint and credentials. The built-in Sidekick can scaffold the request from a plain-English description if you'd rather not fill in the form.
2. Define the action. Give it a label and a description the AI uses to decide when to call it — e.g. "Get the Salesforce account, plan, and renewal date for a customer by email" — set the method and URL (a SOQL query against the Salesforce REST API, with a {{email}} placeholder), and choose Read or Write. A lookup is a Read; updating a case or logging to an opportunity is a Write, which requires confirmation before it runs.
3. The agent uses it live. Once the tool exists, the agent calls it mid-conversation, passes the customer's email, and reasons over the Salesforce record it gets back — answering the question or taking the next step.
Read vs Write, and staying safe
Because a Custom Tool can reach into Salesforce, the guardrails matter: Read tools (lookups) are free to call; Write tools (update a case, log an activity, change a field) require confirmation. Each tool is scoped to one action, permissioned per agent, and every call is audited. Salesforce's OAuth handles credentials, so secrets never live in the conversation.
Teams usually build a few Salesforce tools and group them — a lookup, a "log conversation to opportunity," a "flag account" — alongside tools for their other systems.
Native app vs Custom Tool — which to use
- Use the native Salesforce app if you just want accounts, contacts, and opportunities visible in the Inbox for human agents.
- Use a Custom Tool when you want your AI agent to look up live Salesforce data or take an action from the conversation, you need a specific query or write the native app doesn't expose, or you'd rather work from the live record. The two can coexist.
Beyond the lookup
Once the agent can reach Salesforce, the same pattern covers more: check an account's plan or entitlement before answering, read the open opportunity and owner to route a VIP, log the conversation to the account timeline, or flag churn risk. And because Custom Tools aren't Salesforce-specific, the same approach connects your billing system, data warehouse, or internal API — see the Custom Tools integration directory for 200+ tools, or the full pillar on building custom Intercom integrations.
The bottom line
Intercom's native Salesforce app puts data on the screen. A Custom Tool puts Salesforce in your AI agent's hands — live lookups and actions, decided in context, on any plan, no code. You describe the action, point it at the Salesforce API, and your Intercom agents can use it inside a conversation with Read/Write guardrails and a full audit trail.
Ready to try it? See how Custom Tools work, browse the integration directory, or start a free trial and wire up your first Salesforce tool today.
Frequently asked questions
Does Intercom have a native Salesforce integration? Yes — the Salesforce app in the Intercom App Store surfaces accounts, contacts, leads, and opportunities in the Inbox. It's display-focused; a Custom Tool adds live, agent-driven lookups and actions on any plan.
Do I need an engineer to connect Salesforce to Intercom this way? No. You describe the action in plain English and Macha's Sidekick scaffolds the Salesforce request, OAuth, and parameters. Technical teams can go deeper with full control over the SOQL query, headers, and response mapping.
Can the agent update Salesforce, or only read it? Both. Reads (lookups) run freely; writes — updating a case, logging an activity — require a confirmation step, and you control which agents can use them.
Is it secure? Yes. The tool is scoped to a single action, permissioned per agent, authenticated via Salesforce OAuth, and every call is audited.
Resolve tickets automatically with AI agents
Macha's AI agents work on top of the help desk you already use — no code.
Shopify
Stripe
Slack
Notion
Google Workspace
Confluence

